Process and system for selecting technical solutions to technical problems in an industrial / engineering environment

Abstract

A computer-implemented data processing method of facilitating collaborative development of technical solutions to technical problems in an industrial setting, the collaborative development involving a plurality of users working together to develop the technical solutions, the method comprising steps of: receive from a user, a plurality of technical problem themes related to a technical problem; receive from the user, a plurality of technical criteria against which technical problems can be measured against; inputting the plurality of technical problem themes into a first artificial intelligence data processing engine which processes the plurality of technical problem themes and as a result of the processing, identifies other technical problem themes which are similar to inputted plurality of technical problem themes, where the other technical problem themes have been previously received from at least one other user; inviting a plurality of users to input technical problem statements relating to each of the themes, taking into account the identified other technical problem themes; receiving a first plurality of technical problem statements related to each of the plurality of themes from a plurality of users in response to the inviting; inputting the received first plurality of technical problem statements into a second artificial intelligence data processing engine which processes the first plurality of technical problem statements, assesses their quality based on predefined criteria, and generates improvement recommendations and as a result of the processing, identifies other technical problem statements, where the identified other technical problem statements are similar to the first technical problem statements, and where the other technical problem statements have been previously received from at least one other user.

Description

[0001] PROCESS AND SYSTEM FOR SELECTING TECHNICAL SOLUTIONS TO TECHNICAL PROBLEMS IN AN INDUSTRIAL / ENGINEERING ENVIRONMENT

[0002] Field

[0003] The present disclosure is in the field of computer-implemented methods and systems for leveraging artificial intelligence (Al) and human collaboration to achieve technical advancements in problem identification, solution generation, and innovation acceleration within industrial and engineering environments.

[0004] Background

[0005] Traditional problem-solving approaches in industrial / engineering organisations often suffer from inefficiencies, limited collaboration, and a lack of systematic innovation. These challenges hinder the timely identification and resolution of critical problems, impeding organisational growth and competitiveness. There is a need for a comprehensive system that leverages modern technology, particularly artificial intelligence (Al), to enhance problemsolving efficiency, foster collaboration, and drive innovation within and across organisations.

[0006] Furthermore, existing approaches often fail to harness the full potential of both Al capabilities and human expertise, resulting in suboptimal problem-solving outcomes. There is a need for a novel system that effectively combines Al and human intelligence to address these limitations and foster a more efficient, collaborative, and innovative problem-solving process.

[0007] Additionally, the lack of seamless integration between Al capabilities and human expertise often results in missed opportunities for innovation and suboptimal solutions. The present invention addresses these challenges by introducing a novel system that optimises the interplay between Al and human collaboration to achieve superior technical outcomes.

[0008] It is an object of the present invention to provide a new way of developing and improving systems, processes and technologies to accelerate innovation development and deployment within and between industrial / engineering organisations. Summary of the Invention

[0009] In accordance with a aspects of the present invention, there is provided a computer- implemented method as claimed in claim 1 or claim 32. A corresponding system and computer program are also provided as claimed in other claims.

[0010] Embodiments of the invention may support innovation and development, in particular within an organisation or across multiple organisations in the industrial / engineering environment. Each organisation can leverage the knowledge within its own organisation and / or other organisations to accelerate innovation development and deployment. Embodiments can also incorporate innovation capability assessment and measurement features to provide organisations with actionable insights for continuous improvement. Problems can be identified and addressed that leadership at the organisation were not aware of. Embodiments can also utilise Al to generate and refine ideas based on existing problem statements and previous solutions, further accelerating the innovation process.

[0011] At least some of the users from whom the problem definitions are sourced may be in a first organisation, and at least some of the further users are in at least one predetermined further organisation separate from the first organisation. Thus, the number of users from whom problems and ideas can be sourced can be increased.

[0012] At least some of the users from whom the problem definitions are sourced are in the at least one further organisation. Thus, problems and ideas common to multiple organisations can be identified and technical / industrial / engineering collaborations can be established.

[0013] Embodiments of the present invention can provide a novel approach to problem-solving by establishing a dynamic interplay between Al and human users. Embodiments can leverage Al's computational power, data processing capabilities, and pattern recognition abilities to automate tasks, analyse complex data, and generate novel insights. Simultaneously, the embodiments can empower human users to contribute their domain knowledge, creativity, and critical thinking skills to refine, evaluate, and implement Al-generated solutions. This synergistic collaboration between Al and humans can unlock new levels of efficiency, accuracy, and innovation in solving complex technical problems.

[0014] Brief Description of Figures

[0015] For better understanding of the present invention, exemplary implementations will now be described, by way of example only, with reference to the accompanying drawings in which:

[0016] Figure 1 is an overview of an environment in which a system in accordance with exemplary implementations operate;

[0017] Figure 2 is an overview of the system;

[0018] Figure 3 indicates general steps in implementation of exemplary implementations;

[0019] Figure 4 indicates steps that occur in sourcing problems to be addressed;

[0020] Figure 5 indicates steps in processing problem descriptions to result in problem definitions;

[0021] Figure 6 indicates steps in sourcing proposals to address the problems;

[0022] Figure 7 indicates steps in collating problem descriptions;

[0023] Figure 8 indicates hardware aspect of an example server system on which a system in accordance with exemplary implementations may be stored and run;

[0024] Figures 9a and 9b are a flow chart showing steps in a method according to one implementation of the disclosed technology; and

[0025] Figure 10 is a block diagram showing an example implementation of an artificial intelligence data processing engine, which could be implemented in the system of Fig. 2, according to an exemplary implementation of the disclosed technology.

[0026] Detailed Description of Exemplary Implementations

[0027] Embodiments of the invention relate to a system and process for generating descriptions of problems, and ideas for addressing, partially or wholly, those problems. The system is configured to support sourcing of the problem descriptions from users and sourcing of ideas to address the described problems from other users. Although embodiments are not limited to use with any particular number of users, the system and process are in particular for use by one or more organisations and may make use of the availability of large numbers of users in organisations and / or across communities in which the organisation is active, to support identification of problems and obtaining of ideas to partially or wholly address such problems. The system may be configured to source problems from multiple organisations contemporaneously that may be relevant to the multiple organisations and / or to sourcing of ideas from those multiple organisations.

[0028] An "organisation" herein means an organised group of people with a particular purpose, including but not limited to a business, a government, a healthcare service, university, school or a charity, regardless of whether the organisation is profit making or state funded. A "community" is a group of linked people, for example linked by all residing within a particular geographical area or linked by all working on a shared problem or idea.

[0029] People who use the system are referred to herein as "users" and may be employees, customers or stakeholders of an organisation or members of a community in which an organisation is active, although they are not limited to such.

[0030] The system's unique approach lies in its ability to seamlessly integrate Al and human collaboration at every stage of the problem-solving process. For instance, Al algorithms analyse and categorise problem descriptions, but human users provide the context and domain expertise to refine and prioritise these problems. Similarly, Al generates potential solutions, but human users evaluate and select the most promising options based on their experience and understanding of the specific context. This collaborative approach ensures that the solutions generated are not only technically feasible but also practically relevant and aligned with the organisation's goals.

[0031] Referring to Figure 1, the system 100 is configured for use by users at a plurality of organisations, first to third of the organisations being indicated by the following reference numerals: 102, 104, 106. Each of the organisations includes a plurality of users.

[0032] Referring to Figure 2, the system 100 includes a management module 200, a user data store 202, a problem / idea statement (PSS) store 204, a text processing module 206 and a communications module 208. The management module 200 includes a problem identification submodule 201 and an idea generation submodule 203.

[0033] The user data store 202 is configured to store, for each organisation, information identifying users and, for each of the users, at least one user categorisation tag indicative of the skills and / or knowledge or educational background of the user and / or the function of the user in the organisation and / or a demographic characteristic and / or a group that they work in. Such a group may have an intended purpose of executing a marketing, sales, administrative, research, product development or legal function for the organisation. Intended purposes are not limited to such. The demographic characteristic may be age, for example. Thus, multiple users each have the same user categorisation tag. The management module 200 may be configured to enable or require a user to input the at least one user categorisation tag and / or information enabling generation of the at least one user categorisation tag for that user. Further, the management module 200 may be configured to add and / or validate user categorisation tags by scraping or otherwise obtaining information on users from online sources, for example from social media profiles such as Linkedln®. Embodiments do not require all people in an organisation to be users of the system 100.

[0034] The PSS store 204 is configured to store problem descriptions, problem definitions, idea proposals, processed idea statements and a knowledge base of known failure modes, risks, limitations associated with various technologies, processes, and materials, and a feedback log to store user input on the accuracy and relevance of the system's analysis of unintended consequences.

[0035] The PSS store 204 is further configured to store and manage data from multiple events, including problem themes, problem statements, idea proposals, solution rankings, and user feedback. This data can be used to analyse trends, identify recurring problems, and facilitate connections between users with shared interests or expertise across different events.

[0036] In addition to linking problem definitions and idea statements, the PSS store may also generate and maintain a 'Combined Problem / Solution Map.' This map visually illustrates the relationship between each problem, its sub-problems, and the corresponding proposed solutions.

[0037] Problems and associated sub-problems may define problem chains in a tiered hierarchy which is visualised in the map. The map may show problem statements and technical solutions (also referred to as idea proposals) supplied by users as well as related problem statements / technical solutions identified by the system e.g. using the artificial intelligence processing modules, based on semantic comparison etc. The map can highlight problem and solution gaps per problem chain (e.g. problem statements not linked to technical solutions), allowing users to easily identify areas where further exploration or development is needed. This visual representation enhances the system's usability and facilitates a more holistic understanding of the problem-solving landscape. The map may be output e.g. in graph form via the user interface and may be interactive to allow a user to navigate and inspect individual problem statements / sub-problem statements and associated technical solutions identified for those (sub) problems.

[0038] The problem descriptions set out problems submitted to the system 100 by users in text, audio or visual form, as well as problems automatically identified and catalogued from corporate communications. They are candidate problem definitions. Such problems descriptions may relate to systems or processes within the organisation, be technical or non-technical in nature, or may relate to technological development, for example. For example, a problem description may describe a labour intensive process, the problem being how to automate it. As another example, a problem description may describe how a computer software and / or hardware system is expensive, outdated and lacks flexibility in its configurability.

[0039] The problem descriptions are processed (either by users or the system) to generate the problem definitions. Embodiments are not limited to how this processing is achieved. The problem identification module 201 includes at least one classifier, and is configured to generate category tags for the problem descriptions using the at least one classifier. The at least one classifier may comprise, for example a neural network architecture and be trained to generate at least one category tag further to input of the problem definition. The problem identification module 201 further includes a cost-benefit analysis (CBA) generation component. This component is configured to estimate the costs and benefits associated with potential solutions to each problem definition, generating Al-powered CBA's that can be used to inform the ranking and prioritisation of problems.

