Executing ai consultant sessions

Abstract

A method implements executing AI consultant sessions. The method involves generating a strategy map using a language model during a session from a transcript and a mapping prompt. The method further involves selecting multiple sessions including the session. The method further involves generating an aggregated strategy map from the multiple sessions.

Description

EXECUTING Al CONSULTANT SESSIONSCROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Application 63 / 718,904, filed November 11, 2024, which is incorporated herein by reference.BACKGROUND

[0002] Computer systems capture and display information using data structures and user interfaces. For example, graphical user interfaces (GUIs) may be used to collect and then store information from user inputs into data structures. The user inputs may include natural language provided by the user that is stored in a repository or database. In response to the interaction with the user, information may be manipulated in the data structures and displayed. Challenges include structuring the data and user interfaces to identify specific items, set priorities, and focus on certain items, as well as maintaining a rhythm for performing actions to update the data within the data structures with the user interfaces.SUMMARY

[0003] In general, in one or more aspects, the disclosure relates to a method for executing Al consultant sessions. The method involves generating a strategy map using a language model during a session from a transcript and a mapping prompt. The method further involves selecting multiple sessions including the session. The method further involves generating an aggregated strategy map from the multiple sessions.

[0004] In general, in one or more aspects, the disclosure relates to a system that includes at least one processor and an application that executes on the at least one processor. Executing the application performs generating a strategy map using a language model during a session from a transcript and a mapping prompt. Executing the application further performs selecting multiple sessionsincluding the session. Executing the application further performs generating an aggregated strategy map from the multiple sessions.

[0005] In general, in one or more aspects, the disclosure relates to a non- transitory computer readable medium including instructions executable by at least one processor. Executing the instructions performs generating a strategy map using a language model during a session from a transcript and a mapping prompt. Executing the instructions further performs selecting multiple sessions including the session. Executing the instructions further performs generating an aggregated strategy map from the multiple sessions.

[0006] Other aspects of one or more embodiments may be apparent from the following description and the appended claims.BRIEF DESCRIPTION OF DRAWINGS

[0007] FIG. 1 shows a diagram in accordance with the disclosure.

[0008] FIG. 2 shows a method in accordance with the disclosure.

[0009] FIG. 3, FIG. 4, FIG. 5, FIG. 6, and FIG. 7 show examples in accordance with the disclosure.

[0010] FIG. 8.1 and 8.2 show a computing systems in accordance with the disclosure.

[0011] Similar elements in the various figures may be denoted by similar names and reference numerals. The details of features and elements described in one figure may extend to similarly named features and elements in different figures.DETAILED DESCRIPTION

[0012] Embodiments of the disclosure may implement an artificial intelligence- powered system for strategy mapping and organizational planning processes. The system processes natural language input, including voice and text, to generate structured strategy map nodes. Each node represents an element in a data structure that records an organizational strategy with objectives, projects,or performance indicators. The system further processes multiple facilitated sessions, each producing interim strategy maps and transcripts, and aggregates the outputs into a unified, hierarchical strategy map. The system incorporates a multi-model architecture, including transcription and reasoning models, to iteratively update and refine the strategy map based on ongoing conversations and user interactions.

[0013] Conventional approaches to strategic planning and operational alignment suffer from fragmentation, inefficiency, and reliance on manual facilitation. Organizations may use extensive time and resources to conduct interviews, synthesize insights, and construct coherent strategy maps. Existing tools lack the capacity to dynamically structure unstructured input across multiple sessions or to maintain consistency across evolving strategic frameworks. Manual consolidation of disparate inputs introduces delays, inconsistencies, and cognitive overload for facilitators, particularly in large or distributed organizations.

[0014] The disclosed system addresses the above challenges by automating the transformation of unstructured conversational data into structured strategy maps. The system processes audio input at regular intervals to generate accurate, context-aware transcripts. The system analyzes the transcripts and existing map structures to determine whether updates are warranted. The system aggregates multiple sessions by serializing interim maps and transcripts into a unified prompt, which is processed to generate an aggregated strategy map. A user interface enables selection of sessions for aggregation and includes mechanisms to preserve facilitator-defined elements. The system further incorporates a challenge-response mechanism that introduces alternative perspectives during sessions to enhance strategic rigor.

[0015] Turning to FIG. 1, the system (100) is a computing system shown in accordance with one or more embodiments. The system (100) and corresponding components may utilize the computing systems described in FIG. 8.1 and FIG. 8.2 to implement a management application that operates asan artificial intelligence (Al) powered chief of staff bot. The system (100) includes the repository (102), the server (152), and the user device (190).

[0016] The repository (102) is a collection of devices that form a unit of data storage that stores the data used by the system (100). The repository (102) may include multiple different, potentially heterogenous, storage units and / or devices. The repository (102) stores data utilized by other components of the system (100). The data stored by the repository (102) includes the strategy maps (105).

[0017] The strategy maps (105) are data structures stored on the repository (102). The strategy maps (105) include the objects (108), which may be referred to as strategy map nodes.

[0018] The objects (108) are a set of computer programming objects that store data extracted from user inputs in object records in the repository (102). The data extracted from the user inputs may include textual descriptions stored as objective records within the objects (108). Each of the objects (108) may include text (z.e., an objective record) that describes the object and may include references to other ones of the objects (108). The objects (108) include the user objects (110), relationship objects (112), the objective objects (115), the topic objects (122), and the settlement objects (132). The objects (108) may each be stored at a location within data storage, i.e., within the repository (102).

