Systems and methods for goal-driven journey generation using artificial intelligence

Abstract

Systems and methods for providing omnichannel cognition in a campaign are disclosed. An example platform can include a strategy hub configured to generate and provide a unified and real-time graphical interactive interface displaying an effectiveness of the campaign, wherein the strategy hub is further configured to allow users to dynamically connect strategies to tactics and track key performance indicators (KPIs) while making real-time adjustments and rebalancing of efforts across multiple channels to optimize campaign engagement; a tactic engine configured to utilize a generative artificial intelligence (GenAI) model for analysis and restacking of the tactics to generate an optimal tactic mix that improves engagement efficiency and outcomes based on predicted impact levels of the tactics; and an impact simulator configured to provide a virtual platform for simulating the campaign through virtual trials, to assess an effectiveness of the campaign prior to roll-out of the campaign.

Description

SYSTEMS AND METHODS FOR GOAL-DRIVEN JOURNEY GENERATION USING ARTIFICIAL INTELLIGENCECROSS-REFERENCE TO RELATED APPLICATIONS

[0001] The present application claims priority to U.S. Provisional Application No. 63 / 625,840, filed January 26, 2024, which is incorporated by reference herein in its entirety.BACKGROUND

[0002] When an enterprise is trying to retain as many satisfied customers as possible and have meaningful engagement with potential targets, long-term and short-term considerations or goals often need to be considered and balanced. Although artificial intelligence has been used to make predictions as to which decisions may impact sales and marketing of products or services, more sophisticated analysis is needed to maximize sales and marketing goals (like sales or activity targets) and improve customer interactions / experience over time. In the healthcare space, pharmaceutical sales representatives (pharma reps) may use many different channels to communicate with healthcare professionals (HCPs). For a pharmaceutical company, quantitatively analyzing best practices in light of variations may be challenging, particularly when there is a need to consider and balance across multiple products, channels, representatives, HCPs, and customers.SUMMARY

[0003] There is a need for systems and methods that can utilize results from different machine- learned predictions and user-defined strategies to achieve at least one long-term goal and / or short-term goal that enhance communication effectiveness and engagement between actor types (e.g., pharma reps) and target entities (e.g., HCPs), across a multitude of channels over one or more time periods. The present disclosure can address at least the above need, by using machine learning models with rules and constraints to provide suggested actions to facilitate achieving of both long-term and short-term goals.

[0004] In an aspect, disclosed herein is a platform for providing omnichannel cognition in a campaign, comprising: a strategy hub configured to generate and provide a unified and realtime graphical interactive interface displaying an effectiveness of the campaign, wherein the strategy hub is further configured to allow users to dynamically connect strategies to tactics and track key performance indicators (KPIs) while making real-time adjustments andrebalancing of efforts across multiple channels to optimize campaign engagement; a tactic engine configured to utilize a generative artificial intelligence (GenAI) model for analysis and restacking of the tactics to generate an optimal tactic mix that improves engagement efficiency and outcomes based on predicted impact levels of the tactics; and an impact simulator configured to provide a virtual platform for simulating the campaign through virtual trials, to assess an effectiveness of the campaign prior to roll-out of the campaign. In some embodiments, the tactic engine is configured to generate triggered journeys based on triggering of events. In some embodiments, the triggered journeys comprise an escalated set of triggered journeys, and wherein escalating the set of triggered journeys is based at least on contact channel, degree of urgency, or recommended content. In some embodiments, generating the triggered journeys comprises defining triggering rules for the triggering of events and detecting when at least one triggering event occurs. In some embodiments, the GenAI model is updated after the at least one triggering event occurs, and wherein the updated GenAI model determines at least one new triggered journey for integrating with the existing triggered journey. In some embodiments, the tactic engine is further configured to generate ongoing journeys based on goals of the campaign. In some embodiments, the ongoing journeys comprise a continuous set of ongoing journeys, and wherein continuing the set of ongoing journeys is based at least on a contact channel, a degree of urgency, or a recommended content. In some embodiments, the ongoing journeys comprises defining the goals and detecting when at least one goal is achieved. In some embodiments, the GenAI model is updated after the at least one goal is achieved, and wherein the updated GenAI model determines at least one new goal for integrating with the existing ongoing journey. In some embodiments, the platform further comprises a contextual intelligence engine (CIE) configured to integrate or combine the triggered journeys and the ongoing journeys. In some embodiments, the events are defined by the users using the tactic engine. In some embodiments, the events comprise event definitions, and wherein the tactic engine is configured to automatically convert event logic into a set of trigger rules and a historical table for modeling, wherein the historical table comprises past event trigger indicators. In some embodiments, the GenAI model is configured to determine a new journey in real time using string triggers comprising the set of trigger rules. In some embodiments, the new journey is integrated into an existing triggered journey or an existing ongoing journey. In some embodiments, the platform further comprises an attribute model configured to learn attributes of past engagements with HCPs, wherein the learned attributes are based at least on interaction data and engineered features. In some embodiments, the engineered features are derived from the interaction data and compressed into a new data structure using at least datacompression to reduce computing resources of the platform. In some embodiments, the GenAI model comprises a large language model (LLM) configured to use prompt engineering for generating triggered journeys or ongoing journeys of the campaign based at least on a persona emulation. In some embodiments, the optimal tactic mix is based on a set of constraints comprising a set of global constraints and set of local constraints. In some embodiments, wherein the set of global constraints comprises constraints which apply to a type of actor or a type of target entity. In some embodiments, the set of local constraints comprises constraints which apply to an individual actor of the type of actor or an individual entity of the type of target entity. In some embodiments, the platform further comprises a reporting module configured to generate a structured report structured by metrics specific for a user for determining the effectiveness of the campaign.

[0005] In another aspect, disclosed herein is a computer-implemented method for providing omnichannel cognition in a campaign, the method comprising: generating and providing a unified and real-time graphical interactive interface displaying an effectiveness of the campaign, wherein the interface is further configured to allow users to dynamically connect strategies to tactics and track key performance indicators (KPIs) while making real-time adjustments and rebalancing of efforts across multiple channels to optimize campaign engagement; utilizing a generative artificial intelligence (GenAI) model for analysis and restacking of the tactics to generate an optimal tactic mix that improves engagement efficiency and outcomes based on predicted impact levels of the tactics; and simulating the campaign through virtual trials, to assess an effectiveness of the campaign prior to roll-out of the campaign.

[0006] In another aspect, disclosed herein is a modular platform for enhancing visibility and control over an omnichannel HCP engagement campaign, comprising: a strategy hub configured to generate and provide a unified and real-time graphical interactive interface displaying an effectiveness of the campaign, wherein the strategy hub is further configured to allow users to dynamically connect strategies to tactics and track key performance indicators (KPIs) while making real-time adjustments and rebalancing of efforts across multiple channels to optimize HCP engagement; a tactic engine configured to utilize generative artificial intelligence (Al) including large language models (LLMs) for analysis and restacking of the tactics to generate an optimal tactic mix that improves engagement efficiency and outcomes based on predicted impact levels of the tactics; and an impact simulator configured to provide a virtual platform for simulating the campaign through virtual trials, to assess an effectiveness of the campaign prior to roll-out of the campaign. In some embodiments, the tactic engine maybe configured to generate triggered journeys based on triggering of events. In some embodiments, the tactic engine may be configured to generate ongoing journeys based on goals of the campaign. In some embodiments, the modular platform may further comprise a contextual intelligence engine configured to integrate or combine the triggered journeys and the ongoing journeys. In some embodiments, the events may be defined by the users using the tactic engine. In some embodiments, the events may comprise event definitions, and wherein the tactic engine may be configured to automatically convert event logic into a set of trigger rules and a historical table for modeling, wherein the historical table comprises past event trigger indicators.

[0007] Additional aspects and advantages of the present disclosure will become readily apparent to those skilled in this art from the following detailed description, wherein only illustrative embodiments of the present disclosure are shown and described. As will be realized, the present disclosure is capable of other and different embodiments, and its several details are capable of modifications in various obvious respects, all without departing from the disclosure. Accordingly, the drawings and description are to be regarded as illustrative in nature, and not as restrictive.INCORPORATION BY REFERENCE

[0008] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference. To the extent publications and patents or patent applications incorporated by reference contradict the disclosure contained in the specification, the specification is intended to supersede and / or take precedence over any such contradictory material.BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The novel features of the present disclosure are set forth with particularity in the appended claims. A better understanding of the features and advantages of the present invention will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the invention are utilized, and the accompanying drawings (also “Figure” and “FIG.” herein), of which:

[0010] FIG. 1 illustrates an example of Al-driven system for generating and optimizing HCP contact sequences (e.g., journeys) between actor types and target entities.