[0040] The problem identification module 201 further includes a category generation component. This component is configured to analyse the collected problem descriptions, identifying patterns, trends, and commonalities to automatically generate categories that align with the organisation's goals and priorities.

[0041] Additionally, the problem identification module 201 facilitates problem decomposition, allowing users to break down complex problems into smaller, more manageable subproblems across multiple tiers. The system provides flexibility in adjusting the number of tiers and problems per tier to accommodate varying levels of complexity and granularity. This hierarchical problem decomposition enables a systematic and comprehensive approach to problem-solving.

[0042] In addition to user-submitted problem descriptions, the system 100 includes a corporate communications monitoring module (not shown in the figures for simplicity). This module is configured to:

[0043] • Access various corporate communication channels (chat logs for electronic messaging applications, email archives, transcripts and / or recordings of video calls and conferences, etc.) within the organisation.

[0044] • Utilise NLP techniques to identify and extract potential problem statements from the communication data. This involves identifying patterns, keywords, and phrases that indicate challenges, issues, or areas of improvement.

[0045] • Classify the identified problem statements into relevant categories or themes using pre-defined taxonomies or through the use of machine learning models.

[0046] • The system may employ human-in-the-loop validation, where a subset of automatically identified problems is reviewed by human experts to ensure accuracy and relevance. Feedback from this process can be used to refine the NLP and classification algorithms.

[0047] • The validated problem statements are integrated into the PSS store 204, where they undergo the same processing and analysis as user-submitted problem descriptions.

[0048] By combining Al-powered data analysis with human validation and feedback, the system ensures the accuracy and relevance of identified problems. This collaborative approach enhances the system's ability to uncover hidden issues, anticipate potential challenges, and identify emerging trends.

[0049] The system 100 includes an innovation capability assessment and measurement module. This module is configured to collect and analyse data on various innovation-related factors, such as idea generation rates, solution adoption rates, user engagement, and overall innovation outcomes. By aggregating and analysing this data, the module can generate reports and visualisations that provide organisations with a comprehensive view of their innovation capabilities, strengths, weaknesses, and areas for improvement. This allows leadership to make data-driven decisions regarding resource allocation, process optimization, and strategic planning to foster a culture of continuous innovation.

[0050] The idea proposals are descriptions in a text, audio or video form of how a problem defined in a problem definition can be addressed, partially or wholly. The idea proposals are processed (either by users or the system) to produce idea statements. Embodiments are not limited to how this processing is achieved. The idea generation module 203 may include at least one classifier, and be configured to generate category tags for the idea statements or idea proposals using the at least one classifier. The at least one classifier may comprise, for example a neural network architecture and be trained to generate at least one category tag for the idea statement further to input of the idea statements.

[0051] Classifiers, such as those of the problem identification and idea generation modules 201, 203, that generate tags based on input text are known in the art. In any event, suitable training data may be obtained by initially having humans determine the tags, enabling training of suitable classifiers. For example, to output the subject matter categorisation tags from a classification engine which receives the problem descriptions and idea statements as input, the tags can be represented as input features or labels in a machine learning model. The classification engine can be trained on a dataset that associates problem statements / descriptions or ideas with their corresponding categorisation tags.

[0052] Once the classification engine is trained, it can take a problem statement or idea as input and predict the associated categorisation tags. These tags can be represented as binary variables, with a value of 1 indicating the presence of a particular tag and 0 indicating its absence.

[0053] The output from the classification engine, which consists of the predicted categorisation tags, can then be used as input to a matching algorithm. The matching algorithm can compare the predicted tags of different problem statements or ideas and identify groupings of associated categorisation tags.

[0054] The matching algorithm can employ various techniques such as clustering, similarity measures, or rule-based approaches to group together the problem statements or ideas that share similar categorisation tags. This process allows for the identification of common themes or topics across the problem statements or ideas.

[0055] The groupings generated by the matching algorithm can provide insights into the relationships and associations between different categorisation tags, helping users navigate and explore the problem space based on shared characteristics or themes and connect if desired with those facing similar problems, or developing similar ideas.

[0056] Further, in each case the at least one classifier may be configured to generate at least one primary category tag and at least one secondary category tag, where the primary tag is a broader term than a secondary tag. Again, in the event that suitable training data is unavailable, it can be generated by humans. For example, the classifier may be configured to generate the primary category tags and second category tags as follows. Training Data Preparation:

[0057] The training dataset would be annotated with both primary and secondary tags. Each problem statement or idea should be associated with its primary tag and optionally one or more secondary tags.

[0058] The primary tags should capture the broader categories, while the secondary tags should represent more specific subcategories related to the primary tags.

[0059] Feature Extraction:

[0060] When preparing the input features for the classification engine, primary tags can be represented as separate features / columns.

[0061] Secondary tags can be encoded as additional features / columns. These features can be represented as binary variables, with a value of 1 indicating the presence of a specific secondary tag and 0 indicating its absence.

[0062] Training the Classifier:

[0063] The classifier model should be trained using the prepared dataset with primary and secondary tags.

[0064] The model should be designed to handle multi-label classification, as each problem statement or idea may have multiple secondary tags or none at all.

[0065] Inference and Prediction:

[0066] During inference, the trained classifier takes a problem statement or idea as input and predicts the associated primary tags and secondary tags.

[0067] The output of the classifier will include the predicted primary tags and, if applicable, the predicted secondary tags.

[0068] Managing Primary Tags:

[0069] Primary tags provide the high-level categorisation of the problem statements or ideas. They can be used for broad grouping and initial navigation of the content. The primary tags can be displayed prominently, allowing users to filter and explore the problem statements or ideas based on these top-level categories.

[0070] Managing Secondary Tags:

[0071] Secondary tags offer more detailed information and allow for a deeper level of categorisation within the primary categories.

[0072] The presence of secondary tags can help in refining the search or filtering process by providing more specific subtopics or sub-themes.

[0073] Users can choose to filter or sort the problem statements or ideas based on specific secondary tags, enabling them to explore the content more precisely.

[0074] By incorporating primary and secondary tags into the classification system, the classifier can effectively handle the different levels of categorisation. This allows users to navigate and explore the problem space at both the high-level and more granular levels, facilitating a more nuanced and comprehensive understanding of the problem statements or ideas.

[0075] While the problems descriptions and idea proposals may be submitted in text form, in variant embodiments they may be submitted as audio, video or text and converted to text, video or audio.

[0076] The text processing module 206 is configured to determine if problem descriptions stored in the PSS store 204 relate to the semantically same or a similar problem. The module 206 includes a natural language processing component for converting text into vectors in a latent space (also referred to as embeddings) in a manner that captures semantics and relationships among words, for example using word2vec. The module also includes a trained neural network component, such as a Siamese Manhattan LSTMs, to generate scores indicative of similarity between different problem statements.

[0077] The text processing module 206 is also configured to determine if idea proposals stored in the PSS store 204 relate to the semantically same or a similar idea in the same way. In variant embodiments, separate text processing modules may be provided for processing of the problem descriptions and the idea proposals.

[0078] The text processing module 206 is also configured to identify potential causal relationships between different problem statements, and to identify for a given problem, historical or inprogress projects which are relevant to that specific problem and provide a clear short explanation of how the project information could provide potential benefits, such as who was involved, risk mitigations, lessons learned, possible solutions and approaches to develop and implement at scale. It can also identify potential connections and complementary aspects between the problem and the relevant project and generate novel insights that could inform the problem owner on how to efficiently achieve their desired outcome, including potential unintended consequences. The module can also employ behavioural science techniques to mitigate biases in the analysis and presentation of information, ensuring a more objective and comprehensive evaluation of potential solutions.

[0079] The system can implement various strategies to counteract identified biases in the problem descriptions, idea proposals, and user interactions. This could involve:

[0080] • restating problem descriptions or idea proposals in a more neutral way to avoid leading questions or assumptions

[0081] • encouraging users to consider different viewpoints or stakeholders to broaden the range of potential solutions

[0082] • allowing users to submit ideas anonymously to reduce social pressure and encourage more diverse contributions

[0083] • actively promoting diverse participation and idea generation by seeking input from individuals with different backgrounds, experiences, and expertise

[0084] • incorporating a feedback mechanism where users can rate or flag potentially biassed statements; this data could be used to refine the bias identification and mitigation algorithms over time. The messaging module 208 is configured to send messages to users in accordance with instructions provided by the management module 200 and to receive messages from users and to facilitate feedback from users regarding the identified potential unintended consequences of proposed solutions. Embodiments are not limited to how this messaging is achieved. The messaging module 208 may be configured to send emails to users and receive messages from the users. The messaging module 208 may be configured to integrate with an enterprise messaging system for sending and receiving of messaging within an organisation. The messaging module 208 may provide web-based messaging functionality. The messaging module 208 also enables users to comment on and discuss problems, ideas, and solutions proposed within the system, fostering collaboration and knowledge exchange.

[0085] System 100 is not only designed to address existing problems but also to anticipate and forecast potential future challenges. By leveraging the vast amount of data collected through problem descriptions, idea proposals, and user interactions, the system's Al data processing engines can identify emerging trends, patterns, and correlations that may signal potential future problems. Additionally, by continuously monitoring and assessing innovation capabilities, organizations can identify and address potential barriers to innovation, ensuring a sustained competitive advantage in the long run.

[0086] For instance, the system could:

[0087] • By analysing the frequency and distribution of problem descriptions across various categories, the system could identify recurring themes that may indicate underlying systemic issues or areas of persistent difficulty. This could help organisations proactively address these root causes before they escalate into major problems.

[0088] • The system could analyse the types of solutions proposed for different problems and identify potential resource constraints or limitations that might hinder future innovation efforts. This information could enable organisations to allocate resources more effectively and anticipate potential roadblocks.

[0089] • By tracking the evolution of problem descriptions and idea proposals over time, the system could detect shifts in user priorities or concerns. This could help organisations anticipate changes in the market or technological landscape and adapt their innovation strategies accordingly.

[0090] The system's forecasting capabilities can provide valuable insights into potential future challenges, enabling organisations to take proactive measures to mitigate risks, optimise resource allocation, and stay ahead of the curve in their respective industries.

[0091] The system can also incorporate behavioural science techniques to mitigate biases in the forecasting process, ensuring that predictions are based on objective data analysis rather than subjective interpretations or assumptions.