[0019] Each object of the objects (108) may include references to other objects of the objects (108) that identify the locations of the other objects within the data storage. The objects (108) may be structured with multiple hierarchies. One type of hierarchy is the hierarchy between different types of objects. For example, each of the objective objects (115) may be upstream from (or a parent of) multiple topic objects (122). Each of the topic objects (122) may be upstream from (or a parent of) multiple settlement objects (132).

[0020] Each object of the objects (108) may include properties that represent performance metrics. A performance metric may be a performance indicator ora milestone. A performance indicator may be a numerical value that may be compared against a threshold to determine whether an object represents a strategy node identified as an off-track node. A milestone may include a textual description and a date. When the current date is past the date of the milestone, the strategy node represented by the corresponding object may be identified as an off-track node.

[0021] The user objects (110) identify the users of the system (100). Information stored in the object record of a user object may include a name for the user, contact information (including an email address), etc. Each of the user objects (110) may be associated with one or more of the objective objects (115), the topic objects (122), and the settlement objects (132). Association of the user objects (110) to one or more of the objects (108) (including to one or more of the objective objects (115), the topic objects (122), and the settlement objects (132)) may provide an identification of the users that are responsible for the corresponding objects (108).

[0022] The objective objects (115) store information describing objectives that are associated with the users of the system (100). The information describing the objectives may be stored as text after being extracted from user inputs received by the system (100). The information is stored in objective records of the objective objects (115).

[0023] The relationship objects (112) identify the relationships between the other objects or with tags and labels. For example, one of the relationship objects (212) may identify the hierarchical relationship between one of the pillar objects (216) and one of the project objects (217). Multiple projects represented by the project objects (217) may be related to a tag or label. The tag or label may represent a product that each of the projects affect or are related to. The relationship objects (112) may be stored as edges in a graph database with the other objects (the user objects (210), the objective objects (115), the topic objects (122), and the settlement objects (132), etc.) stored as nodes in the graph database.

[0024] The objective objects (115) may further maintain a hierarchy between the objective objects (115), such as the hierarchy between the pillar objects (116) and the project objects (117). For example, one of the objective objects (115) (e.g., one of the pillar objects (116)) may be a parent that is upstream from other objective objects (115) (e.g., one of the project objects (117)). The hierarchy between the objective objects (115) may correspond to a hierarchy between users of the system. For example, a user Jack may be associated with a first objective object that is the parent to a second objective object that is associated with the user Janice, who is a subordinate of Jack. Additionally, each of the objective objects (115) may be associated with one or more of the topic objects (122).

[0025] The pillar objects (116) are one type of the objective objects (115) that may be superordinate to the project objects (117). The text stored in the object record of one of the pillar objects (116) identifies a pillar (i.e., an objective) that is superordinate to one or more projects corresponding to the project objects (117).

[0026] The project objects (117) are one type of the objective objects (115) that may be subordinate to the pillar objects (116). The text stored in the object record of one of the project objects (117) identifies a project (e.g., a task or set of tasks) that may be part of one of the pillar objects (116).

[0027] The topic objects (122) store information describing topics that relate to the objectives stored with the objective objects (115). One of the topic objects (122) may be associated with one of the objective objects (115) and one of the objective objects (115) may be associated with multiple topic objects (122). The information describing the topics of the topic objects (122) may be stored as text after the text is extracted from user inputs received by the system (100). The topic objects (122) may include several types, including the guidance objects (123), the blocker objects (124), and the update objects (125).

[0028] The guidance objects (123) are one type of the topic objects (122). In an embodiment, the text stored in the object record of one of the guidance objects(123) may include requests for guidance with regards to the objectives stored in the object records for the objective objects (115) to which the guidance objects (123) are associated.

[0029] The blocker objects (124) are one type of the topic objects (122). In an embodiment, the text stored in the object record of one of the blocker objects(124) may include a description of issues preventing progress with the objectives stored in the object records for the objective objects (115) to which the blocker objects (124) are associated.

[0030] The update objects (125) are one type of the topic objects (122). In an embodiment, the text stored in the object record of one of the update objects(125) may include descriptions of updates related to the objectives stored in the object records for the objective objects (115) to which the update objects (125) are associated.

[0031] The settlement objects (132) store information related to the topics of the topic objects (122). In an embodiment, one of the settlement objects (132) may be associated with one of the topic objects (122). Information stored by the settlement objects (132) may be stored as text after the text is extracted from user inputs received by the system (100). The settlement objects (132) may include several types, including the comment objects (133), the decision objects (134), and the action objects (135).

[0032] The comment objects (133) are one type of the settlement objects (132). In an embodiment, the text stored in the object record of one of the comment objects (133) provides additional information about one of the topics of the topic objects (122).

[0033] The decision objects (134) are one type of the settlement objects (132). In an embodiment, the text stored in the object record of one of the decisionobjects (134) is a record of a decision made with regard to one of the topics of the topic objects (122).

[0034] The action objects (135) are one type of the settlement objects (132). In an embodiment, the text stored in the object record of one of the action objects (135) is a record of an action related to one of the topics of the topic objects (122).

