#FutureReadyHealthcare
Share this blog
29 Aug 2024
Customer data platforms (CDPs) play an essential role in consolidating and organizing vast amounts of customer information. These platforms successfully address the needs of individuals working in data or data-associated roles, making Customer 360 data and insights readily available to them. However, a critical gap persists when it comes to extending this accessibility to professionals in non-data roles, such as those in sales and support.
The emerging need for a Gen AI CDP
In our client engagements across several life sciences organizations, a recurring challenge emerges, which is how to effectively unlock the true potential ofcustomer 360 data for personalized engagement.
Traditional CDPs hold a treasure trove of information on healthcare professionals (HCPs) and patients; however, navigating through this data and extracting actionable insights can be a cumbersome task. The challenge lies in bridging the comprehension gap for individuals who may not have an inherent affinity for data intricacies. Even for those familiar with the nuances of data, navigating the expansive landscape of information generated by customer 360 can be overwhelming, often requiring a deep dive into technicalities. This complexity results in the integration of customer 360 with the workflows of sales and support teams becoming a cumbersome and time-consuming process, leading to a reliance on the expertise of individuals directly involved in the CDP project. This frustration has fueled the demand for a more user-friendly and intelligent solution.
Large language models (LLMs) and their associated frameworks offer a promising solution to this predicament. These advanced language models have the potential to revolutionize the accessibility of customer 360, making it more user-friendly for professionals in nondata roles. This vision is realized through the concept of the Generative Artificial Intelligence Customer Data Platform (Gen AI CDP), a solution that combines a traditional CDP with generative AI to unlock the true potential of customer 360 data. It can empower users across different business roles to gain a deeper understanding of customers as well as make data-driven decisions through a user-friendly interface and AI-powered insights.
Gen AI CDPs can analyze a HCP’s historical prescription data and identify their preferred medications and therapeutic areas. This allows sales reps to tailor conversations around the HCP’s specific patient population and needs. Through a user-friendly interface (such as a chatbot or search bar), sales reps can receive real-time insights on recent HCP interactions, research interests, and advertisements they’ve engaged with.
And the power of Gen AI CDPs can extend beyond sales teams. Medical science liaisons (MSLs) can leverage these platforms to tailor scientific presentations to individual HCPs’ research focus areas. Additionally, patient adherence programs can be personalized on the basis of real-time insights into patient behavior and treatment needs.
By bridging the data gap and enabling personalized engagement, Gen AI CDPs can empower life sciences organizations to deliver the right message to the right person at the right time. This translates to:
Benefits for the Internal Stakeholders: Increased sales productivity, enhanced MSL effectiveness, and data-driven decision-making across the organization
Benefits for the Customers: Improved patient care through targeted medication selection and a more positive HCP experience, fostering trust and collaboration
Let’s look at how this transformative solution would work.
Democratizing customer 360: Natural language interaction with a Gen AI CDP
The Gen AI CDP approach offers a holistic and streamlined solution to unleash the true potential of customer 360 data across diverse business roles, fostering effective communication and comprehension. Assuming that the CDP serves as the foundational source of customer 360, let’s dive deeper into the intricacies of each component of the solution and explore how they seamlessly come together to form a cohesive and accessible system.
Interface for user interaction : At the forefront of this innovative solution is a user-friendly interface designed to accept prompts from users and provide intuitive responses. This could manifest as a conversational chatbot interface or an integrated search bar within familiar tools found in customer relationship management (CRM) systems. This interface acts as the gateway for users, offering a natural and accessible means of communication with the CDP.
Customer 360 vector database : The backbone of this approach is the customer 360 vector database, derived from the comprehensive real-time interaction and user-level data stored in the CDP and potentially supplemented by additional relevant datasets. This database serves as the reservoir of customer 360 information and becomes the focal point for user queries. Through the integration of data across key channels, it provides a comprehensive and nuanced perspective, allowing users to extract valuable insights without grappling with the intricacies of raw data.