[0092] Additionally, the system can implement bias mitigation strategies in the forecasting process. For example, the system may use reframing to present forecasts in a neutral manner, encourage users to consider diverse perspectives when interpreting predictions, and promote transparency in the underlying data and assumptions used to generate forecasts.

[0093] The system 100 also includes a user interface module (not shown) which allows interaction with system users, for example to input problem statements and technical solutions (idea proposals), ranking problem statements and solutions, selecting solutions and performing the various other user functions described throughout this disclosure. As mentioned above, the user interface also provides a visualisation interface for visualizing a combined problem / solution map, e.g. as an interactive graph. The user interface typically comprises a web application or other client-server application, with a backend component running on a server (e.g. web server) and a front-end component running on user client devices, e.g. as a web front-end running in a browser.

[0094] In use of the system 100, first an organisation registers with the system 10. The system 100 will then accept registration by users having an email address associated with the organisation. Each user also provides one or more user categorisation tags, which may be selected by the user from a plurality of stored tags. Alternatively, the user provides information enabling the system 100 to generate the at least one tag for the user. Alternatively, an administrator at the organisation may provide registration information for each user to the system 100, including the at least one tag. Referring to Figure 3, in a step 300, problem descriptions are sourced from users. Problem descriptions may be sought from users at the first organisation and optionally at least one of the other organisations. The same or similar problems may be encountered at different organisations. The messaging module will notify users with similar problems / ideas in order to seek collaboration.

[0095] This broad sourcing approach can uncover issues that might otherwise go unnoticed. However, it is important to consider that an over-reliance on user-submitted descriptions might lead to a focus on immediate, visible problems, potentially overlooking deeper, systemic issues that could pose greater challenges in the future. To mitigate this risk, the system could incorporate mechanisms to actively encourage users to consider long-term implications and underlying causes when submitting problem descriptions.

[0096] In a step 302, the problem descriptions are processed to result in a plurality of problem definitions. In a step 304, idea proposals are sourced from users at the first organisations and / or other of the organisations. In a step 306, the idea proposals are processed into idea statements. The idea statements can be made use of at any organisation taking part in the process that has submitted a problem description that the idea statements address.

[0097] Referring to Figure 4, in step 300 of Figure 3, an administrator of the system 100 at the first organisation 102 initiates a step 400 in which a message is sent to users at the first organisation 102 by the system 100 using the stored contact information for the users and handled by the messaging module 208. The message requests submission of problem descriptions identifying problems relevant to the first organisation 102. The message may provide a link to connect users who share one or more problem or idea descriptions.

[0098] The administrator may configure the system 100 to send the message to users at the first organisation 102 having one or more of the user categorisation tags and / or to one or more users selected by the administrator and / or to all users at the first organisation whose contact information is stored in the user data store excluding one or more users selected by the administrator and / or excluding users having one or more user categorisation tags at the first organisation. Embodiments are not limited to how users are selected to be sent the message; the system 100 is preferably configured to provide the administrator with flexibility in this respect.

[0099] The message includes a deadline by which candidate problem descriptions are to be submitted to the system 100. For example, the deadline may allow at least half a day, one day, two days, one week, two weeks or one month from when the messages are sent, assuming that the messages are sent at substantially the same time.

[0100] At step 402, the system 100 receives the problem descriptions. Preferably, the users send the problem descriptions to the system 100 where they are handled by the messaging module 208. The system 100 is configured at step 404 to automatically store the problem descriptions in the PSS store 204.

[0101] Explaining step 302 of Figure 3 in greater detail with reference to Figure 5, at step 500 the system 100 categorises each of the problem descriptions using the classifier. This includes each problem description being input to the classifier and the classifier outputting at least one category tag. The system 100 stores information on correspondence between the category tag and the user categorisation tags.

[0102] The system 100 then processes the problem descriptions to remove duplication. To do this, at step 502 the system 100 uses the text processing module 208 to determine candidate problem descriptions that have the same category tag and that have a minimum degree of semantic similarity. At step 503, the system 100 analyses the relationships between problem descriptions to identify potential causal links, highlighting where one problem may be the cause or consequence of another. At step 504, the system 100 selects the highest ranking problems considering the Al-generated CBA's alongside predetermined criteria and / or a specific problem statement. Alternatively, the system 100 could select all but one of the similar problem descriptions to remove and removes the others. In variant embodiments, the administrator may perform step 502 manually, or step 502 may be automated and step 504 carried out by the administrator at the system 100. In variant embodiments, in this step 504 the administrator may prepare a problem definition based on a plurality of semantically similar problem descriptions, for example for the sake of clarity.

[0103] Thus, the problem descriptions are processed to generate problem definitions. In an optional step, the problem statements are ranked by the system 100. In order to enable the ranking, the system 100 is configured to enable and require each user to submit in association with each problem description an indication of at least one of (any criteria selected by the user): likely costs involved in implementing a solution, if a solution can be found; impact on the organisation of the problem; likely time period with which a solution could be implemented if one could be found. For example, such indications may be provided using a scale. Preferably, when the system 100 generates a problem statement based on multiple problem descriptions, the system determines corresponding indications for the problem statement, based on the indications associated with the problem descriptions. For example, the system 100 may determine numerical averages for each indication for each problem statement using the indications associated with the problem descriptions from which the problem statement derives. The ranking of the problem statements by the system 100 may be carried out using the averages. In addition or alternatively, the ranking may be determined based on the number of problem descriptions from which a problem statement derives. The system 100 may be configured to determine the ranking using a stored algorithm that processes the averages and / or number. Where the ranking is based on the averages, such an algorithm is preferably configured to output a ranked list of the problem statements with the problems ordered by likely cost benefit to the organisation of addressing the problem. In variant embodiments, the administrator may rank the problem statements, based on the averages or otherwise.

[0104] In a variant embodiment, the problem statements are sent in a message to users in the organisation with a request for each of the users to rank (vote) on the problem statements according to criteria set out in the message, for example according to importance to the organisation of addressing the problem. A scale may be used. Replies are received by the system 100 and the system 100 ranks the problem statements accordingly. Ranking of the problem statements is not essential to all embodiments. Explaining step 304 of Figure 3 in greater detail with reference to Figure 6, at step 600 the system 100 determines to which users messages are to be sent. The system 100 then sends the request message to the users with one or more of the problem statements at step 602. The users to whom the message is sent may be determined by the system 100 and include one or more of: users at other of the organisations having a user categorisation tag corresponding to the determined category tag of a problem statement. all users at the first organisation; all users at another of the organisations.

[0105] The system 100 may be configured to enable configuration by a first organisation to preclude predetermined other organisations (for example competitor organisations) from receiving messages with problem statements deriving from the first organisation. Each request message includes a deadline for replies. For example, the deadline may allow at least half a day, one day, two days, one week, two weeks or one month from when the request messages are sent, assuming that the request messages are sent at substantially the same time.

[0106] At step 604, replies with idea proposals are received from users by the system 100 and stored in the PSS store 204 at step 606.

[0107] As to step 306 of Figure 3, with reference to Figure 7, the idea proposals are processed by the text processing module 206 to remove same or similar proposals. In connection with this the text processing module 206 processes at step 700 the proposals relating to a particular problem definition to detect semantic similarity and removes at step 701 those proposals that are similarto greaterthan a minimum extent, that is, the highest ranking ideas can be selected against predetermined criteria. The system 100 then sends at 702 the determined idea statement to organisations as a solution to the corresponding problem statement. While a text processing module 206 is described above and shown in Fig. 2, video or audio processing could also be performed to achieve the same or similar functions as described for the text processing. In an optional step, where multiple idea proposals are received to address a defined problem, the idea proposals are ranked by the system 100. This ranking may be performed by sending to users requests to rank or vote on the idea proposals, and ranking the idea proposals based on the response. The ranking may, additionally or alternatively, be performed by a trained machine learning model. Training data may be obtained by having a human perform the ranking over an initial phase. The ranking may alternatively be performed by humans.

[0108] In another optional step, idea statements, where there are multiple addressing one problem or where they address multiple problems, are ranked by the system 100. This ranking may be performed dependent on the at least one indication associated with the problem definition that the idea statement addresses, using an algorithm stored in the system 100. Additionally, or alternatively, the ranking may be performed by sending to users requests to rank or vote on the idea statements, based on criteria given to the users (e.g. how effectively the idea statement addresses the problem) and ranking the idea statements based on the replies. The ranking may, additionally or alternatively, be performed by a trained machine learning model. Training data may be obtained by having a human perform the ranking over an initial phase. The ranking may alternatively be performed by humans.

[0109] In an example scenario where the method may be implemented, problem descriptions may be sourced from one or multiple separate organisations in the form of companies and idea proposals sourced from the same one or multiple organisations. Particular users to whom requests for problem descriptions or requests for idea proposals are directed may be filtered by user categorisation tag.

[0110] In another example scenario, the method may be implemented to support innovation by organisations in the form of different government departments. For example, an administrator may request users in step 300 to identify problems relating to climate change or energy security. The system's ability to identify causal relationships could be particularly valuable in this context. For instance, it might reveal how a specific policy change (Problem A) has led to unintended consequences in another area (Problem B), allowing policymakers to address the root cause and design more effective interventions. For example, a proposed tax incentive to promote renewable energy (Solution A) could inadvertently discourage investment in other critical areas like infrastructure or education (Problem C), a potential consequence the system would highlight. The method could then be used to source ideas or solutions at step 304 for the identified problems, considering these causal links to prioritise the most impactful solutions.

[0111] Similarly, the method may be implemented to support innovation in healthcare service. For example, problems may be sourced from users having particular user categorisation tags. For example, users working in a particular department, for example radiology, may have a particular tag. Idea proposals may be sourced from users in that department, or other departments where users have predetermined user categorisation tags. By identifying causal relationships between problems, the system could reveal how a shortage of a particular medical supply (Problem A) causes delays in a specific treatment (Problem B), thus prioritising the solution for Problem A and potentially preventing other cascading issues. The system might also flag potential unintended consequences of proposed solutions, such as a new drug intended to treat Problem B inadvertently exacerbating Problem C due to unforeseen side effects.

[0112] In another example scenario, the method may be implemented to support innovation in schools. For example, problems may be sourced from one or more of head teachers; leadership teams; teachers; students; parents, using user categorisation tags. Idea proposals may be sourced from the same users working in other schools for example. The system could identify potential unintended consequences of implementing a new teaching method (Solution X) in one school, such as increased workload for teachers or decreased student engagement in other subjects (Problem Y), thus prompting educators to consider alternative approaches. Additionally, or alternatively, idea proposals may be sourced from users working in other sectors where the nature of the problem is appropriate, that is, depending on the categorisation tag for the problem statement.