[0035] The sessions (142) are discrete conversational interactions between a user and the system that are initiated to collect strategic or operational planning input. Each session includes a sequence of user inputs, which may be in the form of audio or text, and is associated with a specific context, such as a department, project, or strategic objective. The sessions (142) are used to structure the collection of information from multiple stakeholders for the system to isolate and organize inputs for subsequent processing and aggregation.

[0036] The transcripts (145) are structured textual representations of the user inputs captured during the sessions (142). The transcripts (145) quantify the content of the conversations by converting audio or textual input into a consistent, machine-readable format. The transcripts (145) are used by downstream models to extract semantic meaning, identify strategy map nodes, and generate updates to the strategy map based on the content of the sessions (142).

[0037] The prompts (148) are structured textual inputs that direct the behavior of the language models (182) during the processing of the transcripts (145) and other session data. Each of the prompts (148) includes instructions, examples, and context that guide the language models (182) to perform specific tasks, such as identifying objectives, mapping relationships, or generating strategy map nodes. The prompts (148) are used in context to orchestrate the transformation of unstructured session data into structured session data for the system to reason about the content and update the strategy map accordingly.

[0038] Continuing with FIG. 1, the system (100) also may include the server (152). The server (152) is one or more computing systems, which may be part of a distributed computing environment. An example of the server (152) may be the computing system (900) shown in FIG. 8.1.

[0039] The server (152) may host and / or execute one or more processes, programs, applications, etc. For example, the server (152) may execute one or multiple instantiations of the management application (155) using different computing systems and servers. The server (152) may interact with the user device (190) to process information. The server (152) executes the management application (155).

[0040] The management application (155) is a collection of software programs executed by the server (152) to process user input, manage data structures, and generate strategy maps. The management application (155) receives natural language input, applies a series of models to extract structured data, and updates the repository (102) with new or modified strategy map nodes. The management application (155) operates as the central processing component of the system (100), coordinating the flow of data between the user device (190), the repository (102), and the various models used for conversion, reasoning, and execution.

[0041] The transcription prompt (158) is a structured input to a language model that processes audio data (195) to generate a transcript. The transcription prompt (158) includes instructions and contextual information, such as predefined acronyms, to improve the accuracy of the transcription. The transcription prompt (158) is used to convert audio input into text at regular intervals, forming the basis for subsequent reasoning and strategy map generation.

[0042] The session (160) is one of the sessions (142) and is a discrete unit of interaction between a user and the system (100), during which natural language input is collected and processed. The session (160) includes audio data (195),converted text, and intermediate outputs generated by the models executed by the management application (155). The session (160) is used to structure the collection and processing of user input for the system (100) to generate and update strategy maps in a modular and traceable manner.

[0043] The transcript ( 162) is one of the transcripts (145) and is the textual output generated from the audio data (195) using the transcription prompt (158). The transcript (162) quantifies the spoken content of the session (160) by representing the spoken content in a structured, machine-readable format. The transcript (162) is used as an input to the mapping prompt (168) to identify strategy map nodes, determine updates to the strategy map, and generate suggestions in the form of the session updates (170).

[0044] The strategy map (165) is a hierarchical data structure that represents organizational objectives, projects, topics, and settlements. The strategy map (165) includes nodes derived from user input and structured according to relationships identified by the management application (155). The strategy map (165) may also be used as an input to the mapping prompt (168) to identify strategy map nodes, determine updates to the strategy map, and generate suggestions in the form of the session updates (170). The strategy map (165) may be serialized to a structured format such as JavaScript Object Notation (JSON) to support processing by the language models (182) and for storage, transmission, and aggregation across multiple sessions.

[0045] The mapping prompt (168) is a structured input to a language model that processes the transcript (162) and the strategy map (165) to generate session updates (170). The mapping prompt (168) includes instructions and examples that direct the language model to identify new strategy map nodes, update existing nodes, or suggest changes to the structure of the strategy map. The mapping prompt (168) is used to reason over the content of the session (160) and produce actionable outputs for updating the strategy map (165).

[0046] The session updates (170) are structured outputs generated by the language model in response to the mapping prompt (168). The session updates (170) may include changes to the strategy map (165), such as the addition, removal, or modification of nodes, and may include suggestions from the language model to the users to further develop the strategy maps. The suggestions may be in the style of a devil’s advocate with probing questions or suggestions for users to consider during a session. The session updates (170) may be presented to users of the system (100) and be used to revise the strategy map (165) in real time or asynchronously, reflecting the evolution of organizational priorities and actions recorded in the strategy maps (105).

[0047] The aggregation prompt (172) is a structured input to a language model that consolidates multiple sessions (142) (including the session (160)), where each session may include a transcript (e.g., the transcript (162)) and a strategy map (e.g., the strategy map (165)), into a unified representation. The aggregation prompt (172) includes serialized data from the transcripts (162) and strategy maps (165) of the selected sessions (160), along with instructions for synthesizing a coherent and hierarchical strategy map. The aggregation prompt (172) is used to generate the aggregated strategy map (175), which reflects the combined strategic insights from multiple sessions for multiple stakeholders and users.

[0048] The aggregated strategy map (175) is a data structure that integrates the outputs of multiple sessions (142) into a single, unified strategy map. The aggregated strategy map (175) includes strategy map nodes derived from the strategy maps (105) (including from the strategy map (165)) and the session updates (170) and structured according to relationships inferred by the language model using the aggregation prompt (172). The aggregated strategy map (175) is used to record and represent a consolidated view of organizational strategy for cross-functional alignment and decision-making as a data structure stored on the system (100).