[0011] FIG. 2 illustrates an example of Al-driven platform that generates HCP contact sequences (e.g., journeys), where the platform combines the use of deep reinforcement learning (DRL) and large language model (LLM).

[0012] FIG. 3 illustrates an example of trigger-based Al-driven platform that optimizes HCP contact sequences (e.g., journeys), where the platform combines the use of DRL and LLM.

[0013] FIG. 4 illustrates another example of trigger-based Al-driven platform that optimizes HCP contact sequences (e.g., journeys), where the platform combines the use of DRL and LLM.

[0014] FIGs. 5A-5B show a dataset of pharma rep-HCP interactions for fine tuning of the generative Al model. FIG. 5A shows the first portion of the dataset. FIG. 5B shows the second portion of the dataset.

[0015] FIG. 6 shows preprocessing of pharma rep-HCP interaction dataset for fine tuning of the generative Al model.

[0016] FIG. 7 illustrates an example of generating three-interaction journeys.

[0017] FIG. 8 illustrates an example of graphical user interface (GUI) for generating business strategies.

[0018] FIG. 9 illustrates another example of GUI for generating business strategies.

[0019] FIG. 10 illustrates another example of GUI for generating business strategies.

[0020] FIG. 11 shows a computer system that is programmed or otherwise configured to implement methods provided herein.DETAILED DESCRIPTION

[0021] While various embodiments of the invention have been shown and described herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only. Numerous variations, changes, and substitutions may occur to those skilled in the art without departing from the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed.Overview

[0022] The present disclosure is directed to a versatile Al-driven platform that may facilitate or enhance interactions and engagements between actor types and target entities. The interactions and engagements may relate to sales, marketing, advertising or promotion of one or more products or services to the target entities. The platform may be built as an Al-driven engine having several key elements that are combined in unique ways to solve real-world business problems. As an example, machine learning, optimization and algorithms may becollectively combined as described herein, to explain various actions and rules to stakeholders and customers in a transparent manner. The key elements of the Al-driven engine include (1) machine learning models, (2) rule / expert systems having rules and feature sets, (3) constrained optimization, (4) information solicitation and learning feedback loop, (5) natural language processing, and (6) contextual business intelligence and insights. The combination of the above elements will be apparent in light of the description of the examples herein.

[0023] The systems and methods as described herein surpasses existing approaches by providing an all-encompassing solution tailored to tackle global and local constraints. Its adaptability may extend to seamless integration with existing rule-based, segmentation-based, and machine learning (ML) / artificial intelligence (Al) and optimization next-best-action engines. For example, one advantage of the systems and methods as described herein is the capacity to enhance for evolving objectives, outlined by user-defined KPIs. Existing reinforcement learning approaches frequently encounter difficulties in navigating complex real-world situations and adjusting to changing environments. The systems and methods as described herein may signify a noteworthy advancement in overcoming these challenges, providing a more advanced and flexible solution.

[0024] The Al-driven platform may be an omnichannel, as the platform may orchestrate actions over a plurality of promotional / communication channels to optimize interactions and engagement between actor types and target entities. The orchestration may include coordinating different actions across multiple channels in a synchronized fashion, to ensure that the experience for actor types as well as target entities is seamless, impactful and rewarding for parties on all sides.

[0025] Given there are often multiple available channels between actor types and target entities, and numerous variables / constraints involved, this presents an omnichannel challenge which may be classified as an optimization problem. The steps towards solving this challenge include understanding the user’s needs, creating engaging user experiences, and ensuring flawless execution. Although the majority of this disclosure is described in the context of interactions between life science companies, pharmaceutical sale representatives and healthcare professionals, it should be noted that the present disclosure is not limited thereto, and may be applicable to a variety of other industries or fields involving user experience and sales / marketing / promotion of products. The Al-driven platform may be configured to ensure that every customer experience is tailored to their preferences and needs, thereby helping customers (e.g., life science companies) to strengthen their relationships with target entities (e.g., HCPs) in achieving better or improved patient care.

[0026] In some embodiments, the Al-driven platform may comprise a tactic engine that process long-term goals and / or short-term goals provided by actor types and generate business strategies. The business strategies may be used to facilitate or enhance interactions and engagement between the actor types and target entities. Unlike traditional business plans including abundant detailed rules, the strategies may comprise an execution plan generated by the tactic engine. Some non-limiting examples of the execution plan may comprise how to work with a given group of users, how to treat a group of users effectively for a given period of time and / or under given circumstances. When the actor types determine obj ect targets, goals, audience (e.g., audience clinical practitioners (ACPs)), which are used as input to the tactic engine, the engine may generate one or more strategies targeted to the goals and audience.

[0027] The input to the Al-driven platform may comprise one or more of KPIs, goals, targets, objectives, content, audience, events, or event triggers. Each type of input may be defined by users interactively. In some embodiments, the Al-driven platform may provide a graphical user interface (GUI) for users to define the input and review the output generated therefrom by the platform.

[0028] The output may comprise a plurality of target entity contact suggestions. A sequence of target entity contact suggestions may refer to a journey. For example, a journey may comprise sequences of contact with information about when and what channel to use with the contact (e.g., email, in-person visit, virtual meeting), what content to use (i.e., topic to be conveyed), and degree of urgency. A collection of journeys may refer to a tactic. Journeys that correspond to similar audiences may be grouped as a tactic.

[0029] The output generated by the Al-driven platform may be presented to users via the GUI. Users may respond to the output at the planning stage before the output is executed. In addition or alternatively, users may provide feedback and modify / adjust the output. When the output is modified / adjusted, the Al-driven platform may take into consideration the modification and generate a new journey.

[0030] The actor types may comprise one or more enumerated actors including marketing / sales representatives, district managers, or medical science liaisons that are associated with the promotion of the product to the target entities. In other embodiments, the actor types may comprise automated systems including marketing automation systems or web portal management systems.

[0031] The target entities may comprise one or more HCPs, healthcare organizations (HCOs), or healthcare institutional accounts. Examples of HCPs may include doctor, nurse, nursepractitioner, physician assistant, technician, physical therapist, occupational therapist, health aide, respiratory therapist, or clinical psychologist.

[0032] In some embodiments, a product may be a health product. The product may be a pharmaceutical product. The product may comprise a drug, a therapeutic, a health treatment, a medical device, or any health or healthcare product to treat one or more health or disease- related conditions. A drug may be a prescription drug or an over-the-counter (OTC) drug. A health treatment may be a therapy or surgery.

[0033] In some embodiments, long-term goals of a user (e.g., pharma company) may include foundational targeting, such as prioritization of HCPs (or HCP segments) based on long-term opportunities for engagement or sales to the associated HCOs. A long-term goal may be updated as frequently as necessary, for example every 6-12 months by the user (e.g., pharma company).

[0034] In some embodiments, short-term goals or metrics may include (1) Al-based monitoring of sales data to identify any significant deviations, (2) automatic detection of new prescribers or newly diagnoses patients, based on recent drug product sales or medical claims; (3) strategic actions to progress actor types and / or target entities through configurable, complex journeys involving their interactions, or (4) pre-defined HCP engagement sequences that are designed to drive day-to-day re engagement with one or more HCPs. The short-term goals or metrics may be updated on a shorter timescale than the long-term goals. For example, the short-term goals or metrics may be updated daily, weekly or monthly.

[0035] The Al-driven platform may generate one or more models. The models may comprise one or more of an out-of-box (OOB) model, a generative Al (GenAI or GA) model, or a custom model. In some embodiments, the Al-driven platform may process one or more input and generate a generic model. In other embodiments, when users define some or all of the input, the platform may process these specified input and generate a custom model. These models may be executed, which may generate journeys.