[0113] In another example scenario, the method may be used to support innovation at for-profit companies that have common problems or areas where they may improve, such as banks or water companies. For a water company, the system might identify how a new water treatment process (Solution X), while effective at removing a specific contaminant, could unintentionally lead to increased energy consumption or the generation of harmful byproducts (Problem Y).

[0114] In another example scenario, the system 100 is to be used by a small company (for example start-up) community comprising a plurality of such companies. Steps 300 to 306 may be performed for all companies in the community. The system 100 may be administered by an administrator of the community. For example, many such communities are based in serviced offices run by management companies or universities, and they may administer the system 100. The system could help a startup identify how a rapid expansion strategy (Solution X) might lead to quality control issues or unsustainable financial burdens (Problem Y), prompting them to reconsider their growth trajectory.

[0115] Embodiments are not limited to where users are in organisations. In another example scenario, an organisation in the form of a local government may use the method of steps 300 to 306 to improve an aspect of the corresponding geographical area, for example services or the environment. The organisation may source problem descriptions from users in the local area, using user categorisation tags indicating that the user lives in the predetermined geographical area. Users may be encouraged to register by conventional promotional methods, thereby providing suitable contact information, or by filtering by postcode (zip code) where databases of contact information are available. The users may submit problem descriptions for receipt by the system 100 at step 402 relating to an aspect relevant to the relevant area. For example, a proposed solution to improve traffic congestion (Solution A) could unintentionally increase noise pollution in residential areas (Problem B), a consequence the system would alert local officials to.

[0116] In some embodiments of the invention, in order to source idea proposals from users, the system 100 is configured to provide a search interface by which users can search problem and / or idea definitions. For example, the system 100 may enable a user, who may have particular knowledge or skills, to search for problem definitions that align with those knowledge and skills. The user can then submit idea proposals for a found problem definition or collaborate with those with ideas. If a user does not find a problem or idea from the search, System 100 could provide an interface that utilises Al to create a problem or idea from the search.

[0117] The system 100 may be configured with a reward system for users, with rewards being allocated to users depending on, for example, number of submissions of problem descriptions, number of submissions of idea proposals, ultimate rank of such idea statements. The rewards may be in the form of points that may, for example, convert to money. The reward system could be extended to be gamified to make it fun for users to engage with the system.

[0118] The system 100 may facilitate or integrate with events designed to brainstorm problems, ideas, and solutions. These events could be workshops, hackathons, or other collaborative gatherings where users can interact face-to-face, fostering a more dynamic exchange of ideas and enhancing the overall collaborative problem-solving process. The system 100 could provide tools and functionalities to support these events, such as agenda planning, idea submission, and real-time feedback mechanisms.

[0119] In addition to delivering data, the system 100 can provide an ecosystem visualisation feature. This interactive visualisation can map the entire innovation landscape, showcasing connections between problems, ideas, solutions, organisations, themes, categories, and potential investors. By utilising Al-powered algorithms, the system can dynamically cluster and connect these elements based on their relationships, relevance, and potential impact. This visual representation provides a holistic view of the innovation ecosystem, allowing users to easily identify collaboration opportunities, funding potential, and areas for further exploration. Furthermore, this interconnected map can be utilised to enhance other Al functionalities within the system. For example, the visualisation can inform the Al's recommendations for problem-solution matching, investor matching, and identification of knowledge gaps, leading to more accurate and relevant suggestions.

[0120] The system 100 includes an Al-guided implementation and onboarding module. This module provides personalised guidance and support to new users, helping them navigate the system's features, understand its functionalities, and effectively contribute to the problem-solving process. The module may utilise interactive tutorials, contextual help, and intelligent recommendations to ensure users can quickly and efficiently onboard and leverage the system's capabilities.

[0121] The system 100 is designed to support multiple events over time, maintaining a central repository of data generated from each event. This data includes problem themes, problem statements, idea proposals, solution rankings, and user feedback. The system 100 can analyse this data across events to identify trends, recurring problems, and successful solution patterns. Furthermore, the system 100 can leverage this data to connect users with similar interests or expertise across different events, fostering a broader network of collaboration and knowledge sharing.

[0122] A key feature of the system 100 is its ability to facilitate gap identification. By analysing the collective knowledge and data generated through the collaborative process, the system can identify gaps in knowledge, technology, or potential solutions. These identified gaps can then be used to guide further research and development efforts, potentially leading to the generation of novel and patentable inventions. The system 100 can also assist in the patent generation process by providing relevant data and insights to support the drafting of patent applications. Furthermore, the Al-guided implementation and onboarding module can assist in the seamless integration of new features and functionalities, ensuring users can readily adopt and utilise them to enhance their innovation efforts.

[0123] The PSS store links problem definitions and idea statements. In some embodiments, in response to submission of a problem description by a user, the system 100 may search for same or similar problems and retrieve an idea statement for an identified same or similar problem. The system 100 may provide the retrieved idea statement to the user.

[0124] System 100 is not only designed to address existing problems but also to anticipate and forecast potential future challenges. By leveraging the vast amount of data collected through problem descriptions, idea proposals, and user interactions, the system's Al data processing engines can identify emerging trends, patterns, and correlations that may signal potential future problems.

[0125] For instance, the system could:

[0126] • Identify Recurring Problem Themes: By analysing the frequency and distribution of problem descriptions across various categories, the system could identify recurring themes that may indicate underlying systemic issues or areas of persistent difficulty. This could help organisations proactively address these root causes before they escalate into major problems.

[0127] • Predict Resource Bottlenecks: The system could analyse the types of solutions proposed for different problems and identify potential resource constraints or limitations that might hinder future innovation efforts. This information could enable organisations to allocate resources more effectively and anticipate potential roadblocks.

[0128] • Monitor Shifting Priorities: By tracking the evolution of problem descriptions and idea proposals over time, the system could detect shifts in user priorities or concerns. This could help organisations anticipate changes in the market or technological landscape and adapt their innovation strategies accordingly.

[0129] The system's forecasting capabilities can provide valuable insights into potential future challenges, enabling organisations to take proactive measures to mitigate risks, optimise resource allocation, and stay ahead of the curve in their respective industries.

[0130] The system 100 can function as a complete Al innovator. By leveraging advanced Al capabilities, the system can autonomously generate problems, connect disparate ideas, propose novel solutions, identify relevant projects, papers, and patents, secure funding, assess capabilities, incubate ideas, establish governance frameworks, navigate the complexities of implementation, and even identify potential investors to scale innovations. This holistic approach to Al-driven innovation streamlines the entire process, from ideation to implementation and scaling, significantly accelerating the pace of technological advancement. Referring to Figure 8, the system 100, as well as the processes described above, may be stored and run on a server system 800. The server system 800 includes a bus 802 or other communication component for communicating information between components of the server system 900. The server system 800 includes a processor 804 or processing circuit for processing information. The server system 800 also includes a main memory 806, such as a random access memory (RAM) or other dynamic storage device, operatively coupled to the bus 802 for storing information and instructions to be executed by the processor 804. The main memory 806 can also be used for storing intermediate information during execution of instructions by the processor. The system may further include a read only memory (ROM) 808 or other static storage device operatively coupled to the bus 802, for storing static information and instructions for the processor 804. A storage device 810, such as a solid state device, magnetic disk or optical disk, is coupled to the bus 802 for persistently storing information and instructions.

[0131] The system 800 may be coupled via the bus 802 to an electronic display 812, such as a liquid crystal display, for displaying information to a user. An input device 816, for example a keyboard and cursor control device (e.g. a mouse), or touchscreen, for communicating information and command selections to the processor 804. The processes described herein may be implemented by the server system 800 in response to the processor 804 executing a computer program contained in the main memory 806. Such a computer program can be read into the main memory 806 from another computer-readable storage medium, such as the storage device 810. The processor 802 may be a at least two processors in a multi-processing arrangement configured to execute the computer program in the main memory 806. In alternative implementations special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit) may be used in place of or in combination with the computer program. Implementations are not limited to implementation in hardware or in software or in a combination thereof.

[0132] The computer program (also known as software, software application, script, or code) can be written in any form of programming language, including compiled or interpreted languages, declarative or procedural languages, and can be deployed in any form, including as a standalone program or as a module, component, subroutine, object, or other unit suitable for use in a computing environment. The computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communication network.

[0133] The user devices with which users may send and receive messages may be a smartphone, personal computer, laptop, tablet, or other user device configured for communication with the server system 800.

[0134] The flow chart of Figs. 9a and 9b shows an additional series of process flow steps according to an exemplary implementation of the disclosed technology, which could be implemented in the system 100.

[0135] This flow chart will be described using an example technical problem area of how to create a new zero emission jet engine.

[0136] First, the user would access a graphical user interface showing a wide range of technical clusters relating to different technical problems being faced by other users of the platform.

[0137] The user would be able to assess whether or not a cluster already exists that has the same problem using an Al assisted search facility.

[0138] These clusters would have high level details on the problem being solved by the cluster. Should a cluster not exist that already focusses on the chosen problem, the user can establish their own cluster for the solving of the chosen problem.

[0139] • Problem identification and prioritisation module (for problem owners) o At step 901, the user would configure their cluster (with suggestions from the Al, including automatically generated categories) by identifying a series of problem themes in relation to the problem statements they wish to receive, for example, using our designated example of creating a new zero emission jet engine, the series of problem themes could be:

[0140] ■ Combustion System Design ■ Advanced Materials

[0141] ■ Exhaust Gas Treatment

[0142] ■ Power Electronics and Electric Propulsion Integration

[0143] ■ Thermal Management At step 902, the user can also set out criteria for users to measure problems against, for example:

[0144] ■ Impact (1-5): Evaluate the potential impact of solving each problem statement. Consider the significance of the problem in achieving a zero-emission jet engine, the magnitude of the improvement it would bring, and the environmental benefits it would provide.

[0145] ■ Feasibility (1-5): Assess the feasibility of solving each problem statement. Consider factors such as the availability of resources (technological, financial, and human), existing research or technologies that can be leveraged, and any potential obstacles or limitations that may arise during the solution development process.