[0049] The language models (182) are machine learning models executed by the management application (155) to process natural language input and generate structured outputs. Different models may be used for processing different prompts, which may be based on the type of input and the complexity of the task. A model that processes the transcription prompt (158) may be a multimodal language model capable of processing both audio and text data to generate the transcript (162). The model that processes the mapping prompt (168) may be the same as the model used to process the aggregation prompt (172) and may be a large language model (LLM) trained to reason over structured and unstructured data. The language models (182) are used throughout the system (100) to transform unstructured user input into structured data for use in the strategy map (165) and aggregated strategy map (175).

[0050] The user device (190) is a computing system that interacts with the server (152) to transmit user input and receive and present system output. The user device (190) includes hardware and software components for capturing audio data (195), executing the user application (192), and displaying the strategy map (165). The user device (190) is used to initiate sessions (160), transmit requests to the server (152), and present the outputs of the management application (155) to a user.

[0051] The user application (192) is a software program executed on the user device (190) that manages user interaction with the system (100). The user application (192) collects user input, transmits requests to the server (152), and displays responses, including the strategy map (165) and aggregated strategy map (175) as well as suggestions from the session updates (170). The user application (192) is used to control the flow of interaction between the user and the system (100) for the user to participate in sessions (160) and review session updates (170).

[0052] The audio data (195) is the raw sound input captured during a session (160) using the user device (190). The audio data (195) quantifies the spoken content of the user and is used as input to the transcription prompt (158) forconversion into the transcript (162). The audio data (195) is used to initiate the processing pipeline of the management application (155), forming the basis for generating structured strategy map nodes and updates.

[0053] FIG. 2 shows a flowchart of a method for executing Al consultant sessions. The method of FIG. 2 may be implemented using the systems described in the other figures, and one or more of the steps may be performed on, or received at, one or more computer processors. The system may include at least one processor and an application that, when executing on the at least one processor, performs the method. A non-transitory computer readable medium may include instructions that, when executed by one or more processors, perform the method. The outputs from various components (including models, functions, procedures, programs, processors, etc.) for performing the method may be generated by applying a transformation to inputs using the components to create the outputs without using mental processes or human activities.

[0054] Turning to FIG. 2, the method (200) generates an aggregated strategy map from multiple sessions. The process (200) may include multiple steps (e.g., Block 202 through Block 210) that may execute on the components described in the other figures, including those of FIG. 1, FIG. 8.1, and FIG. 8.2.

[0055] Block 202 involves generating a strategy map using a language model during a session from the transcript and a mapping prompt. The transcript is derived from natural language input, such as audio or text, and the mapping prompt includes structured instructions for identifying objectives, topics, and actions. The language model processes the transcript and the mapping prompt to extract semantic meaning and generate structured nodes representing elements of the strategy map, which are then assembled into a hierarchical format.

[0056] Block 205 involves selecting multiple sessions, including the session.Each session includes a transcript and a corresponding strategy map generatedduring a prior interaction. The selection may be performed through a user interface that displays a list of available sessions, allowing a user to choose multiple sessions for consolidation based on relevance, time frame, or participant.

[0057] Block 208 involves generating an aggregated strategy map from the multiple sessions. The selected sessions are serialized into a structured format, such as JSON, and processed by a language model using an aggregation prompt. The model synthesizes the content of the transcripts and strategy maps from the selected sessions to produce a unified, hierarchical strategy map that reflects the combined strategic insights.

[0058] The method (200) may involve executing a transcription prompt with audio data to generate the transcript. The audio data is captured during a session and passed to a multimodal language model along with the transcription prompt. The model processes the audio input to produce a text-based transcript that accurately reflects the spoken content of the session.

[0059] The method (200) may involve executing a transcription prompt with a list of acronyms to generate the transcript. The transcription prompt includes a predefined list of acronyms relevant to the organization or domain. The language model uses the list to disambiguate and transcribe specialized terms or abbreviations present in the audio data.

[0060] The method (200) may involve executing a transcription prompt at a transcription interval selected from a range of 15 to 60 seconds. The system segments the audio data into time-based intervals and triggers the transcription prompt at each interval. The periodic execution updates the transcript incrementally and remains synchronized with the ongoing session.

[0061] The method (200) may involve executing the mapping prompt to generate a session update based on a current strategy map and the transcripts. The mapping prompt is constructed using the current state of the strategy map (z.e., the current strategy map) and the most recent transcript segment. The languagemodel processes the prompt to identify new or modified strategy map nodes and outputs a session update containing proposed changes.

[0062] The method (200) may involve revising the strategy map using the session update. The session update includes structured data representing additions, deletions, or modifications to the strategy map. The system applies the update by modifying the underlying data structures and refreshing the visual representation of the strategy map.

[0063] The method (200) may involve presenting a suggestion using the session update. The suggestion may be displayed in the user interface as a proposed change, such as a new objective or action item. The user may review, accept, or reject the suggestion, thereby maintaining control over the evolving strategy map.

[0064] The method (200) may involve executing a mapping prompt at a suggestion interval that is greater than a transcription interval for executing a transcription prompt. The system maintains separate timers for transcription and mapping prompts. The mapping prompt is executed less frequently to allow sufficient transcript content to accumulate before reasoning about updates to the strategy map.