[0036] The Al-driven platform may comprise a reward function, the definition of which may be determined by the input to the platform. The reward function may comprise deep reinforcement learning (DRL) model and LLM. The DRL model may process the input and generate a journey that may be a good candidate of sequences. The LLM may learn the journey and generate a new and / or better journey. In some embodiments, the LLM may process other input that may be relevant to the generated journey. The LLM may be modified by one or more reinforcement learning steps including initial fine tuning, final fine tuning, and humanfeedback fine tuning. In some embodiments, the LLM may be a local model after the initialfine tuning. The local LLM may ensure the security and privacy of the model, as well as the high efficiency. The final fine tuning may comprise quality reinforcement learning and toxicity detection. The toxicity may generally refer to irrelevant or problematic output generated by a machine learning model. For example, an LLM may perceive patterns or objects that are nonexistent or imperceptible to human observers and create outputs that are nonsensical or inaccurate. The final fine tuning may improve the capability of the LLM in generating journeys with high quality.

[0037] In some embodiments, the DRL model may process historical data comprising sequences of past interactions between the actor types and target entities. The DRL model may also comprise a simulator that generates hypothetical or counterfactual sequences. The DRL model may augment data with hypothetical sequences of actions and / or estimations. The LLM may learn the reward associated with the augmented data and generate a new / better journey of sequences. In other words, the journey of sequences may be generated from actual sequences from historical data as well as counterfactual sequences generated by the simulator.

[0038] The initial fine tuning of the LLM may comprise large-scale prompt engineering. The LLM may be trained by a plurality of prompts in an iterative manner. In some embodiments, the iterative training may comprise a plurality of iterations using a plurality of records and examples. The number of iterations may be about 10, about 50, about 100, about 500, about 1,000, about 5,000, about 10,000, about 50,000, about 100,000, about 500,000, about 1,000,000, or more. The number of records and examples may be about 10, about 50, about 100, about 500, about 1,000, about 5,000, about 10,000, about 50,000, about 100,000, about 500,000, about 1,000,000, or more. In some embodiments, some of the actual sequences of historical data may be turned into a prompt. The LLM may learn the prompt and generate a new / better journey of sequences. As the LLM may undergo fine tuning processes that add more refinements to the content of the sequences, the generated journey of sequences may better fit the goals provided by users.

[0039] The Al-driven platform herein may integrate with a full range of data sources, analytics, and marketing technology. The platform may comprise a contextual intelligence engine (CIE) that is configured to optimize actions across different segments of target entities, across multiple channels, and across multiple temporal periods. The CIE may include one or more machine learning modules, customer journey management (CJM) tools, and tools for execution management workflows. The CIE may be referred to interchangeable herein as an optimization module, an optimization engine, an optimizer, and the like. In some embodiments, an optimization module, an optimization engine, or an optimizer may comprise a component orsubcomponent that is integrated into as part of the CIE. In other embodiments, the optimization module, the optimization engine, or the optimizer itself may be the CIE. In further embodiments, the CIE may comprise a plurality of optimization modules, optimization engines or optimizers.

[0040] The CIE may be configured to process data, rules, campaigns and models from any source. The CIE may also utilize a common data model and framework for omnichannel action evaluation, and recommendations / optimization incorporating feedback from multiple channels. In some embodiment, the output generated by the Al-driven platform may be processed by the CIE. For example, the journeys generated by the platform may be executed by the CIE.

[0041] The CIE may synthesize suggestions across multiple brands and products while incorporating practical considerations for each user. For example, the CIE may be configured to combine / synthesize recommendations across multiple brands / products. The CIE may be used for, or used by brand managers. For example, the CIE may be used to generate visual reports, notifications and recommendations to brand managers. Brand managers may use the CIE to create one or more campaigns for different brands / products that are targeted to the same or different segments of HCPs. A detailed description of CIE can be found in PCT Patent Application No. PCT / US2022 / 050155, Publication No. WO / 2023 / 091519, filed November 16, 2022, which is incorporated by reference in its entirety.Al-driven platform

[0042] FIG. 1 illustrates an example of Al-driven system for generating and optimizing HCP contact sequences (e.g., journeys) between actor types and target entities. The system 100 (e.g., Al-driven platform) allows users (e.g., actor types) to define input to the Al-driven platform, including objectives, targets, goals, KPIs, and the like (see 110 of FIG. 1). The system 100 provides a GUI for users to provide input and review the journeys generated by the platform. The system 100 performs reinforcement learning to generate HCP contact sequences (see 120 of FIG. 1). The DRL model may process historical data comprising sequences of actual interactions between the actor types and target entities, as well as hypothetical or counterfactual sequences. The LLM may learn from the output generated by the DRL model and generate HCP contact sequences with high quality.

[0043] The system 100 (e.g., Al-driven platform) may utilize data stored in the customer relationship management (CRM) system. These histories encapsulate detailed past engagements with HCPs across various communication channels. In some embodiments, these data may be organized into two data structures comprising interaction content data andengineered features. The interaction content data may capture the specifics of each interaction, including the content exchanged. The engineered features may comprise derived attributes that enhance the understanding of the data. The Al-driven platform may package these data in a way that may optimizes the efficiency and reliability during the development and deployment of the Al-driven models.Content targeting model

[0044] As further illustrated, the system 100 comprises a content targeting model (CTM) with a diverse counterfactual explanations (DiCE) model (see 130 of FIG. 1). The CTM and DiCE models may be configured to facilitate content management. Some non-limiting examples of journey content may comprise suggestions to pharma reps that appropriate for generating impactful content in different contexts. The suggestions can be targeted towards different HCPs with a probability of success. The predicted probabilities of success can provide a confidence in the HCP of prescribing a drug had their engagement history (e.g., communication history) been different (for example, if the HCP had responded differently to a survey question in the past).

[0045] In some embodiments, content topics may be tagged (see 130 of FIG. 1). Tagging may include tagging emails from pharma reps, tagging emails from HCPs, tagging visit notes in which a rep indicated that the HCP is likely to be comfortable prescribing a drug, or tagging a dataset with a high level description of its content. A tag may indicate that an email relates to setting up a meeting with an expert. Another tag may indicate than an email is about patient compliance material. Another tag may indicate that an email is about dosing. The content items may be tagged in ways that are useful for business analytics. The content may be tagged by relevant (e.g., expert) personnel, for example a scientist or business analyst at the pharma company. In some embodiments, the tagging of the content may be partially or fully automated. The tagging may be subsequently reviewed and curated if necessary, by subject matter experts and / or customers, to meet a company’s business needs and also for regulatory compliance.Generative Al model

[0046] The system 100 further comprises an OOB GenAI model and a custom GenAI model (see 140 of FIG. 1). In some embodiments, the system 100 (e.g., Al-driven platform) may process one or more input and generate a generic model. In other embodiments, when users define some or all of the input, the system 100 (e.g., Al-driven platform) may process these specified input and generate a custom model. These models may be executed, which may generate the journeys that can be executed by CIE (see 150 of FIG. 1) and reviewed by users via the GUI (see 160 of FIG. 1).

[0047] The system 100 (e.g., Al-driven platform) may define the state for the reinforcement learning model. The state may encapsulate characteristics of HCPs, their behavioral attributes or dispositions, and contextual information relevant to their positions in the HCP contact journeys generated by the platform. By defining a comprehensive state, the Al-driven platform may gain a holistic understanding of the environment, enabling more nuanced decisionmaking.

[0048] In some embodiments, a deep learning model may be trained to capture the state. The deep learning model may be trained on one or more of packaged CRM data, learning patterns, and representations that may encapsulate the complexities of HCP engagements. The training process may involve adjusting the model's parameters to minimize the difference between its predictions and actual outcomes, ensuring it becomes adept at understanding the underlying structures within the data and changing dynamics in the data due to simulated actions taken by agent in the model.

[0049] The system 100 (e.g., Al-driven platform) may utilize a reward function to quantify the success of pharma rep-HCP interactions. The reward function may capture key HCP values, key business and financial goals, propensities, constraints, KPIs, and the like. The reward function may align with the central objectives of the pharma rep-HCP engagement strategy, reflecting desired outcomes and incentivizing actions that may contribute positively to these objectives or penalizing actions that may contribute negatively.Behavioral simulation model

[0050] The system 100 (e.g., Al-driven platform) may comprise a behavioral simulation model that is trained to predict the responses of pharma reps and HCPs to actual historical actions and / or to counter-factual actions. The behavioral simulation model may facilitate the exploration of actual historical actions and hypothetical scenarios (e.g., counter-factual actions) within a decision space. The decision space may span various options, including interaction channels, content or topic categories, and the timing of HCP contacts. In some embodiments, the training process of the behavioral simulation model may comprise learning the dynamics of how different actions influence the behavior of pharma reps and / or HCPs.