[0146] ■ Alignment with Strategic Goals (1-5): Evaluate how well each problem statement aligns with the strategic goals and objectives of the organisation or project. Consider whether solving the problem will contribute directly to the overall mission and vision of developing a zero-emission jet engine. User would have the option to then put a value or ranking on each problem theme (this is akin to a bounty that can be received by the problem solver upon selection of their idea). At step 903, these themes are fed into a first artificial intelligence data processing module (e.g., a classifier) which can then identify similar themes in other cluster instances so the user has an awareness of how many other organisations are facing a similar problem in other sectors or organisations. The classifier will work on both receiving inputs in the form of themes, processing the themes to look for similarities and then output the clusters that are already active in these areas. Filters can restrict view by country, by sector, by size etc. At step 904, stakeholders (internally and or externally of an organisation) would then be invited to suggest problem statements relating to each of these themes, for example:

[0147] ■ Combustion System Design

[0148] • How can we design an efficient combustion system for a zeroemission jet engine that minimises greenhouse gas emissions while maintaining the required thrust and performance characteristics. The challenge lies in developing a combustion system that can effectively burn alternative fuels, such as hydrogen or biofuels, without compromising engine efficiency or introducing additional pollutants.

[0149] ■ Advanced Materials

[0150] • How can we develop lightweight and high-performance materials that can withstand the extreme operating conditions of a zero-emission jet engine. The materials need to have excellent heat resistance, mechanical strength, and corrosion resistance to ensure long-term reliability and safety while minimising the weight of the engine components.

[0151] ■ Exhaust Gas Treatment

[0152] • How can we design an effective and compact exhaust gas treatment system for a zero-emission jet engine. This system should be capable of capturing and neutralising or separating pollutants, such as nitrogen oxides (NOx) and particulate matter, generated during the combustion process. The challenge lies in developing a system that can achieve high pollutant removal efficiency while minimising weight, space requirements, and impact on engine performance.

[0153] ■ Power Electronics and Electric Propulsion Integration • How can we optimise the integration of power electronics and electric propulsion systems into a zero-emission jet engine. This includes designing efficient power distribution networks, developing advanced motor control algorithms, and integrating energy storage systems to support the electric propulsion system. The challenge lies in achieving seamless integration, minimising power losses, and ensuring reliable operation under demanding flight conditions.

[0154] ■ Thermal Management

[0155] • How can we implement effective thermal management strategies for a zero-emission jet engine to maintain optimal operating temperatures and prevent overheating of critical components. This involves designing advanced cooling systems, developing efficient heat transfer mechanisms, and optimising airflow management within the engine. The challenge lies in balancing the cooling requirements with the overall engine efficiency and minimising the weight and complexity of the thermal management system. Stakeholders and users can also break the problem statements down in to sub problem themes and statements, across multiple tiers and adjust the number of tiers and problems per tier, for example by adding new branches to the main problem:

[0156] ■ Combustion System Design

[0157] • Fuel Injection System Design o Designing a fuel injection system that can effectively deliver alternative fuels, such as hydrogen or biofuels, into the combustion chamber of a zero-emission jet engine. The challenge lies in ensuring precise fuel atomization, proper mixing with air, and consistent fuel flow rates to achieve optimal combustion efficiency and minimize emissions. • Combustion Chamber Design o Developing combustion chamber configurations and geometries that promote efficient fuel-air mixing and complete combustion of alternative fuels. The challenge lies in designing combustion chambers that can accommodate different fuel properties and combustion characteristics while maintaining stable and controlled combustion processes.

[0158] • Combustion Control and Monitoring o Optimising the combustion process through advanced combustion control algorithms and sensor technologies. This involves developing algorithms to monitor and adjust fuel-air ratios, ignition timing, and flame stability in real-time. The challenge lies in achieving optimal combustion efficiency, minimizing emissions, and ensuring reliable and stable combustion under varying operating conditions.

[0159] • Emission Reduction Strategies o Addressing combustion-related emissions, such as nitrogen oxides (NOx) and unburned hydrocarbons, by implementing post-combustion treatments or alternative combustion techniques. The challenge lies in developing effective methods, such as lean-burn combustion, exhaust gas recirculation, or catalytic converters, to reduce or eliminate these pollutants while maintaining engine performance and thermal efficiency. At step 905, problem statements are then received from stakeholders in response to the invitations. The system's Al module analyses each problem statement for clarity, relevance, and completeness, providing a quality score and suggesting potential improvements. During the receiving process, an artificial intelligence processing model could enable users to more effectively communicate the problem statements through:

[0160] • generating text based on already inputted text - potentially the whole statement from just the title

[0161] • specifying the gaps in the statements that the user has not covered

[0162] • asking questions of the user to clarify the text

[0163] • suggesting evidence to backup claims in the statement. The Al in step 905 would be trained using NLP techniques based on other problem statements, internal and external data. At step 906, all problems (and possibly also the sub problems) are fed into a second artificial intelligence data processing module (again, possibly a classifier) so it can:

[0164] ■ help to cluster a large number of similar problems that are worded differently by users but essentially all mean roughly the same thing (i.e. taking multiple inputs and reducing to an output of a smaller number of consolidated statements)

[0165] ■ take a large number of multiple problems statements and articulate similarities so they can be placed into groups, and automatically generate tags for the groups, for example:

[0166] • After analysing numerous problem descriptions related to the development of a zero-emission jet engine, the Al-powered category generation component might automatically identify themes such as "Sustainable Aviation Fuels," "Electric Propulsion Systems," and "Lightweight Materials Research." These themes can then guide the focus of problem-solving efforts and idea generation.

[0167] ■ connect problem owners either in private (say within the organisation) or in public (outside of the organisation)

[0168] ■ generate a quality score for each problem statement based on predefined criteria and provide recommendations for improving clarity, focus, or scope ■ generates Al-powered CBA's for each problem statement, estimating the potential costs and benefits of addressing the problem. At step 906a, the system, using the second artificial intelligence data processing engine (classifier), will actively search for other users who have previously inputted problem statements that are semantically similar or related to the current user's problem statement. This matching will be based on the content, keywords, and overall meaning of the problem statements. At step 906b:

[0169] ■ IF the identified matching users are from a different organisation than the current user, THEN:

[0170] • The system will present the current user with the profiles and organisation tags of the matching users.

[0171] • The profiles will include relevant information such as the user's name, job title, expertise, and interests.

[0172] • The organisation tags will include relevant information such as the organisation's name, industry, size, and location.

[0173] ■ ELSE (if the matching users are from the same organisation), proceed to the next step. At step 906c, the system will then prompt the current user to decide whether they want to connect with any of the identified matching users. At step 906d:

[0174] ■ IF the current user chooses to connect with a matching user, THEN:

[0175] • The system will automatically merge the clusters (groups of stakeholders working on similar problems) of both users around the specific shared problem. This merging will create a single, consolidated cluster where all stakeholders can collaborate and share knowledge.

[0176] ■ ELSE (if no connection is established), proceed to the next step. At step 906e, the system will analyse the newly merged cluster and the broader ecosystem of clusters to identify potential opportunities for optimization. This includes: • Identifying redundant or overlapping clusters: If the merging of clusters results in redundant or highly overlapping groups, the system will suggest merging or restructuring them for better efficiency.

[0177] • Suggesting cluster reassignments: If certain users or stakeholders would be better suited in different clusters based on their expertise or interests, the system will suggest potential reassignments to optimise the distribution of knowledge and resources across the ecosystem.

[0178] • Balancing cluster sizes: If certain clusters become too large or too small, the system will suggest splitting or merging them to maintain a balanced and efficient ecosystem.

[0179] The ecosystem optimization process can be either automatic (driven by the Al) or semi-automatic (with user input and approval). The goal is to ensure that the cluster ecosystem remains dynamic, adaptable, and optimised for effective collaboration and problem-solving. At step 907, a user and problem owners (or the classifier) then rank each (or certain) problems according to a set of criteria chosen by the user as shown below, including the Al-generated CBA's. The Al-generated quality score can be included as an additional criterion in the ranking process.

[0180] ■ Problem: Designing an efficient combustion system for a zeroemission jet engine.

[0181] • Criteria: a. Environmental Impact: 8 / 10 b. Energy Efficiency: 9 / 10 c. Safety and Reliability: 7 / 10

[0182] • Average Score: (8 + 9 + 7) / 3 = 8

[0183] ■ Problem: Developing lightweight materials for the jet engine components.

[0184] • Criteria: a. Weight Reduction: 9 / 10 b. Durability: 8 / 10 c. Manufacturing Feasibility: 7 / 10

[0185] • Average Score: (9 + 8 + 7) / 3 = 8

[0186] ■ Problem: Enhancing thermal management in the engine to improve efficiency. • Criteria: a. Heat Dissipation: 9 / 10 b. Temperature Control: 8 / 10 c. Thermal Efficiency: 9 / 10

[0187] • Average Score: (9 + 8 + 9) / 3 = 8.7

[0188] ■ Problem: Minimising noise emissions from the jet engine.

[0189] • Criteria: a. Noise Reduction: 9 / 10 b. Regulatory Compliance: 8 / 10 c. Performance Impact: 7 / 10

[0190] • Average Score: (9 + 8 + 7) / 3 = 8

[0191] ■ Problem: Developing advanced control systems for efficient engine operation.

[0192] • Criteria: a. Control Accuracy: 9 / 10 b. Real-time Responsiveness: 9 / 10 c. Integration Compatibility: 8 / 10

[0193] • Average Score: (9 + 9 + 8) / 3 = 8.7 At step 908, the user would then be able to select against the chosen criteria (1, 2 or all 3), to create a ranked list of problem statements. User could also allocate problem owners to said problems (if the suggester is not happy to own the problem through the lifecycle) and mark them as either 'open' (public) or 'closed' (private). The user will then have a list of ranked problems that they can then select for the next ideation phase. The classifier could identify other users facing similar problems (where present) and invite user to 'match'. Should certain problems be matched, say from two different organisations, the user profiles or said organisations and users (tags) will be shared prior to the user receiving an invitation to connect. The classifier could manage this. Upon connection, both organisations' clusters around the specific problem are merged so their respective stakeholders are combined around the problem and subsequent ideation, creating opportunities for collaboration. o Al will input into the classifier its problems, sub problems and can be instructed by the user to support ranking and criteria scoring according to the kind of innovation portfolio the user requires (i.e. ratio of incremental vs architectural) o Matched users then become part of a cluster where they can communicate with each other around the specific problem and collaborate on ideation (this would then merge their communities so they see all public problems within that cluster) o User selects problems, that triggers the ideation phase

[0194] • Idea generation and prioritisation module (for problem solvers) o Upon receipt of the final selected problems the user will move to ideation o At step 909, the user would set the criteria including scoring criteria by which the ideas are to be ranked and evaluated by both the users and the Al module. These criteria may include factors such as feasibility, environmental impact, cost-effectiveness, innovation potential, alignment with strategic goals, and the like. In a concrete example, the criteria may include:

[0195] ■ Feasibility: Evaluate the feasibility of implementing each idea based on technical considerations, available resources, and potential challenges.

[0196] ■ Environmental Impact: Assess the potential environmental impact of each idea in terms of emissions reduction, energy efficiency, and sustainability.