[0065] The method (200) may involve executing a mapping prompt at a suggestion interval selected from a range of 1 to 5 minutes. The suggestion interval is configurable and determines how often the system evaluates the transcript and current strategy map to generate new session updates. The interval balances responsiveness with computational efficiency.

[0066] The method (200) may involve selecting a checkbox for a node of the strategy map to prevent updates to the node responsive to execution of the mapping prompt. The user interface includes a control element, such as a checkbox, associated with each strategy map node. When selected, the system marks the node as immutable, and the language model is instructed to preserve the node during subsequent updates.

[0067] The method (200) may involve generating the strategy map with multiple objective objects, topic objects, and settlement objects. The language model identifies and classifies extracted content into distinct object types based on semantic roles. Each object is instantiated with properties and relationships, forming a structured and navigable strategy map.

[0068] The method (200) may involve presenting the aggregated strategy map. The aggregated strategy map is rendered in the user interface using a hierarchical layout. The user may explore the map, inspect individual nodes, and trace relationships across sessions to view records of the consolidated strategic direction.

[0069] Turning to FIG. 3, the figure illustrates an example of the computer- implemented user interface (300) that processes conversational input to generate and revise a strategy map. The user interface (300) includes the map view (302) and the chat view (352), which operate concurrently to structure and analyze strategic planning discussions. The map view (302) displays the strategy map (305) composed of the objective object (312), the topic objects (315, 318, 320, 322), and the action objects (325, 328, 330, 332, 335, 338, 340). The chat view (352) includes suggestion elements (355, 358, 360) and the system response (362) that are generated by the system to guide the user through the planning process.

[0070] The strategy map (305) is generated and updated by the system using structured data extracted from user input. The objective object (312) includes a textual description such as “Determine which Al-powered features to patent to maximize value and defensibility.” The system identifies topic objects (315, 318, 320, 322) as subordinate nodes to the objective object (312) based on semantic relationships in the transcript. For example, the topic object (315) includes “Identify Core Al Features,” the topic object (318) includes “Assess Patentability Core Features,” the topic object (320) includes “Determine patent breadth and enforcement costs,” and the topic object (322) includes “Prioritize Al features for patent protection.”

[0071] The system further identifies action objects as subordinate nodes to each topic object. The topic object (315) is linked to the action objects (325, 328), which include “List all Al-powered features of the product” and “Analyze each feature’s contribution to the product’s value proposition.” The topic object (318) is linked to the action objects (330, 332), which include “Conduct a prior art search for each core Al feature” and “Consult with a patent attorney to evaluate novelty, nonobviousness, and technical feasibility.” The topic object (320) is linked to the action objects (335, 338), which include “Estimate the scope of protection each patent would offer” and “Assess the potential costs associated with enforcing each patent in relevant jurisdictions.” The topic object (322) is linked to the action object (340), which includes “Develop a scoring system to rank Al features based on their value proposition, patentability, and enforcement cost.”

[0072] The system may apply color coding to visually distinguish topic-object groupings. The topic object (315) and the action objects (325, 328) are color coded brown to indicate their grouping around identifying core Al features. The topic object (318) and the action objects (330, 332) are color coded blue to indicate their grouping around assessing patentability. The topic object (320) and the action objects (335, 338) are color coded green to indicate their grouping around evaluating patent breadth and enforcement costs. The topic object (322) and the action object (340) are color coded yellow to indicate their grouping around prioritization for patent protection.

[0073] The chat view (352) is generated by the system to present interactive suggestions and feedback, which may be in the style of a devil’s advocate. The suggestion elements (355, 358, 360) include questions such as “QI: What Al features are core to our product and contribute most significantly to its value proposition?”, “Q2: Which of these features are novel, not obviousness, and technically feasible to patent?”, and “Q3: Considering potential patent breadth and enforcement costs, which patents would offer the strongest market protection?” Each suggestion element includes a refresh button that may beselected to replace the current question with a new one. The system response (362) includes a message such as “Up here, I have identified some questions you should explore. You can refresh one of them if it isn’t quite what you’re looking for.”

[0074] The system includes a listening element in the chat view (352) to indicate that audio input is being captured. The system executes a transcription prompt using a multimodal model at regular intervals, such as at 30 second intervals, to convert audio input into transcripts. The transcripts are processed using a mapping prompt executed with a language model at longer intervals, such as at 3 minute intervals, to update the strategy map (305) and generate suggestions such as the questions presented in the suggestion elements (355, 358, 360). The session may be a blue sky session in which a strategy map is generated from open-ended discussion, or a collaborative session in which multiple users discuss and revise the strategy map.

[0075] Turning to FIG. 4, the figure illustrates an example of a computer- implemented user interface (400) that processes conversational input to generate and revise a strategy map based on transcribed speech and structured reasoning. The user interface (400) includes a map view (402) and a chat view (452), which operate concurrently to structure strategic planning content and guide user interaction. The map view (402) displays a strategy map (405) composed of objective objects (408, 410) and topic objects (412, 415, 418, 420, 422). The chat view (452) includes a transcribed user prompt (455) and a system response (458), which are generated and processed by the system to update the strategy map (405) and prompt further discussion.