[0051] The behavioral simulation model may comprise a plurality of control parameters that may be used as levers / dials to make granular adjustments to the behavior of the platform. Examples of those parameters may include return on investment (ROI), certain thresholds, matrix parameters, and constraints at different levels. The control parameters may be adjusted to determine the behavior of the system, along with actual empirical data about HCPs and machine learning models.

[0052] The DRL model may be trained to estimate the expected business reward for all feasible actions and HCPs. The DRL model may be trained using a controlled mixture of actual and simulated (counter-factual) interaction history data. These history data may encapsulate detailed past engagements with HCPs across various communication channels. A ratio between actual and simulated (counter-factual) interaction history data within the controlled mixture may be configurable. For example, the ratio between actual and simulated data may be about 1000: 1, about 500: 1, about 100: 1, about 50: 1, about 10: 1, about 5: 1, or the like. The use of simulated (counter-factual) actions as part of the training data set may allow the DRL model to learn from a diverse set of scenarios, and generalize well to different situations over a period of time. Moreover, adding simulated (counter-factual) actions may resolve the issue that there is insufficient training dataset to optimize a machine learning model.

[0053] The trained DRL model may be executed iteratively to generate HCP contact sequences (e.g., journeys) over a desired number of days based on actual historical and / or simulated actions. The iterative process may refine the model’s understanding over time, incorporate feedback and adjust recommendations based on the evolving dynamics of the HCP engagement landscape in different situations. To further improve the quality of the generated journeys, they may be used as input to fine tune the LLM. The fine-tuning process enhances the details of the topics associated with each step in the journey. The LLM may leverage its language generation capabilities, enrich the content and context of the journeys, and make them more meaningful and personalized for HCPs and assist sales representatives for better interactions.

[0054] In some embodiments, one or more specialized prompts (or engineered prompts) may be used for the fine-tuned LLM to generate finalized HCP contact sequences. The use of prompts may improve the capability of the LLM in generating journeys that align with particular criteria, for example, context relevancy, tone control and toxicity reduction. In some cases, the prompts may be based on a persona setting or persona emulation when generating journeys. A persona setting or persona emulation can refer to imitating or assuming a certain personality, style, or character. A persona setting or persona emulation can improve responses of the LLM when determining a campaign by causing the captain to be more engaging, interesting, or tailored to a certain context or audience, e.g. engagements between reps and HCPs. For example, using a persona setting or personal emulation can involve emulating the speaking style of a particular rep with a past history of successful engagement with HCPs.

[0055] The LLM may be used as a crucial facilitator for practical and comprehensive management of contact suggestions’ contents and topics. In contrast to existing approaches that are often insufficient for the intricate crafting of personalized sequences, the LLM mayintroduce a dimension of linguistic comprehension and generation. This may enable the system 100 not only to customize the content of HCP contact suggestions but to do so in a way that holds significance and appropriateness for intended customers. The linguistic generation abilities of the LLM may play a role in developing contextually abundant and compelling sequences, elevating the general efficacy of HCP engagements.

[0056] Furthermore, the system 100 may be designed to account for both global and local constraints during the modelling and execution process. Global constraints generally refer to constraints that go across a population of actor types or target entities. Some non-limiting examples may comprise reps that cannot work more than seven or eight hours a day or doctors who do not want to have more than a given amount of virtual meetings in a quarter. Local constraints generally refer to constraints about individual entities, for example, if a given rep had a visit or any contact last week, he / she will not be recommended for another visit in the near future. The platform may consider customer objectives and constraints (e.g., global) while also addressing specific local factors and limitations (e.g., local). This capability may be crucial for crafting journeys that align with broader strategic goals while remaining flexible enough to adapt to unique circumstances at the individual or segment level.

[0057] In some embodiments, after tone control and toxicity filter are applied to the LLM for final tuning, the generated journeys may be combined with other systems, including rule-based system, ML / AI driven next-best-action system. The integration may ensure the generated journeys align with broader engagement strategies and adhere to any additional predefined rules or constraints.

[0058] The input data may comprise educational or research data, market data, strategies, analytics, content, context, campaigns / rules, or models from one or more data sources. As examples, the educational or research data may include scientific literature that a target entity may be interested in to keep apprised (or learned) in the target entity’s field, e.g., field of cancer research. Campaigns may be determined based on the educational or research data for purposes of educating a target entity instead of for purposes of marketing or sales. The input data may include product sales information, information regarding the communication frequency of a rep to an HCP, types of products promoted, or brand or marketing strategy rules. The market data may include sales numbers for products, company revenue numbers, sales statistics for particular HCPs, sales statistics aggregated from multiple HCPs, or pharmaceutical industry metrics. The strategies may include achieving quarterly sales targets, achieving quarterly revenue targets, or achieving quarterly profit targets. The analytics may include interaction data for specific pharma reps. The context may include contextual information that is specific toeach channel, rep or HCP. The content may include customized content that is suggested to reps, and for targeting towards individual HCPs or segments. The campaigns may include advertising campaigns or product launch campaigns. The models may include sales models or forecasts or revenue models or forecasts.

[0059] The sources of the input data may include databases for scientific literature, research literature, customer relationship management (CRM), marketing automation systems, website target management systems, customer data platforms (CDP), market segmentation, sales, a pharmaceutical company's proprietary databases, and / or third party provided / generated data. In some embodiments, the input data may include field notes taken by reps in their interactions or visits with HCPs.

[0060] A plurality of temporal goals for attainment may be defined based at least in part on the input data. The plurality of temporal goals may be classified into at least two categories of different time scales. There may be at least 2, at least 3, at least 4, at least 5, at least 10, at least 15, at least 20, at least 30, at least 40, at least 50, or at least 100 temporal goals. There may be more than 2, more than 3, more than 4, more than 5, more than 10, or more than 20 time scales. The categories may include both long-term and short-term. In some cases, the plurality of temporal goals comprise at least one long-term goal on a first time scale ranging from about six to twelve months, and a plurality of short-term goals on a second timescale that may be daily, weekly, or monthly. The plurality of temporal goals may be expressed in a form of actions and / or responses to actions on the different timescales.

[0061] In some embodiments, at least one long-term goal may be provided on a first timescale. The first timescale may range from about six to twelve months. The first timescale may be at least 2 months, at least 3 months, at least 4 months, at least six months, or at least a year. A plurality of short-term goals may be provided on a second timescale. The second timescale may be daily, weekly, or monthly.Contextual intelligence engine

[0062] The CIE may be configured to select target, channel, content and timing to optimize the customer experience / journey (e.g., interact! ons / engagem ent between reps and HCPs). The CIE may be configured to process a plurality of OC action candidates including rules, campaigns and ML model outputs, which are subject to constraints, filtered and consolidated before being provided to an optimization model. The optimization model may be configured to take into account a number of factors (e.g., educational value, value, urgency, priority, cost, content affinity, and channel affinity) in selecting the appropriate target entity, channel, content and timing to optimize the customer experience / journey (e.g., interactions / engagement betweenreps and HCPs). The CIE may be flexible and scalable, in that any number and type of channel, account and action candidates may be added or removed at any given time. The customer experience / joumey may be optimized for transparency, where the next best and alternative actions are presented to the actor types (e.g., reps), with explanations of the effects of those actions (having quantifiable component values), and aided by simulation and visual reporting to the actor types (e.g., reps and stakeholders).

[0063] The CIE may comprise an objective function. A plurality of temporal goals and a set of constraints may be input to the optimization model. The channels may enable or facilitate communications between one or more actor types and the one or more target entities. The channels may include communicating with an HCP via email, short message system (SMS), telephone, web chat, in-person, or any other communication channel. There may be at least 1, 2, 3, 4, 5, 10, 15 or more types of communication channels.

[0064] The set of constraints may comprise action-type constraints, channel constraints, channel capacity constraints, pacing constraints, educational or scientific constraints, and / or channel fatigue constraints across a plurality of channels. An action-type constraint may include a restriction on an action or communication that may be performed by a rep when engaging with an HCP. A channel constraint may include a constraint on a particular communication method (e.g., email or text) between the rep and HCP. A channel capacity constraint may include a volume or frequency of communications that are routine / standard for each channel. A pacing constraint may include a timing or frequency by which the rep sends communications to an HCP. A channel fatigue constraint may include overuse of certain channels that may cause the rep and / or HCP to shun or shy away from those channels.