[0197] ■ Cost-effectiveness: Consider the cost-effectiveness of implementing each idea, taking into account factors such as research and development costs, manufacturing expenses, and long-term operational savings. Stakeholders (internally and / or externally) would then be invited at step 910 to suggest ideas for technical solutions relating to each of these problems. Suggested ideas are then received from users / stakeholders in response to the invitations at step 911 and the system evaluates each idea for potential unintended consequences within the given technical system. The Al module also scores each idea based on the predefined scoring criteria (feasibility, environmental impact, cost-effectiveness, etc.), taking into account the quality of the idea, its potential benefits, and its alignment with the project goals. The system also allows users to provide comments and feedback on each other's ideas and solutions. At step 911a, IF an idea is deemed unsuccessful based on the Al evaluation and user scoring, THEN:

[0198] ■ The Al will generate personalised feedback for the user who submitted the idea.

[0199] ■ This feedback will highlight the specific reasons for the low score, suggest potential improvements, and recommend alternative approaches or resources that the user can explore.

[0200] ■ The feedback will be tailored to the user's expertise, interests, and the specific context of the problem statement, providing actionable insights for the user to refine their ideas.

[0201] The personalised feedback aims to encourage continuous learning and improvement among users. By providing specific guidance and recommendations, the Al empowers users to refine their ideas and contribute more effectively to the problem-solving process. For example, ideas could be received against the following selected problems:

[0202] ■ Problem 1: Designing an efficient combustion system for a zeroemission jet engine.

[0203] • Idea 1: Advanced Fuel Injection System Design o Develop a state-of-the-art fuel injection system that incorporates advanced atomisation techniques, such as micro / nano-scale fuel droplets, to improve fuel-air mixing and enhance combustion efficiency. This can be achieved through innovative nozzle designs and precise control mechanisms that optimise fuel injection timing and spray patterns

[0204] • Idea 2 Computational Fluid Dynamics (CFD) Simulation and Modelling: o Utilise CFD simulations and modelling techniques to gain a deeper understanding of the combustion process and optimise the combustion system design. By accurately simulating the flow field, turbulence, and combustion dynamics, engineers can identify areas for improvement, optimise the shape and arrangement of combustion chamber components, and achieve more efficient and cleaner combustion.

[0205] • Idea 3: Multi-Fuel Combustion Capability o Design a versatile combustion system that can accommodate a wide range of alternative fuels, including hydrogen, biofuels, and synthetic fuels. By incorporating adaptable combustion technologies, such as variable fuel injection strategies and combustion chamber geometries, the engine can optimise the combustion process for different fuels, ensuring high efficiency and minimal emissions across various operating conditions.

[0206] ■ Problem Statement 2: Developing lightweight and high-performance materials for a zero-emission jet engine.

[0207] • Idea 1: Advanced Composite Materials o Investigate and develop advanced composite materials, such as carbon fibre-reinforced polymers, with enhanced thermal stability, high strength-to-weight ratio, and excellent corrosion resistance. These materials can be utilised in critical engine components, such as turbine blades and compressor discs, to reduce weight while maintaining structural integrity and durability.

[0208] • Idea 2: Ceramic Matrix Composites (CMCs) o Explore the use of CMCs for high-temperature components, such as turbine nozzles and combustor liners. CMCs offer excellent heat resistance and mechanical properties, enabling engines to operate at higher temperatures, improve fuel efficiency, and reduce emissions. Develop improved manufacturing techniques and design methodologies to optimise CMC integration into the engine.

[0209] • Idea 3: Additive Manufacturing (3D Printing) o Leverage additive manufacturing technologies to produce intricate, lightweight, and custom-engineered components. 3D printing allows for complex geometries and the integration of internal cooling channels, reducing weight while maintaining structural integrity. Additionally, it offers the possibility of creating hybrid materials by combining different alloys and composites, further enhancing performance and weight savings. The ideas from a wide range of inputs will be received by the classifier which will perform a similar function as the problem identification phase (identify similar ideas in other cluster instances so the user has an awareness of how many other organisations are developing similar ideas in the same sector, other sectors or organisations). The classifier will work on both receiving inputs in the form of idea statements, processing the ideas to look for similarities and then output the clusters that are already active in these areas. Another module will then invite participation in existing clusters where desired and marked as public or ask the user if they would like to establish their own cluster (either public or private). A visual representation would show the size, region and membership of the clusters.) Users (and / or the ranking engine provided by Al) then score the ideas, for example:

[0210] ■ Idea 1: Advanced Fuel Injection System Design

[0211] • Feasibility: 8

[0212] • Environmental Impact: 9

[0213] • Cost-effectiveness: 7

[0214] • Average Score: (8 + 9 + 7) / 3 = 8

[0215] ■ Idea 2: Computational Fluid Dynamics (CFD) Simulation and Modelling

[0216] • Feasibility: 7

[0217] • Environmental Impact: 8

[0218] • Cost-effectiveness: 6

[0219] • Average Score: (7 + 8 + 6) / 3 = 7

[0220] ■ Idea 3: Multi-Fuel Combustion Capability

[0221] • Feasibility: 9

[0222] • Environmental Impact: 9

[0223] • Cost-effectiveness: 8

[0224] • Average Score: (9 + 9 + 8) / 3 = 8.67 (rounded to 2 decimal places) At step 912, a third artificial intelligence data processing module (for example, a classifier) will then receive as inputs, the received ideas. The Al module assesses potential negative impacts of each idea on the system as a whole, considering factors like resource depletion, safety concerns, or ethical implications. The classifier could also receive inputs from: ■ Ideas that already exist in other clusters that solve the same problem but are in the process of being developed.

[0225] ■ Existing products and services from a wide range of companies that already have a product that solves the problem, this forms the backbone of the 'innovation marketplace'. The user will be able see how many products are available, the degree to which they will solve their problem, and an option to connect with the company who has the product. Thus preventing wasted effort in generating ideas for problems that have already been solved.

[0226] ■ The knowledge base in PSS store 204 to assess potential unintended consequences, such as identifying known failure modes or risks associated with the proposed technical solutions.

[0227] ■ User comments and feedback to identify potential strengths, weaknesses, and areas for improvement in proposed ideas and solutions. It can then generate Al-powered comments and suggestions to further refine and develop these ideas. At step 912a, the system presents users with an analysis of potential unintended consequences for each proposed solution, allowing them to refine or revise their ideas. At step 913, a ranking of the received ideas is then carried out, combining the Al scores with the user scores according to a weighted average or other aggregation method defined by the user. The system generates a summary of insights for each problem-solution pair, highlighting key benefits, potential risks, and estimated impact on the organisation's goals. At step 914, the user can then be prompted by the platform to select the highest ranking ideas (technical solutions) either based on the ranking from the participants or the ranking of the classifier. The system may then output the selected technical solution(s), e.g. via messages sent to one or more other users of the system and / or may flag the solution(s) as the selected one(s), e.g. by highlighting it in the problem / solution map or elsewhere in the application interface. At step 914a, users may provide feedback on the system's analysis of unintended consequences, for example indicating whether the identified risks are accurate, relevant, or require further clarification At step 915, selected solutions undergo a circular evolution process. They are re-evaluated by the Al and user communities, potentially generating new problem statements or sub-problems. This process allows for continuous improvement and refinement of solutions, ensuring that they address the evolving needs and challenges of the technical landscape. Categories and tags applied by the classifie r(s) to the problem statements and / or idea proposals (technical solutions) could include such items as:

[0228] ■ Sector Categories:

[0229] • Aerospace

[0230] • Automotive

[0231] • Maritime

[0232] ■ Job profile categories:

[0233] • Chief Engineer

[0234] • Engineer Manager

[0235] • Engineering Apprentice

[0236] ■ Organisational Categories:

[0237] • More than 10,000 employees

[0238] • More than 1,000 employees

[0239] • Less than 50 employees

[0240] ■ Technical Categories:

[0241] • Combustion System Design

[0242] • Materials Development

[0243] • Exhaust Gas Treatment System

[0244] ■ Engineering Disciplines:

[0245] • Mechanical Engineering

[0246] • Electrical Engineering

[0247] • Aerospace Engineering

[0248] ■ Environmental Impact: • Emissions Reduction

[0249] • Energy Efficiency

[0250] • Sustainable Design

[0251] ■ Research and Development:

[0252] • Advanced Fuel Injection

[0253] • Computational Fluid Dynamics

[0254] • Multi-Fuel Combustion

[0255] ■ Implementation Strategies:

[0256] • Innovative Technologies

[0257] • Advanced Manufacturing

[0258] • Integration Solutions

[0259] ■ Cost Considerations:

[0260] • Cost-effective Solutions

[0261] • Long-term Savings

[0262] • Value for Money

[0263] ■ Feasibility Assessment:

[0264] • Technical Feasibility

[0265] • Resource Availability

[0266] • Potential Challenges

[0267] ■ Sustainability Focus:

[0268] • Carbon Neutrality

[0269] • Renewable Energy Sources

[0270] • Lifecycle Analysis

[0271] The classifiers (or artificial intelligence data processing engines) discussed above, can perform their processing with dynamically changing data, such that they continually improve their categorisation results over time (e.g, re-training is done continually). Specifically, training a classifier using data from both within the organisation and publicly available data will likely involve several steps so here is a detailed technical description of the potential process. Data Collection:

[0272] • Gather relevant problem statements or ideas from within the organisation, ensuring they cover a wide range of topics and categories.

[0273] • Obtain publicly available data related to the problem domain from reputable sources. This can include research papers, industry reports, online forums, or open datasets.

[0274] Data Annotation:

[0275] • Prepare the collected data for annotation by establishing a tagging scheme that includes primary and secondary tags.

[0276] • Assign primary and secondary tags to each problem statement or idea in the dataset, ensuring consistency and accuracy in the annotation process.

[0277] • It is important to have a diverse and representative set of annotations that cover various primary and secondary tags.

[0278] Data Pre-processing:

[0279] • Pre-process the collected data by performing tasks such as text cleaning, tokenisation, and removal of stop words to prepare the text data for the classifier.

[0280] • Convert the text data into a numerical representation using techniques like TF-IDF (Term Frequency-Inverse Document Frequency) or word embeddings such as Word2Vec or GloVe.

[0281] Dataset Splitting:

[0282] • Split the dataset into training, validation, and testing sets. Typically, an 80-10-10 split is used, but the ratio can be adjusted based on the available data size.

[0283] • Ensure that the distribution of primary and secondary tags is maintained across the splits to prevent bias in the model. Model Selection:

[0284] • Select an appropriate classifier model based on the nature of the problem, available data size, and complexity of the classification task.