[0076] The strategy map (405) is generated by the system using structured data extracted from user speech and processed through transcription and mapping prompts. The objective objects (408, 410) and the topic objects (412, 415, 418, 420, 422) are generated in response to descriptions received from the user through the conversational interface. The objective object (408) includes the text “Leverage Al to enhance decision plus driving user engagement anddecision making” and is associated with the personal identifier “Alex B.” The objective object (410) includes the text “Expansion strategy for the US” and is also associated with the personal identifier “Alex B.” The system identifies topic objects (412, 415, 418) as subordinate nodes to the objective object (408), and topic objects (420, 422) as subordinate nodes to the objective object (410), based on semantic relationships in the transcript. The topic object (412) is a recommendation-type topic associated with the personal identifier “Mary” and includes the text “Consider how you plan to differentiate your efforts. For example, identify unique Al capabilities that can provide a competitive edge in user engagement.” The topic object (415) is a directive-type topic associated with the personal identifier “Mary” and includes the text “File patent application for existing Al features.” The topic object (418) is a planning-type topic associated with the personal identifier “Joe” and includes the text “Develop a roadmap for future Al features.” The topic object (420) is a recommendation-type topic associated with the personal identifier “Mary” and includes the text “Consider what additional capabilities you might need. For example, develop partnerships with local businesses to enhance market penetration and support.” The topic object (422) is a planning-type topic associated with the personal identifier “Diana” and includes the text “Create go-to-market approach for the US focused on pilots on the East Coast.”

[0077] The chat view (452) is generated by the system to capture and process conversational input. The transcribed user prompt (455) includes the text “Yeah, so Mary is working on the patent application for our Al features, our existing Al features. Joe is working on a road map for our future Al features. Diana is working on creating the go-to-market approach for the US focused on pilots on the East Coast.” The system processes the transcribed user prompt (455) using a mapping prompt to identify named individuals and their associated priorities. The system response (458) includes the text “I added Mary, Joe, and Diana and their priorities. Are there any other people and their priorities that you’d like to add or should we proceed?” The system response(458) is generated based on the updated strategy map (405) and is presented to prompt further input.

[0078] The system executes a transcription prompt using a multimodal model at regular intervals, such as every 30 seconds, to convert audio input into transcripts. The transcripts are processed using a mapping prompt executed with a language model at longer intervals, such as every 3 minutes, to update the strategy map (405) and generate system responses such as the prompt in the system response (458). The session may be a blue sky session in which a strategy map is generated from open-ended discussion, or a collaborative session in which multiple users discuss and revise the strategy map.

[0079] Turning to FIG. 5, the figure illustrates an example of a computer- implemented user interface (500) that processes conversational input and session data to generate, revise, and manage strategy maps. The user interface (500) includes a map view (502) and a header view (552), which operate in parallel to structure strategic planning content and control session-level actions. The map view (502) displays a strategy map (505) composed of objective objects (508, 512) and topic objects (510, 515), which are generated by the system in response to descriptions received from the user through the conversational interface. The header view (552) includes interface elements (555, 558, 560) that trigger system actions for reviewing, importing, and duplicating strategy maps.

[0080] The strategy map (505) is generated by the system using structured data extracted from user speech and processed through transcription and mapping prompts. The objective object (508) includes the text “Implement Al in the company” and is linked to the topic object (510), which includes the text “Build products using Gemini.” The objective object (512) includes the text “Optimize the costs for this company” and is linked to the topic object (515), which includes the text “Reduce infrastructure costs.” The system identifies relationships between objective objects and topic objects based on semanticstructure in the transcript and maps the objects into a hierarchical format for display in the map view (502).

[0081] The header view (552) includes interface elements that initiate sessionlevel operations. The interface element (555) is a button labeled “Review map session.” When selected, the system displays an interface that includes the transcripts generated during the session, the corresponding strategy map (505), and a timeline slider. The timeline slider may be used to view the evolution of the strategy map over time in conjunction with the transcript history.

[0082] The interface element (558) is a button labeled “Import from brainstorming.” When selected, the system presents an interface to select strategy maps or transcripts from other sessions. The selected content is imported into the current session and processed to update the strategy map (505) using the same transcription and mapping prompts.

[0083] The interface element (560) is a button labeled “Duplicate map.” When selected, the system duplicates the current strategy map (505) and stores the duplicate as a new map instance. The duplicated map may be used as the basis for discussion with multiple different users, each contributing to separate sessions. The system tracks each session independently and may later aggregate the duplicated maps using an aggregation prompt to generate a consolidated strategy map.

[0084] Turning to FIG. 6, the figure illustrates an example of a computer- implemented user interface (600) that processes session-level data and user selections to manage and aggregate strategic planning sessions. The user interface (600) includes a session list (602) and interface elements (605, 608), which are used to select, transfer, and combine session content into strategy maps. The system displays the session list (602) as a scrollable or searchable list of session names. Each session name corresponds to a previously recorded session that includes transcripts, strategy maps, and associated metadata. Thesystem receives user selection input from the session list (602) to retrieve and display the contents of the selected session for review or further processing.

[0085] The interface element (605) is a button labeled “Save to Strategy”. When selected, the system presents an interface that includes the current strategy map and a destination strategy map. The system receives user input identifying specific elements from the current strategy map, such as objective objects, topic objects, or action objects, and transfers the selected elements to the destination strategy map. The system may update the destination strategy map by inserting the transferred elements and adjusting the hierarchical relationships.