[0065] The CIE may be used to generate a set of multi-dimensional actions with predicted action values for at least a subset of the plurality of channels. The predicted action values may collectively maximize an impact value of the objective function while balancing a feasibility of the plurality of temporal goals with respect to one another. The set of multi-dimensional actions with predicted action values may be generated based on one or more of the following: a value of the one or more target entities, a relative impact value of the actions, a probability of success of the actions, and a timing value of the actions. The predicted action values may include scores (e.g., of 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 out of 10), or probabilities, or combinations of scores and probabilities. The subset of channels may include at least 1, at least 2, at least 3, at least 4, or at least 5 of the channels. The balancing may be implemented based on the predicted action values.

[0066] The objective function may be configured to return a value function that may be a product of a targeted entity value, action impact values, engagement probabilities, and success probabilities for achieving the plurality of goals, and time values of the actions.

[0067] The set of recommended user multi-dimensional actions may be designed to complement each other so as to enable cross-channel optimization across the subset of channels, based at least in part on an execution capacity of the enumerated actors, capacity limits set of the automated systems, utilization or consumption of content, or usefulness / relevance of the content.

[0068] The CIE may be configured to iteratively search for actions to maximize expected action values by applying and varying the plurality of constraints and parameters to (1) influence the predicted action values and (2) monitor changes in the predicted action values. The optimization model may be configured to iteratively apply and vary a plurality of parameters to influence the predicted action values and monitor changes in the predicted action values. The optimization may apply at least 2, at least 3, at least 4, at least five, or at least 10 iterations. The plurality of parameters comprise one or more of the following: cost, impact, urgency, recency, channel propensity, physical proximity, representative (rep) engagement, time to engage, account priority, or factor priority.

[0069] The CIE may be configured to search through a range of predicted action values for the plurality of channels when the plurality of parameters is being varied subject to the set of constraints. This may be performed to identify a subset of predicted action values that form a basis for the set of multi-dimensional actions.

[0070] The set of multi-dimensional actions with predicted action values for at least a subset of the plurality of channels may be provided to the CRM system to facilitate optimization of the customer joumey / experience. Those actions with predicted action values may be used in field sales (e.g., remote calls by reps to HCPs, face-to-face meetings between reps and HCPs, and emails from reps to HCPs), by field medical personnel (e.g., MSLs and patient support), in marketing automation (e.g., HQ email, messaging to customers, mailers), digital platforms (e.g., social media, web portals, or third party conferences / events), or for educational and scientific engagements (e.g., educating target entities about current scientific understanding of diseases and treatments thereof).Generating and optimizing journeys

[0071] FIG. 2 illustrates an example of Al-driven platform that generates HCP contact sequences (e.g., journeys), where the platform combines the use of DRL model and LLM. The DRL model may process input defined by users and generate a journey that may be a goodcandidate of sequences. The LLM may learn the sequences and generate a new and / or better journey. The LLM may be modified by one or more reinforcement learning steps including initial fine tuning, final fine tuning, and human-feedback fine tuning. In some embodiments, the LLM may be a local model after the initial fine tuning. The local LLM may ensure the security and privacy of the model. Moreover, as the local LLM has 50-time fewer parameters compared to a global model, it may have a significantly high efficiency. The final fine tuning may comprise quality reinforcement learning, tone control, and toxicity detection. The final fine tuning may improve the capability of the LLM in generating journeys with high quality.

[0072] The Al-driven platform as described herein may be used in a variety of use cases. A non-limiting example may be the Al-driven platform generates a new business strategy which is executed on an ongoing basis. A non-limiting example may be the Al-driven platform generates an educational or scientific program for educating target entities bout current understandings of diseases and treatments thereof. Another example may be the Al-driven platform generates a new strategy which is executed in a given period of time. For instance, the generated strategy may be a campaign strategy that is not executed after the campaign is complete. Another example may be a trigger-based journey generation, where users define a triggering event and requirements for the generated j ourneys. When the triggering event occurs, the Al-driven platform may automatically generate a journey following the requirements defined by the users. In some embodiments, in a string trigger scenario, the Al-driven platform may generate a journey based on a triggering event, which is executed on an ongoing basis. During the execution, another triggering event may occur. The Al-driven platform may generate another journey and integrate it with the existing ongoing journey. The Al-driven platform may have system awareness (or journey awareness) that identify the priorities of different types of journeys. For example, a given type of trigger-based journeys may have a higher priority than existing ongoing journeys. The system awareness may be stored and used during the execution of the journeys by, for example, CIE (150 in FIG. 1). The journey generation based on triggering events will be described in more detail below in accordance with FIGs. 3 and 4.

[0073] FIG. 3 illustrates an example of trigger-based Al-driven platform that optimizes HCP contact sequences (e.g., journeys), where the platform combines the use of DRL and LLM. In some cases, the CIE is configured to integrate or combine triggered journeys and an ongoing journey.

[0074] The Al-driven platform may allow users to provide one or more input such that the platform can generate trigger-based journeys. Users may provide the definition of a triggeringevent and requirements for the generated journeys when the event occurs. In some embodiments, the logic of the triggering event may be automatically converted to a trigger rule and a historical table for modeling including past event trigger indicators. For example, the triggering rule may be “if an HCP responds to an email informing that he / she is interested in a conference being planned, then provide the HCP with a URL of the website where the HCP can find information about the conference.” The historical table may list historical data in the past year with indicators of similar events that occurred to all HCPs.

[0075] Other input to the Al-driven platform may comprise audience definition including an audience definition table (ADT). ADT may comprise a two-dimensional segmentation-based table with weights. For example, the ADT may comprise a table of tier x customer journey (CJ) priority segments defined the audience in a flexible manner. Content constraints may also be used as input to the Al-driven platform. The content constraints may comprise one or more of a content assignment table (CAT) and a specification of approved content or a subset of approved content. KPIs and other related goal / objective metrics may also be used as input to the Al-driven platform. KPIs and goal / objective metrics with weights may be incorporated into custom GenAI modeling via a custom deep reinforcement learning reward function. The modeling dataset may also be enriched by a history of KPI values. In addition, other metadata including the name of the trigger and description of its use may also be input to the platform.

[0076] As illustrated in FIG. 3, the event logic and KPIs with weights may be converted to an execution ready table comprising event indicators and KPI evaluations. The table may need to be updated on a regular basis, for example, daily, weekly, or monthly. When a triggering event occurs, the execution ready table may indicate current possible interactions, e.g., what interactions may be relevant to the triggering event. In other embodiments, the event logic and KPIs with weights may be converted to a modeling dataset comprising event indicators and KPI evaluations. The data may need to be updated on a regular basis, for example, weekly or monthly. The modeling dataset may be used by the DRL and LLM models. The KPI information may be used by the reward function.

[0077] The Al-driven platform may allow different approaches for trigger-based GenAI modeling. In some embodiments, the modeling may be on a regular basis, for example, daily, weekly, or monthly. Triggering events may be surveyed and classified as existing or new. Only new types of triggering events may require GenAI model development. In other embodiments, new types of triggering events may trigger the modeling. Upon the completion of modeling, the journey generation may commence.

[0078] FIG. 4 illustrates another example of trigger-based Al-driven platform that optimizes HCP contact sequences (e.g., journeys), where the platform combines the use of DRL and LLM. As illustrated, the Al-driven platform allows monthly model update and execution. Journeys generated from the platform may be executed on a monthly basis. CIE with system awareness (or journey awareness) may execute the journeys. Moreover, the Al-driven platform allows trigger-based journey generation and execution. The monthly model update and execution as well as trigger-based journey generation and execution may be executed by the CIE. The CIE with system awareness (or journey awareness) may execute both modeling in parallel. When the model is developed, it is a fixed object until redeveloped. When a triggering event occurs, the CIE may receive suggestions from the monthly journeys and trigger-based journeys.

[0079] In some embodiments, the CIE may execute basic journeys. The CIE may prioritize journeys for execution. For example, journeys may have priority scores assigned by the DRL, users, or a combination thereof. The scores may be used for prioritizing and weighing suggestions from multiple sources in the CIE. The priority scores may be in a numerical form or categorical forms, for example, high, medium, or low. The priority scores may be generated based on quality score of the DRL and / or model-estimated reward. The score may also vary by, for example, journey steps. The CIE may have system awareness (or journey awareness). In other words, the CIE may be aware of whether or not a suggestion is part of a journey and prioritize it accordingly. The CIE may use an Influencer- Adjustment mechanism to prioritize one or more suggestions as needed. The priority of journeys may be adjustable. When ajourney step fails to pass the CIE optimization on day N, its priority may be automatically increased, for example, on day N+l, N+2, . . ., N+M. Based on whether the step passes the optimization, its priority may be increased to an absolute priority (e.g., urgent).