[0285] • Commonly used models for text classification include Naive Bayes, Support Vector Machines (SVM), Random Forests, or deep learning models such as Convolutional Neural Networks (CNN) or Transformer-based models like BERT or GPT.

[0286] Model Training:

[0287] • Train the selected model using the training dataset.

[0288] • Fine-tune the model by adjusting hyperparameters, such as learning rate, batch size, and regularisation techniques, to improve performance.

[0289] • Monitor training metrics such as accuracy, precision, recall, and Fl score to assess the model's progress.

[0290] Model Evaluation:

[0291] • Evaluate the trained model using the validation dataset to assess its performance on unseen data.

[0292] • Analyse evaluation metrics and make adjustments as necessary, such as tuning hyperparameters or exploring different model architectures.

[0293] Incorporating Publicly Available Data:

[0294] • Combine the annotated data from the organisation with the publicly available data to create an augmented dataset.

[0295] • Pre-process and format the publicly available data to match the input format of the model.

[0296] • Merge the augmented dataset with the existing dataset from the organisation, ensuring an appropriate balance of the data sources.

[0297] Retraining and Fine-tuning:

[0298] • Retrain the model using the augmented dataset, including both the organisation's data and the publicly available data. • Repeat the training and evaluation steps, optimising the model's performance based on the combined dataset.

[0299] Model Deployment:

[0300] • Once the model achieves satisfactory performance, deploy it for inference on new problem statements or ideas.

[0301] • Set up an appropriate pipeline to pre-process the input data and make predictions using the trained model.

[0302] It would naturally be crucial to ensure compliance with data privacy and security regulations when incorporating publicly available data, taking into account any licensing requirements or restrictions on data usage. Also, data storage could be added.

[0303] The functions described above with respect to the text processing module 206 (e.g., determining if two problem descriptions are the same or similar) and the messaging module 208 (e.g., sending and receiving messages from users) can be combined with the functionality described above in conjunction with Figs. 9a / 9b, to improve the overall functionality of the described technology.

[0304] The classifiers mentioned above, for example, in conjunction with the problem identification submodule 201, the solution generation submodule 203 and / or the text processing module 206, may be trained separately for different domains to produce at least two separate domain modules. The domain modules may be trained in a specific manner as described below in conjunction with Fig. 10.

[0305] When training the classifier with the described architecture involving multiple Al domain modules (1001, 1002 and 1003 in Fig. 10, only three are shown in the Fig.10 but many more of them could be included), an inter-domain mediation module (1004 in Fig. 10), a crossdomain inference module (1005), a retraining engine (1006), and an analysis engine (1007), the process could be outlined as follows: Data Collection:

[0306] • Gather training data for each domain separately, ensuring that each dataset represents the specific domain knowledge.

[0307] • Collect a diverse range of problem statements or ideas from within the organisation for each domain.

[0308] Training of Al Domain Modules (1001, 1002, 1003):

[0309] • Train the first Al domain module (1001) using the first set of domain knowledge, utilising the collected problem statements or ideas specific to that domain.

[0310] • Train the second Al domain module (1002) using the second set of domain knowledge, using the corresponding problem statements or ideas.

[0311] • Repeat for any further domain modules (1003)

[0312] Inter-domain Mediation Module (1004):

[0313] • Train the inter-domain mediation module 1004 to infer connections between the different domain modules.

[0314] • The module should be capable of inferring terminology translations or mappings between the vocabularies used in each domain.

[0315] Cross-Domain Inference Module 1005:

[0316] • Train the cross-domain inference module 1005, which uses information from multiple domain modules to make inferences.

[0317] • This module can take input from both the first 1001 and second 1002 Al domain modules (and any further domain modules) and leverage the inferred terminology translations from the inter-domain mediation module 1004.

[0318] • The module also cross-references proposed solutions with the knowledge base to identify potential risks or unintended consequences that might arise due to known limitations or failure modes of specific technologies or approaches.

[0319] • The cross-domain inference module 1005 not only leverages terminology mappings but also analyses co-occurrence and temporal patterns in problem descriptions to suggest causal relationships. Retraining Engine 1006:

[0320] • Use the cross-domain inference module 1005 to generate revised training data based on the inferences made across domains.

[0321] • The retraining engine 1006 incorporates the revised training data into the training process for each Al domain module (1001, 1002, 1003), improving their performance.

[0322] • The retraining engine also incorporates user feedback from the feedback log, adjusting the models' weights and parameters to improve the accuracy of future unintended consequence analysis.

[0323] Analysis Engine 1007:

[0324] • Train the analysis engine 1007 to suggest solutions for a specific domain, leveraging information from another domain.

[0325] • This engine utilises the cross-domain inferences, Al-generated categories, the problem-solution matching algorithm to generate recommendations and insight summaries within a particular domain.

[0326] • The architecture allows for the integration of multiple domain modules, enabling the sharing of information and knowledge across domains. The inter-domain mediation module 1004 facilitates communication and translation of terminologies, while the cross-domain inference module 1005 leverages the collective knowledge from multiple domains to make more informed inferences.

[0327] • The analysis engine 1007 utilises the problem-solution matching algorithm to generate recommendations and insight summaries.

[0328] The retraining engine 1006 ensures that the training data is updated based on the crossdomain inferences, improving the performance and adaptability of the Al domain modules 1001, 1002, 1003. Finally, the analysis engine 1007 incorporates information from different domains to suggest solutions specific to a particular domain, benefiting from the insights gained from other domains.

[0329] The analysis engine 1007 is further configured to identify gaps in knowledge, technology, or potential solutions based on the analysis of data. The engine can also facilitate the generation of patentable inventions by providing relevant data and insights based on the identified gaps.

[0330] The above described features regarding the training of the classifiers can be applied to a wide field of industrial / engineering settings, in order to find appropriate technical / engineering solutions to technical / engineering problems. Examples will now be provided.

[0331] Network Engineering:

[0332] • The first Al domain module in network engineering is trained on a dataset specific to networking protocols, network infrastructure, and cybersecurity.

[0333] • The second Al domain module in network engineering is trained on a dataset focused on network performance optimization, network traffic management, and quality of service.

[0334] • The inter-domain mediation module infers connections between the two domain modules, enabling terminology translation between networking protocols and performance optimization strategies.

[0335] • Example connection: The inter-domain mediation module identifies the relationship between network protocols and network traffic prioritisation techniques, enabling the optimization of network performance while considering cybersecurity measures.

[0336] Automotive Engineering:

[0337] • The first Al domain module in automotive engineering is trained on a dataset specific to vehicle dynamics, engine performance, and emissions control.

[0338] • The second Al domain module in automotive engineering is trained on a dataset focused on autonomous driving systems, sensor fusion, and advanced driver assistance systems (ADAS).

[0339] • The system could compare a proposed lightweight material against the knowledge base, highlighting potential risks like reduced durability under certain conditions or increased manufacturing complexity.

[0340] • The inter-domain mediation module infers connections between the two domain modules, enabling terminology translation between vehicle dynamics and ADAS algorithms. • Example connection: The inter-domain mediation module establishes the link between vehicle dynamics parameters and sensor fusion algorithms, allowing the optimization of sensor data processing for enhanced autonomous driving capabilities.

[0341] Chemical Engineering:

[0342] • The first Al domain module in chemical engineering is trained on a dataset specific to chemical reactions, process optimization, and safety protocols.

[0343] • The second Al domain module in chemical engineering is trained on a dataset focused on environmental impact assessment, waste management, and sustainable practices.

[0344] • The inter-domain mediation module infers connections between the two domain modules, facilitating terminology translation between reaction kinetics and sustainability metrics.

[0345] • When evaluating a new chemical process, the system could access the knowledge base to identify potential safety hazards or environmental impacts based on similar past failures.

[0346] • Example connection: The inter-domain mediation module identifies the correlation between reaction kinetics and environmental impact, enabling the selection of sustainable reaction conditions for improved process efficiency and reduced ecological footprint.

[0347] By establishing these connections between the disciplines, the system can leverage the knowledge and insights from one domain to enhance the understanding and decisionmaking in another domain. For example, in network engineering, the insights from performance optimization strategies can inform the development of more secure network protocols. In automotive engineering, the understanding of vehicle dynamics can contribute to the enhancement of sensor fusion algorithms for autonomous driving. In chemical engineering, the integration of sustainability metrics into reaction kinetics can guide the design of environmentally friendly processes. Overall, these connections enable a multidisciplinary approach, promoting innovation and synergy between different engineering disciplines.

[0348] Further connections could include:

[0349] Network Engineering and Automotive Engineering:

[0350] • Example connection: The first Al domain module in network engineering, trained on network protocols and cybersecurity, can be connected with the second Al domain module in automotive engineering, trained on autonomous driving systems and sensor fusion.

[0351] • Connection example: The inter-domain mediation module identifies the connection between network security protocols and the secure transmission of sensor data in autonomous vehicles, ensuring the integrity and confidentiality of data during communication.

[0352] Network Engineering and Chemical Engineering:

[0353] • Example connection: The first Al domain module in network engineering, trained on network performance optimization, can be connected with the second Al domain module in chemical engineering, trained on process optimization and safety protocols.

[0354] • Connection example: The inter-domain mediation module identifies the correlation between network traffic management techniques and the optimization of chemical processes, allowing for efficient resource allocation and improved safety measures within chemical plants.

[0355] Automotive Engineering and Chemical Engineering:

[0356] • Example connection: The first Al domain module in automotive engineering, trained on vehicle dynamics and emissions control, can be connected with the second Al domain module in chemical engineering, trained on environmental impact assessment and sustainable practices.

[0357] • Connection example: The inter-domain mediation module establishes the link between vehicle emissions and environmental impact, enabling the development of sustainable practices in automotive manufacturing and the implementation of eco- friendly materials and processes.

[0358] Automotive Engineering, Chemical Engineering, and Network Engineering:

[0359] • Example connection: All three Al domain modules in automotive engineering, chemical engineering, and network engineering can be interconnected through the cross-domain inference module.

[0360] • Connection example: The cross-domain inference module leverages knowledge from all domains to make comprehensive inferences, such as optimally routing autonomous vehicles based on network traffic conditions, considering environmental impact, and ensuring data security during transmission.

[0361] By establishing these connections between network engineering, automotive engineering, and chemical engineering, the system can leverage the knowledge and insights from each discipline to enhance the understanding and decision-making across domains. These connections enable a holistic approach, promoting collaboration and innovation between different engineering fields, and ultimately leading to advancements in areas such as autonomous driving, sustainable manufacturing, and secure network infrastructure.