[0086] The interface element (608) is a button labeled “Combine Sessions”. When selected, the system presents an interface to select multiple sessions from the session list (602). The system serializes the transcripts and strategy maps from the selected sessions into a structured format, such as JSON. The system processes the serialized data using an aggregation prompt executed with a language model to generate an aggregated strategy map. The aggregated strategy map reflects the combined strategic insights from the selected sessions and is displayed for further review or revision.

[0087] Turning to FIG. 7, the figure illustrates an example of a computer- implemented user interface (700) that processes session selection input to generate an aggregated strategy map. The user interface (700) includes a selection view (702) and an interface element (705), which operate together to receive user selections and initiate session aggregation. The selection view (702) presents a list of sessions, each displayed with a corresponding checkbox. The system receives user input through the selection view (702) to identify multiple sessions for aggregation. Each session may include a transcript and a strategy map generated during a prior interaction.

[0088] The interface element (705) is a button labeled “Combine Selected.” When selected, the system retrieves the transcripts and strategy maps corresponding to the sessions selected in the selection view (702). The system 1serializes the transcripts and strategy maps into a structured format, such as JSON, and incorporates the serialized data into an aggregation prompt. The aggregation prompt includes instructions for synthesizing the information from each of the selected sessions, including the strategy maps and the transcripts generated with the selected sessions. The system executes the aggregation prompt using a language model to generate an aggregated strategy map. The aggregated strategy map reflects the combined strategic insights from the selected sessions and is displayed for further review, revision, or export.

[0089] One or more embodiments may be implemented on a computing system specifically designed to achieve an improved technological result. When implemented in a computing system, the features and elements of the disclosure provide a significant technological advancement over computing systems that do not implement the features and elements of the disclosure. Any combination of mobile, desktop, server, router, switch, embedded device, or other types of hardware may be improved by including the features and elements described in the disclosure.

[0090] For example, as shown in FIG. 8.1, the computing system (800) may include one or more computer processor(s) (802), non-persistent storage device(s) (804), persistent storage device(s) (806), a communication interface (808) (e.g., Bluetooth interface, infrared interface, network interface, optical interface, etc.), and numerous other elements and functionalities that implement the features and elements of the disclosure. The computer processor(s) (802) may be an integrated circuit for processing instructions. The computer processor(s) (802) may be one or more cores, or micro-cores, of a processor. The computer processor(s) (802) includes one or more processors. The computer processor(s) (802) may include a central processing unit (CPU), a graphics processing unit (GPU), a tensor processing unit (TPU), combinations thereof, etc.

[0091] The input device(s) (810) may include a touchscreen, keyboard, mouse, microphone, touchpad, electronic pen, or any other type of input device. Theinput device(s) (810) may receive inputs from a user that are responsive to data and messages presented by the output device(s) (812). The inputs may include text input, audio input, video input, etc., which may be processed and transmitted by the computing system (800) in accordance with one or more embodiments. The communication interface (808) may include an integrated circuit for connecting the computing system (800) to a network (not shown) (e.g., a local area network (LAN), a wide area network (WAN), such as the Internet, mobile network, or any other type of network) or to another device, such as another computing device, and combinations thereof.

[0092] Further, the output device(s) (812) may include a display device, a printer, external storage, or any other output device. One or more of the output device(s) (812) may be the same or different from the input device(s) (810). The input device(s) (810) and output device(s) (812) may be locally or remotely connected to the computer processor(s) (802). Many different types of computing systems exist, and the aforementioned input device(s) (810) and output device(s) (812) may take other forms. The output device(s) (812) may display data and messages that are transmitted and received by the computing system (800). The data and messages may include text, audio, video, etc., and include the data and messages described above in the other figures of the disclosure.

[0093] Software instructions in the form of computer readable program code to perform embodiments may be stored, in whole or in part, temporarily or permanently, on a non-transitory computer readable medium, such as a solid state drive (SSD), compact disk (CD), digital video disk (DVD), storage device, a diskette, a tape, flash memory, physical memory, or any other computer readable storage medium. Specifically, the software instructions may correspond to computer readable program code that, when executed by the computer processor(s) (802), is configured to perform one or more embodiments, which may include transmitting, receiving, presenting, and displaying data and messages described in the other figures of the disclosure.

[0094] The computing system (800) in FIG. 8.1 may be connected to, or be a part of, a network. For example, as shown in FIG. 8.2, the network (820) may include multiple nodes (e.g., node X (822) and node Y (824), as well as extant intervening nodes between node X (822) and node Y (824)). Each node may correspond to a computing system, such as the computing system shown in FIG. 8.1, or a group of nodes combined may correspond to the computing system shown in FIG. 8.1. By way of an example, embodiments may be implemented on a node of a distributed system that is connected to other nodes. By way of another example, embodiments may be implemented on a distributed computing system having multiple nodes, where each portion may be located on a different node within the distributed computing system. Further, one or more elements of the aforementioned computing system (800) may be located at a remote location and connected to the other elements over a network.

[0095] The nodes (e.g., node X (822) and node Y (824)) in the network (820) may be configured to provide services for a client device (826). The services may include receiving requests and transmitting responses to the client device (826). For example, the nodes may be part of a cloud computing system. The client device (826) may be a computing system, such as the computing system shown in FIG. 8.1. Further, the client device (826) may include or perform all or a portion of one or more embodiments.