[0080] In other embodiments, the CIE may combine ongoing journeys with trigger-based journeys. In the daily execution process of a trigger-based journey, the CIE may determine that a suggestion is grounded in an HCP journey with an existing active journey underway. For example, suggestions from a new journey (e.g., trigger-based journey) may be automatically weighed higher than suggestions from a current journey.

[0081] In some embodiments, the systems and methods as described herein may have a workflow for developing and optimizing journeys. The workflow may comprise the steps of defining the use case; pretraining a model or choosing an existing model; adapting and aligning the model; and integrating the application.

[0082] The use case may be defined as using GenAI or LLM model to generate journeys. The journey may comprise sequences of contact channels, content or topic conveyed to audience, and degree of urgency. The task of journey generation may comprise generation of sequential recommendations in the form of text-to-text generation.

[0083] The next step is pretraining a model or choosing an existing model. In some embodiments, an existing language model may be used to manage sequence-to-sequence transfer tasks. A non-limiting example may comprise text-to-text transfer transformer (T5) that is configured to a variety of natural language processing tasks. The use of existing models provides the flexibility of using large language model as a starting point for optimization, as large models are more capable of carrying out complex tasks.

[0084] The existing model may be optimized for the defined use case. The step of adapting and aligning the model may comprise prompt engineering, fine-tuning and aligning with human feedback. Compared to the pre-training process where the model uses a vast amount of unstructured textual data via self-supervised learning, fine tuning is a supervised learning process where a dataset with labeled examples can be used to update the weights of the model. In some embodiments, the labeled examples may comprise prompt completion pairs to optimize the model and improve its capability of accomplishing given tasks.

[0085] FIG. 5 shows a dataset of pharma rep-HCP interaction for fine tuning of the generative Al model. The dataset comprises previous pharma rep-HCP interactions that are used as input sequences for the model, and current interactions that are target sequences for the model. For each HCP, the input sequences comprise lagged contact channels (e.g., web interactive, email, and visit), lagged degree of urgency (e.g., not urgent, somewhat urgent, and urgent), and lagged content of the sequence. The target sequences comprise contact channels, degree of urgency, and content of the sequence.

[0086] FIG. 6 shows preprocessing of pharma rep-HCP interaction dataset for fine tuning of the generative Al model. The current interactions and recommended interactions (prompt response) are converted into explicit instructions for the LLM. As illustrated, the training prompt (current interaction) lists the previous contact channel, degree of urgency, and content for a given HCP. The fine tuning target lists the recommended interaction with an HCP journey trend, contact channel, degree of urgency, and content. For example, a given HCP (HCP unique identification number OOlAOOOOOOkZOgOIAS) interacted with an actor type regarding specific content of “SBR Guideline; Skin - sustained efficacy; Axial Spondyloarthritis Lamina” (contact channel: visit; degree of urgency: low). After the fine tuning of the model using thishistorical or current interaction data, the model may recommend the interaction with content of “Cosentyx minimally as good as Humira.”

[0087] FIG. 7 illustrates an example of generating three-interaction journeys. As illustrated, the Al-driven platform generates three-interaction journeys for a given HCP. Each journey includes a contact channel, degree of urgency, and recommended content. For example, for pain and fatigue that is not urgent, the platform recommends in person visit. For fast and sustained response using Cosentyx® vs anti-TNFs that is somewhat urgent, the platform recommends virtual interaction. For the content of “Cosentyx® minimally as good as Humira” that is urgent, the model recommends email which may bring quick attention from the HCP.Actionable user interfaces

[0088] FIGs. 8-10 are examples of GUIs for users to define input to the Al-driven platform, which in turn generates journeys. FIG. 8 illustrates an example of GUI for generating business strategies. Users are allowed to define a goal, confirm the audience of the to-be-generated journey, establish the message mix, and verify the KPIs. Some or all of the information may be used as input to the Al-driven platform to generate journeys. FIG. 9 illustrates another example of GUI for generating business strategies. Users are allowed to select a brand / product as part of the goal confirmation. The platform may combine / synthesize recommendations across brands / products. FIG. 10 illustrates another example of GUI for generating business strategies. When a particular brand / product is selected as a confirmed goal, the Al-driven platform can generate corresponding journeys. Users are allowed to review the journey progression, customize j oumey messaging, and other tasks.Data analytics and reporting

[0089] Any of the data infrastructure systems herein may include one or more server computer systems for physically storing and processing data. The server computer systems may be housed at one or more facilities. If data storage is decentralized, the data infrastructure system may use a cloud storage model. Data stored within the data infrastructure system may be managed by one or more types of database (e.g., a relational database or a non-relational database), using languages such as SQL or NoSQL. Some of the databases in the database may be implemented using various standard data structures, such as an array, hash, (linked) list, struct, structured text file (e.g., XML), table, JavaScript™ Object Notation (JSON), NOSQL and / or the like. Such data structures may be stored in memory and / or in (structured) files. The database may comprise object-oriented databases. Object databases may include a number of object collections that are grouped and / or linked together by common attributes; they may be related to other object collections by some common attributes. Object-oriented databases mayperform similarly to relational databases with the exception that objects are not just pieces of data but may have other types of functionality encapsulated within a given object. The database may include a graph database that uses graph structures for semantic queries with nodes, edges and properties to represent and store data. Also, the database may be implemented as a mix of data structures, objects, and relational structures. Databases may be consolidated and / or distributed in variations through standard data processing techniques. Portions of databases, e.g., tables, may be exported and / or imported and thus decentralized and / or integrated. The data infrastructure system may also include networking equipment for providing data to various parts of the system, including to end users. The networking equipment may provide access to the Internet or to a local network (e.g., a local area network (LAN)), or to a wide area network (WAN). The networking equipment may include routers, switches, cabling, access points, gateways, and other networking hardware. The data infrastructure system may also include software to authenticate and authorize users for access to various parts of the system. The data infrastructure system may also include software to pre-process data (e.g., to compress or deduplicate the data).

[0090] Any of the data analytics systems herein may process the data to generate analytics and predictions useful for increasing effectiveness of communications between pharma reps and HCPs. The data analytics system may include computer hardware and software for generating machine-learned predictions regarding communication methods, and additional hardware and software to convert these predictions into scores or probabilities which convey information to pharma reps about best communication practices. The data analytics system may interface with various portions of the data infrastructure system. For example, data infrastructure components may send intermediate or final results to various portions of the data analytics system. The data analytics system may access databases within the data infrastructure system to extract data used for machine learning processes. The data analytics system may implement one or more machine learning pipelines to train machine learning models, automate machine learning tasks, provide results to various components of the system, and perform other data processing tasks. A machine learning pipeline may include a registry with one or more machine learning models to be implemented by the system.

[0091] Any of the customer relationship management (CRM) systems herein may provide or generate structured electronic reports to pharma reps or their management, to show various metrics with respect to educating HCPs, targeting HCPs, automate sales and marketing processes, and provide additional data to such users and other stakeholders within a user interface. The user interface may be a graphical user interface. The CRM system may receiverecommended actions for optimizing performance with respect to long-term and short-term goals and present them to one or more users.

[0092] Any of the machine learning models may receive pre-processed data or may pre-process data. For example, a machine learning model may alter the dimensionality of the input data. The machine learning model may create an encoding or a representation of the input data. The machine learning models may be conducted over multiple epochs or iterations. In some embodiments, the machine learning model may reserve a validation set of data to validate the training of the algorithm, to determine whether more training is necessary. The trained model may then be tested on data to generate the predictions.Developing machine learning models

[0093] The machine learning model may use supervised machine learning (ML) algorithms. A supervised ML algorithm may be trained using labeled training inputs, i.e., training inputs with known outputs. The training inputs may be provided to an untrained or partially trained version of the ML algorithm to generate a predicted output. The predicted output may be compared to the known output, and if there is a difference, the parameters of the ML algorithm may be updated. A semi-supervised ML algorithm may be trained using a large number of unlabeled training inputs and a small number of labeled training inputs.