Claims

CLAIMS1. A computer-implemented data processing method of facilitating collaborative development of technical solutions to technical problems in an industrial setting, the collaborative development involving a plurality of users working together to develop the technical solutions, the method comprising steps of:(a) receiving a first plurality of technical problem statements from a plurality of users;(b) inputting the received first plurality of technical problem statements into a first artificial intelligence data processing engine which processes the first plurality of technical problem statements and as a result of the processing, identifies other technical problem statements, where the identified other technical problem statements are similar to the first technical problem statements, and where the other technical problem statements have been previously received from at least one other user;(c) receiving from a user a group of selected technical problem statements, selected from the first plurality of technical problem statements and the other technical problem statements;(d) inviting a plurality of users to input technical solutions to the selected technical problem statements;(e) receiving, from a plurality of users, a plurality of technical solutions to the selected technical problem statements in response to the inviting step (d);(f) inputting the received plurality of technical solutions into a second artificial intelligence data processing engine which processes the plurality of technical solutions and as a result of the processing, identifies other technical solutions, where the identified other technical solutions are similar to the received plurality oftechnical solutions, and where the other technical solutions have been previously received from at least one other user; and(g) receiving from a user at least one technical solution, selected from the plurality of technical solutions received in step (e) and other technical solutions identified in step (f), as a selected technical solution to at least one of the selected technical problem statements.

2. The method of claim 1, further comprising providing an output to one or more users indicating the selected technical solution(s).

3. The method of claim 1 or 2, comprising: receiving, from a user, a plurality of technical criteria against which technical problems can be measured; and ranking the plurality of technical problem statements and the other technical problem statements identified in step (b) against the plurality of technical criteria, wherein the step (c) of receiving from a user a group of selected technical problem statements is performed based on the results of the ranking step.

4. The method of any of the preceding claims, comprising: receiving from a user a plurality of technical criteria against which users can measure technical solutions to the group of selected technical problem statements; and ranking the received technical solutions and the other technical solutions identified in step (f) against the technical criteria; wherein the step (g) of receiving from a user at least one selected technical solution is performed based on the result of the ranking step.

5. The method of any of the preceding claims, comprising: receiving from a user, a plurality of technical problem themes related to a technical problem;inputting the plurality of technical problem themes into a third artificial intelligence data processing engine which processes the plurality of technical problem themes and as a result of the processing, identifies other technical problem themes which are similar to inputted plurality of technical problem themes, where the other technical problem themes have been previously received from at least one other user; inviting a plurality of users to input technical problem statements relating to each of the themes, taking into account the identified other technical problem themes identified using the third artificial intelligence data processing engine; and receiving the first plurality of technical problem statements in step (a) in response to the inviting step.

6. The method of any of the preceding claims wherein at least one of the first, second and third artificial intelligence data processing engines includes a plurality of domain modules, where each of the plurality of domain modules has been trained with training data representing domain knowledge gathered for each of a respective plurality of technical domains.

7. The method of claim 6, wherein at least one of the first, second and third artificial intelligence data processing engines further includes an inter-domain mediation module, trained to infer connections between at least two of the plurality of domain modules.

8. The method of claim 7, wherein the connections include mappings between vocabularies used in each technical domain.

9. The method of claim 8, wherein at least one of the first, second and third artificial intelligence data processing engines further includes a cross domain inference module, which leverages the mappings between the vocabularies to make inferences across domains.

10. The method of claim 9, wherein at least one of the first, second and third artificial intelligence data processing engines further include a retraining engine which generates revised training data for the plurality of domain modules based on the inferences made across domains by the cross domain inference module.

11. The method of claim 10, wherein at least one of the first, second and third artificial intelligence data processing engines further includes an analysis engine which utilises the inferences made across domains to suggest a technical solution in a first technical domain using a first domain module which has been trained via the retraining engine, to leverage information from a second technical domain in suggesting the technical solution.

12. The method of any of the preceding claims, wherein receiving the first plurality of technical problem statements comprises, for one or more of the problem statements, using an artificial intelligence processing model to generate and output one or more improvement suggestions for the problem statement, optionally including one or more of: generating additional problem statement text based on provided problem statement text; identifying missing information in a problem statement; generating one or more questions for a user for clarifying the problem statement; suggesting additional information useful for a user in expanding the problem statement.

13. The method of any of the preceding claims, wherein receiving the first plurality of technical problem statements comprises processing a plurality of communications associated with a communication system, and deriving one or more of the problem statements automatically from the communications, wherein the communications are optionally electronic communications transmitted via an electronic mail or messaging system.

14. The method of claim 13, wherein the problem statements are derived by applying natural language processing and / or machine learning techniques to the communications.

15. The method of claim 13 or 14, comprising retrieving the communications from one or more of: chat logs of a messenger application, email archives, and recordings and / or transcripts of video calls.

16. The method of any of the preceding claims, wherein the technical problem statements comprise a hierarchy of problem statements arranged across multiple tiers, wherein a given problem statement is associated with one or more subproblem statements.

17. The method of any of the preceding claims, wherein identifying other technical problem statements that are similar to the first technical problem statements comprises performing a semantic analysis and / or semantic comparison of the problem statements, optionally based on vector embeddings of the problem statements.

18. The method of any of the preceding claims, wherein identifying the other technical problem statements comprises identifying technical problem statements which have potential causal relationships with one or more of the first technical problem statements.

19. The method of any of the preceding claims, comprising automatically assigning categories to problem descriptions and / or technical solutions, preferably based on analysis of the problem descriptions and / or technical solutions.

20. The method of claim 19, comprising using a trained classification engine to assign a set of category tags to one or more problem descriptions and / or technical solutions.

21. The method of claim 19 or 20, comprising applying a matching algorithm which compares categories or category tags assigned to problem descriptions and / or technical solutions to identify groupings of problem descriptions and / or technical solutions; and optionally comprising using one or more of: clustering, similarity measures, or rule-based approaches to group together the problem statements or technical solutions that share similar or matching category tags.

22. The method of any of the preceding claims, further comprising processing the problem statements, optionally by the first artificial intelligence data processing engine, to score each problem statement based on a set of predefined scoring criteria to assess the quality of the problem statements, wherein ranking of the problem statements is optionally performed in dependence on the scores assigned to the problem statements.

23. The method of any of the preceding claims, further comprising processing the problem statements, optionally by the first artificial intelligence data processing engine, to identify historical or in-progress projects within an organisation or from external sources that are relevant to each problem statement, and optionally further comprising providing based on the identified projects an indication of one or more of: potential benefits, involved users, risk mitigations, lessons learned, and possible solutions.

24. The method of claim 23, wherein the processing identifies and outputs information pertaining to potential connections and / or complementary aspects between the problem statements and the identified projects.

25. The method of any of the preceding claims, comprising generating an output recommending potential solutions for one or more of the problem statements using a problem-solution matching algorithm, optionally using an artificial intelligence data processing engine.

26. The method of claim 25, wherein the problem-solution matching algorithm selects recommended potential solutions based on one or more of: technical feasibility, resource constraints, and user feedback.

27. The method of any of the preceding claims, wherein the other technical solutions identified in step (f) are further determined based on one or more of: a knowledge base of known failure modes, an assessesment of potential unintended consequences of proposed solutions, analyses of user comments and feedback, Al-generated cost-benefit analyses for each potential solution.

28. The method of any of the preceding claims, comprising generating a combined problem-solution map including problem statements and technical solutions, and preferably outputting a visualisation of the problem-solution map on a user interface.

29. The method of claim 28, wherein the map includes one or more of: the first plurality of problem statements, the identified other technical problem statements identified using the first artificial intelligence data processing engine in step (b), the received technical solutions, and the identified other technical solutions identified using the second artificial intelligence data processing engine in step (f).

30. The method of claim 29, wherein the map illustrates one or more of: relationships between problem statements, for example as a tiered hierarchy of problem statements, the hierarchy optionally defining problem chains comprising problems and associated sub-problems; relationships between problem statements and associated technical solutions;solution gaps for one or more problem chains, wherein a solution gap preferably identifies one or more problem statements in a problem chain without an associated technical solution.

31. The method of any of the preceding claims, comprising generating using a machine learning model a cost benefit analysis for each problem statement, estimating the potential costs and benefits of addressing the problem.

32. A computer-implemented data processing method of facilitating collaborative development of technical solutions to technical problems in an industrial setting, the collaborative development involving a plurality of users working together to develop the technical solutions, the method comprising steps of:(a) receiving from a user, a plurality of technical problem themes related to a technical problem;(b) receiving from the user, a plurality of technical criteria against which technical problems can be measured;(c) inputting the plurality of technical problem themes into a first artificial intelligence data processing engine which processes the plurality of technical problem themes and as a result of the processing, identifies other technical problem themes which are similar to the inputted plurality of technical problem themes, where the other technical problem themes have been previously received from at least one other user;(d) inviting a plurality of users to input technical problem statements relating to each of the themes, taking into account the identified other technical problem themes from step (c);(e) receiving a first plurality of technical problem statements related to each of the plurality of themes from a plurality of users in response to the inviting step (d);(f) inputting the received first plurality of technical problem statements into a second artificial intelligence data processing engine which processes the first plurality of technical problem statements and as a result of the processing, identifies other technical problem statements, where the identified other technical problem statements are similar to the first technical problem statements, and where the other technical problem statements have been previously received from at least one other user;(g) ranking the results of the step (f) against the plurality of technical criteria;(h) receiving from a user a group of selected technical problem statements, selected from the first plurality of technical problem statements and the other technical problem statements, based on the results of the ranking step;(i) receiving from a user a plurality of technical criteria against which users can measure technical solutions to the group of selected technical problem statements;(j) inviting a plurality of users to input technical solutions to the selected technical problem statements;(k) receiving from a plurality of users, a plurality of technical solutions to the selected technical problem statements in response to the inviting step (j);(l) inputting the received plurality of technical solutions into a third artificial intelligence data processing engine which processes the plurality of technical solutions and as a result of the processing, identifies other technical solutions, where the identified other technical solutions are similar to the received plurality of technical solutions, and where the other technical solutions have been previously received from at least one other user;(m) ranking the results of step (I) against the technical criteria received at step (i); and(n) receiving from a user at least one technical solution as a selected technical solution to at least one of the selected technical problem statements, as a result of the step (m).

33. The method of claim 32, further comprising the further steps or features of any of claims 2 or 6 to 31.

34. A system comprising means, optionally in the form of one or more processors with associated memory, adapted for carrying out the steps of the method according to any preceding method claim.

35. A computer program or non-transitory computer readable medium comprising instructions for carrying out the steps of the method according to any preceding method claim, when said instructions are executed on a computer system.

Images