[0096] The computing system of FIG. 8.1 may include functionality to present data (including raw data, processed data, and combinations thereof), such as results of comparisons and other processing. For example, presenting data may be accomplished through various presenting methods. Specifically, data may be presented by being displayed in a user interface, transmitted to a different computing system, and stored. The user interface may include a graphical user interface (GUI) that displays information on a display device. The GUI may include various GUI widgets that organize what data is shown, as well as how data is presented to a user. Furthermore, the GUI may present data directly to the user, e.g., data presented as actual data values through text, or rendered bythe computing device into a visual representation of the data, such as through visualizing a data model.

[0097] As used herein, the term “connected to” contemplates multiple meanings. A connection may be direct or indirect (e.g., through another component or network). A connection may be wired or wireless. A connection may be a temporary, permanent, or a semi-permanent communication channel between two entities.

[0098] The various descriptions of the figures may be combined and may include, or be included within, the features described in the other figures of the application. The various elements, systems, components, and steps shown in the figures may be omitted, repeated, combined, or altered as shown in the figures. Accordingly, the scope of the present disclosure should not be considered limited to the specific arrangements shown in the figures.

[0099] In the application, ordinal numbers (e.g., first, second, third, etc.) may be used as an adjective for an element (i.e., any noun in the application). The use of ordinal numbers is not to imply or create any particular ordering of the elements, nor to limit any element to being only a single element unless expressly disclosed, such as by the use of the terms “before”, “after”, “single”, and other such terminology. Rather, ordinal numbers distinguish between the elements. By way of an example, a first element is distinct from a second element, and the first element may encompass more than one element and succeed (or precede) the second element in an ordering of elements.

[0100] Further, unless expressly stated otherwise, the conjunction “or” is an inclusive “or” and, as such, automatically includes the conjunction “and”, unless expressly stated otherwise. Further, items joined by the conjunction “or” may include any combination of the items with any number of each item, unless expressly stated otherwise.

[0101] In the above description, numerous specific details are set forth in order to provide a more thorough understanding of the disclosure. However, it willbe apparent to one of ordinary skill in the art that the technology may be practiced without these specific details. In other instances, well-known features have not been described in detail to avoid unnecessarily complicating the description. Further, other embodiments not explicitly described above can be devised which do not depart from the scope of the claims as disclosed herein. Accordingly, the scope should be limited only by the attached claims.

Claims

CLAIMSWhat is claimed is:

1. A method comprising: generating a strategy map using a language model during a session from transcript and a mapping prompt; selecting a plurality of sessions comprising the session; and generating an aggregated strategy map from the plurality of sessions.

2. The method of claim 1, further comprising: executing a transcription prompt with audio data to generate the transcript.

3. The method of claim 1, further comprising: executing a transcription prompt with a list of acronyms to generate the transcript.

4. The method of claim 1, further comprising: executing a transcription prompt at a transcription interval selected from a range of 15 to 60 seconds.

5. The method of claim 1, further comprising: executing the mapping prompt to generate a session update based on a current strategy map and the transcript; and revising the strategy map using the session update; and presenting a suggestion using the session update.

6. The method of claim 1, further comprising: executing a mapping prompt at a suggestion interval that is greater than a transcription interval for executing a transcription prompt.

7. The method of claim 1, further comprising: executing a mapping prompt at a suggestion interval selected from a range of 1 to 5 minutes.

8. The method of claim 1, further comprising: selecting a checkbox for a node of the strategy map to prevent updates to the node responsive to execution of the mapping prompt.

9. The method of claim 1, further comprising: presenting the aggregated strategy map.

10. The method of claim 1, further comprising: generating the strategy map with a plurality of objective objects, topic objects, and settlement objects.

11. A system comprising: a computer processor; and an application that, when executing on the computer processor, performs operations comprising: generating a strategy map using a language model during a session from transcript and a mapping prompt, selecting a plurality of sessions comprising the session, and generating an aggregated strategy map from the plurality of sessions.

12. The system of claim 11, wherein the application performs operations further comprising: executing a transcription prompt with audio data to generate the transcript.

13. The system of claim 11, wherein the application performs operations further comprising: executing a transcription prompt with a list of acronyms to generate the transcript.

14. The system of claim 11, wherein the application performs operations further comprising: executing a transcription prompt at a transcription interval selected from a range of 15 to 60 seconds.

15. The system of claim 11, wherein the application performs operations further comprising: executing the mapping prompt to generate a session update based on a current strategy map and the transcript; revising the strategy map using the session update; and presenting a suggestion using the session update.

16. The system of claim 11, wherein the application performs operations further comprising: executing a mapping prompt at a suggestion interval that is greater than a transcription interval for executing a transcription prompt.

17. The system of claim 11, wherein the application performs operations further comprising: executing a mapping prompt at a suggestion interval selected from a range of 1 to 5 minutes.

18. The system of claim 11, wherein the application performs operations further comprising: selecting a checkbox for a node of the strategy map to prevent updates to the node responsive to execution of the mapping prompt.

19. The system of claim 11, wherein the application performs operations further comprising: presenting the aggregated strategy map.

20. A non-transitory computer readable medium comprising instructions executable by a computer processor to perform: generating a strategy map using a language model during a session from transcript and a mapping prompt; selecting a plurality of sessions comprising the session; and generating an aggregated strategy map from the plurality of sessions.

Images