[0094] Any of the algorithms described herein may be neural networks. Neural networks may employ multiple layers of operations to predict one or more outputs. Neural networks may include one or more hidden layers situated between an input layer and an output layer. The output of each layer may be used as input to another layer, e.g., the next hidden layer or the output layer. Each layer of a neural network may specify one or more transformation operations to be performed on input to the layer. Such transformation operations may be referred to as neurons. The output of a particular neuron may be a weighted sum of the inputs to the neuron, adjusted with a bias and multiplied by an activation function, e.g., a rectified linear unit (ReLU) or a sigmoid function.

[0095] Training a neural network may involve providing inputs to the untrained neural network to generate predicted outputs, comparing the predicted outputs to expected outputs, and updating the algorithm’ s weights and biases to account for the difference between the predicted outputs and the expected outputs. Specifically, a cost function may be used to calculate a difference between the predicted outputs and the expected outputs. By computing the derivative of the cost function with respect to the weights and biases of the network, the weights and biases may be iteratively adjusted over multiple cycles to minimize the cost function. Trainingmay be complete when the predicted outputs satisfy a convergence condition, e.g., a small magnitude of calculated cost as determined by the cost function.

[0096] As described herein, an optimization module may balance the plurality of machine learning outputs to provide an optimum output per HCP that balances a plurality of long-term and short-term goals. The optimization module may balance these goals by producing an output based on a weighing of several considerations. The output may or may not apply weights to a plurality of quantitative values associated with machine learned predictions relating to sales or communication effectiveness between reps and HCPs, before calculating an average, sum, or product of the quantitative values. The optimization module may have weight categories associated with such predictions, such as urgency, value, content affinity, channel affinity, feasibility, and cost. Following the weighting, the optimization module may implement a set of constraints. The short-term goals maximized by the optimization module may include assisting with access challenges, reacting to HCP-led engagement, sharing new clinical data, or coordinating around an event. The long-term goals optimized for may include targeting and segmentation of HCPs and providing a more engaging customer experience. The optimization module may also be used to rank HCPs by priority to interact, balancing short-term engagement needs with long-term commercial opportunities.

[0097] The optimization module may optimize using an objective function. The objective function may be a product of various terms, including a perceived value of an HCP, an impact of communicating via a particular channel, a probability of a successful sale, a target sales value, a recency of a prior communication, and a frequency of communication. Implementing the objective function may produce a score relating to a value for a particular action taken by a pharma rep to an HCP. Following producing such scores, the optimization module may place constraints restricting actions that may be taken. Constraints may include restrictions on types of actions, types of communication channels time constraints on pharma reps, constraints on contact frequencies, or other constraints.Computer systems

[0098] In another aspect, disclosed herein is a computer-implemented method for providing omnichannel cognition in a campaign, the method comprising: generating and providing a unified and real-time graphical interactive interface displaying an effectiveness of the campaign, wherein the interface is further configured to allow users to dynamically connect strategies to tactics and track key performance indicators (KPIs) while making real-time adjustments and rebalancing of efforts across multiple channels to optimize campaign engagement; utilizing a generative artificial intelligence (GenAI) model for analysis andrestacking of the tactics to generate an optimal tactic mix that improves engagement efficiency and outcomes based on predicted impact levels of the tactics; and simulating the campaign through virtual trials, to assess an effectiveness of the campaign prior to roll-out of the campaign.

[0099] The present disclosure provides computer systems that are programmed to implement methods of the disclosure. FIG. 11 shows a computer system 1100 that is programmed or otherwise configured to machine learning model predictions subject to rules / constraints to optimize long-term and short-term goals. The computer system 1100 may regulate a regulate various aspects of producing suggestions of the present disclosure, such as, for example, performing machine learning analysis. The computer system 1100 may be an electronic device of a user or a computer system that is remotely located with respect to the electronic device. The electronic device of the user may be a mobile electronic device.

[0100] The computer system 1100 includes a central processing unit (CPU, also “processor” and “computer processor” herein) 1105, which may be a single core or multi core processor, or a plurality of processors for parallel processing. The computer system 1100 also includes memory or memory location 1110 (e.g., random-access memory, read-only memory, flash memory), electronic storage unit 1115 (e.g., hard disk), communication interface 1120 (e.g., network adapter) for communicating with one or more other systems, and peripheral devices 1125, such as cache, other memory, data storage and / or electronic display adapters. The memory 1110, storage unit 1115, interface 1120 and peripheral devices 1125 are in communication with the CPU 1105 through a communication bus (solid lines), such as a motherboard. The storage unit 1115 may be a data storage unit (or data repository) for storing data. The computer system 1100 may be operatively coupled to a computer network (“network”) 1130 with the aid of the communication interface 1120. The network 1130 may be the Internet, an internet and / or extranet, or an intranet and / or extranet that is in communication with the Internet. The network 1130 in some cases is a telecommunication and / or data network. The network 1130 may include one or more computer servers, which may enable distributed computing, such as cloud computing. The network 1130, in some cases with the aid of the computer system 1100, may implement a peer-to-peer network, which may enable devices coupled to the computer system 1100 to behave as a client or a server.

[0101] The CPU 1105 may execute a sequence of machine-readable instructions, which may be embodied in a program or software. The instructions may be stored in a memory location, such as the memory 1110. The instructions may be directed to the CPU 1105, which may subsequently program or otherwise configure the CPU 1105 to implement methods of thepresent disclosure. Examples of operations performed by the CPU 1105 may include fetch, decode, execute, and writeback.

[0102] The CPU 1105 may be part of a circuit, such as an integrated circuit. One or more other components of the system 1100 may be included in the circuit. In some cases, the circuit is an application specific integrated circuit (ASIC).

[0103] The storage unit 1115 may store files, such as drivers, libraries and saved programs. The storage unit 1115 may store user data, e.g., user preferences and user programs. The computer system 1100 in some cases may include one or more additional data storage units that are external to the computer system 1100, such as located on a remote server that is in communication with the computer system 1100 through an intranet or the Internet.

[0104] The computer system 1100 may communicate with one or more remote computer systems through the network 1130. For instance, the computer system 1100 may communicate with a remote computer system of a user (e.g., a mobile user device). Examples of remote computer systems include personal computers (e.g., portable PC), slate or tablet PC’s (e.g., Apple® iPad, Samsung® Galaxy Tab), telephones, Smart phones (e.g., Apple® iPhone, Android-enabled device, Blackberry®), or personal digital assistants. The user may access the computer system 1100 via the network 1130.

[0105] Methods as described herein may be implemented by way of machine (e.g., computer processor) executable code stored on an electronic storage location of the computer system 1100, such as, for example, on the memory 1110 or electronic storage unit 1115. The machine executable or machine readable code may be provided in the form of software. During use, the code may be executed by the processor 1105. In some cases, the code may be retrieved from the storage unit 1115 and stored on the memory 1110 for ready access by the processor 1105. In some situations, the electronic storage unit 1115 may be precluded, and machine-executable instructions are stored on memory 1110.

[0106] The code may be pre-compiled and configured for use with a machine having a processer adapted to execute the code, or may be compiled during runtime. The code may be supplied in a programming language that may be selected to enable the code to execute in a pre-compiled or as-compiled fashion.

[0107] Aspects of the systems and methods provided herein, such as the computer system 1100, may be embodied in programming. Various aspects of the technology may be thought of as “products” or “articles of manufacture” typically in the form of machine (or processor) executable code and / or associated data that is carried on or embodied in a type of machine readable medium. Machine-executable code may be stored on an electronic storage unit, suchas memory (e.g., read-only memory, random-access memory, flash memory) or a hard disk. “Storage” type media may include any or all of the tangible memory of the computers, processors or the like, or associated modules thereof, such as various semiconductor memories, tape drives, disk drives and the like, which may provide non-transitory storage at any time for the software programming. All or portions of the software may at times be communicated through the Internet or various other telecommunication networks. Such communications, for example, may enable loading of the software from one computer or processor into another, for example, from a management server or host computer into the computer platform of an application server. Thus, another type of media that may bear the software elements includes optical, electrical and electromagnetic waves, such as used across physical interfaces between local devices, through wired and optical landline networks and over various air-links. The physical elements that carry such waves, such as wired or wireless links, optical links or the like, also may be considered as media bearing the software. As used herein, unless restricted to non-transitory, tangible “storage” media, terms such as computer or machine “readable medium” refer to any medium that participates in providing instructions to a processor for execution.

[0108] Hence, a machine readable medium, such as computer-executable code, may take many forms, including but are not limited to, a tangible storage medium, a carrier wave medium or physical transmission medium. Non-volatile storage media include, for example, optical or magnetic disks, such as any of the storage devices in any computer(s) or the like, such as may be used to implement the databases shown in the drawings. Volatile storage media include dynamic memory, such as main memory of such a computer platform. Tangible transmission media include coaxial cables; copper wire and fiber optics, including the wires that comprise a bus within a computer system. Carrier-wave transmission media may take the form of electric or electromagnetic signals, or acoustic or light waves such as those generated during radio frequency (RF) and infrared (IR) data communications. Common forms of computer-readable media therefore include for example: a floppy disk, a flexible disk, hard disk, magnetic tape, any other magnetic medium, a CD-ROM, DVD or DVD-ROM, any other optical medium, punch cards paper tape, any other physical storage medium with patterns of holes, a RAM, a ROM, a PROM and EPROM, a FLASH-EPROM, any other memory chip or cartridge, a carrier wave transporting data or instructions, cables or links transporting such a carrier wave, or any other medium from which a computer may read programming code and / or data. Many of these forms of computer readable media may be involved in carrying one or more sequences of one or more instructions to a processor for execution.

[0109] The computer system 1100 may include or be in communication with an electronic display 1135 that comprises a user interface (UI) 1140 for providing, for example, suggestions to pharma reps re communications to HCPs and also content. Examples of UI’s include, without limitation, a graphical user interface (GUI) and web-based user interface.

[0110] Methods and systems of the present disclosure may be implemented by way of one or more algorithms. An algorithm may be implemented by way of software upon execution by the central processing unit 1105. The algorithm may, for example, calculate the optimal engagement strategy across multiple channels between reps and HCPs.Terms and Definitions[OHl] Unless otherwise defined, all technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this present disclosure belongs.

[0112] As used herein, the singular forms “a,” “an,” and “the” include plural references unless the context clearly dictates otherwise. Any reference to “or” herein is intended to encompass “and / or” unless otherwise stated.

[0113] As used herein, the term “about” in some cases refers to an amount that is approximately the stated amount.

[0114] As used herein, the term “about” refers to an amount that is near the stated amount by 10%, 5%, or 1%, including increments therein.

[0115] As used herein, the term “about” in reference to a percentage refers to an amount that is greater or less the stated percentage by 10%, 5%, or 1%, including increments therein.

[0116] As used herein, the phrases “at least one”, “one or more”, and “and / or” are open-ended expressions that are both conjunctive and disjunctive in operation. For example, each of the expressions “at least one of A, B and C”, “at least one of A, B, or C”, “one or more of A, B, and C”, “one or more of A, B, or C” and “A, B, and / or C” means A alone, B alone, C alone, A and B together, A and C together, B and C together, or A, B and C together.

[0117] As used herein, whenever the term “at least,” “greater than,” or “greater than or equal to” precedes the first numerical value in a series of two or more numerical values, the term “at least,” “greater than” or “greater than or equal to” applies to each of the numerical values in that series of numerical values. For example, greater than or equal to 1, 2, or 3 is equivalent to greater than or equal to 1, greater than or equal to 2, or greater than or equal to 3.

[0118] As used herein, whenever the term “no more than,” “less than,” or “less than or equal to” precedes the first numerical value in a series of two or more numerical values, the term “no more than,” “less than,” or “less than or equal to” applies to each of the numerical values inthat series of numerical values. For example, less than or equal to 3, 2, or 1 is equivalent to less than or equal to 3, less than or equal to 2, or less than or equal to 1.

[0119] While preferred embodiments of the present invention have been shown and described herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only. It is not intended that the invention be limited by the specific examples provided within the specification. While the invention has been described with reference to the aforementioned specification, the descriptions and illustrations of the embodiments herein are not meant to be construed in a limiting sense. Numerous variations, changes, and substitutions will now occur to those skilled in the art without departing from the invention. Furthermore, it shall be understood that all aspects of the invention are not limited to the specific depictions, configurations or relative proportions set forth herein which depend upon a variety of conditions and variables. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in practicing the invention. It is therefore contemplated that the invention shall also cover any such alternatives, modifications, variations or equivalents. It is intended that the following claims define the scope of the invention and that methods and structures within the scope of these claims and their equivalents be covered thereby.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A platform for providing omnichannel cognition in a campaign, comprising: a strategy hub configured to generate and provide a unified and real-time graphical interactive interface displaying an effectiveness of the campaign, wherein the strategy hub is further configured to allow users to dynamically connect strategies to tactics and track key performance indicators (KPIs) while making real-time adjustments and rebalancing of efforts across multiple channels to optimize campaign engagement; a tactic engine configured to utilize a generative artificial intelligence (GenAI) model for analysis and restacking of the tactics to generate an optimal tactic mix that improves engagement efficiency and outcomes based on predicted impact levels of the tactics; and an impact simulator configured to provide a virtual platform for simulating the campaign through virtual trials, to assess an effectiveness of the campaign prior to rollout of the campaign.

2. The platform of claim 1, wherein the tactic engine is configured to generate triggered journeys based on triggering of events.

3. The platform of claim 2, wherein the triggered journeys comprise an escalated set of triggered journeys, and wherein escalating the set of triggered journeys is based at least on contact channel, degree of urgency, or recommended content.

4. The platform of claim 3, wherein generating the triggered journeys comprises defining triggering rules for the triggering of events and detecting when at least one triggering event occurs.

5. The platform of claim 4, wherein the GenAI model is updated after the at least one triggering event occurs, and wherein the updated GenAI model determines at least one new triggered journey for integrating with the existing triggered journey.

6. The platform of claim 2, wherein the tactic engine is further configured to generate ongoing journeys based on goals of the campaign.

7. The platform of claim 6, wherein the ongoing journeys comprise a continuous set of ongoing journeys, and wherein continuing the set of ongoing journeys is based at least on a contact channel, a degree of urgency, or a recommended content.

8. The platform of claim 7, wherein generating the ongoing journeys comprises defining the goals and detecting when at least one goal is achieved.

9. The platform of claim 8, wherein the GenAI model is updated after the at least one goal is achieved, and wherein the updated GenAI model determines at least one new goal for integrating with the existing ongoing journey.

10. The platform of claim 6, further comprising a contextual intelligence engine (CIE) configured to integrate or combine the triggered journeys and the ongoing journeys.

11. The platform of claim 2, wherein the events are defined by the users using the tactic engine.

12. The platform of claim 11, wherein the events comprise event definitions, and wherein the tactic engine is configured to automatically convert event logic into a set of trigger rules and a historical table for modeling, wherein the historical table comprises past event trigger indicators.

13. The platform of claim 12, wherein the GenAI model is configured to determine a new journey in real time using string triggers comprising the set of trigger rules.

14. The platform of claim 13, wherein the new journey is integrated into an existing triggered journey or an existing ongoing journey.

15. The platform of claim 1, further comprising an attribute model configured to learn attributes of past engagements with HCPs, wherein the learned attributes are based at least on interaction data and engineered features.

16. The platform of claim 15, wherein the engineered features are derived from the interaction data and compressed into a new data structure using at least data compression to reduce computing resources of the platform.

17. The platform of claim 1, wherein the GenAI model comprises a large language model (LLM) configured to use prompt engineering for generating triggered journeys or ongoing journeys of the campaign based at least on a persona emulation.

18. The platform of claim 1, wherein the optimal tactic mix is based on a set of constraints comprising a set of global constraints and set of local constraints.

19. The platform of claim 18, wherein the set of global constraints comprises constraints which apply to a type of actor or a type of target entity.

20. The platform of claim 19, wherein the set of local constraints comprises constraints which apply to an individual actor of the type of actor or an individual entity of the type of target entity.

21. The platform of claim 1, further comprising a reporting module configured to generate a structured report structured by metrics specific for a user for determining the effectiveness of the campaign.

22. A computer-implemented method for providing omnichannel cognition in a campaign, the method comprising: generating and providing a unified and real-time graphical interactive interface displaying an effectiveness of the campaign, wherein the interface is further configured to allow users to dynamically connect strategies to tactics and track key performance indicators (KPIs) while making real-time adjustments and rebalancing of efforts across multiple channels to optimize campaign engagement; utilizing a generative artificial intelligence (GenAI) model for analysis and restacking of the tactics to generate an optimal tactic mix that improves engagement efficiency and outcomes based on predicted impact levels of the tactics; and simulating the campaign through virtual trials, to assess an effectiveness of the campaign prior to roll-out of the campaign.