Vector search with retrieval-augmented generation or framework (RAG) : Central to the user interaction process is the vector search, leveraging the powerful RAG. This RAG-based pipeline plays a crucial role in enhancing the understanding of semantics in user prompts. By augmenting the original prompt with a customer 360 vector database, RAG ensures that queries are refined and contextually enriched, paving the way for more accurate and relevant search results. Using RAG enables the ability to augment the context of LLM. Finetuning, which is the process of further training a pretrained LLM on a specific dataset, is another approach to optimization. However, RAG and fine-tuning address different aspects to improve the model’s performance. If there is a need for flexibility in data sources and for valuing explainability, RAG becomes the preferred choice. Fine-tuning would be preferred for working with stable, labeled data, where the needs are extremely specific. Therefore, for use cases that center around customer 360 accessibility, creating a RAG pipeline becomes our go-to choice.
LLM for generative responses: : Adding a layer of intelligence to the system is the inclusion of a LLM. This LLM functions as a generative force, producing contextually relevant responses based on the information retrieved from user queries and augmented by the Customer 360 vector database. Its capability to understand and generate human-like responses ensures that users receive meaningful insights, fostering a more natural and engaging interaction with the data.
Let’s look at how it all comes together:
The user initiates a query through the user-friendly interface, which is then processed by the vector search pipeline, leveraging RAG to refine and augment the original prompt. This enhanced query is directed to the LLM. The LLM, acting as an intelligent assistant, utilizes the customer 360 information retrieved to generate insightful and contextually relevant responses. The seamless integration of these components ensures a fluid and intuitive user experience, democratizing access to customer 360 insights across diverse business roles.
Securing pharma insights: HIPAA and beyond in Gen AI CDPs
The Gen AI CDP prioritizes data security and compliance for the life sciences industry through HIPAA (Health Insurance Portability and Accountability Act)-compliant encryption, which safeguards protected/personal health information (PHI) and HCP information. Granular access controls adhering to HIPAA’s principle of least privilege minimize unauthorized access. Audit trails track user activity for breach investigations. The CDP can be configured to support further HIPAA compliance through de-identification techniques and data minimization. While the FDA doesn’t directly regulates marketing, integration with content management systems (CMSs) can ensure ethical and truthful promotion to HCPs.
The future is now: Salesforce paves the way for conversational CDPs
On June 6, 2024, Salesforce announced the general availability of the Data Cloud Vector Database for making unstructured and structured data accessible through Gen AI in its CDP platform: Salesforce Data Cloud . This approach leverages the power of the Einstein 1 platform to empower organizations to unlock the power of data cloud.
Here’s a breakdown of the process:
Data collection : Data Cloud gathers customer information from various sources, including structured formats like Excel files and unstructured formats like audio files.
Vector database creation: An AI model analyzes the collected data and generates vectors, essentially creating a searchable database of both structured and unstructured information.
Einstein copilot: This Salesforce product utilizes user-submitted requests and performs semantic searches within the Data Cloud Vector Database.
RAG pipeline: Based on the retrieved results, the RAG pipeline generates an augmented prompt. This prompt can be directed to various AI models, depending on the specific needs of the search.
Model selection: The augmented prompt can be directed to models hosted within the secure Salesforce Trust Boundary, external models hosted on the client’s side, or models with a shared trust boundary.
Answer generation: The chosen model generates answers relevant to the user’s initial query.
This AI-powered approach within the Salesforce Data Cloud Vector Database allows for a more comprehensive understanding of customer data, facilitating the retrieval of valuable insights from diverse sources.
Conclusion
The integration of LLMs with CDPs presents a transformative opportunity for life sciences organizations. While CDPs have excelled at aggregating and managing comprehensive customer data, the translation of this information into actionable insights through a conversational interface can unlock the full potential of customer 360.
Imagine a scenario where a life sciences organization leverages the Gen AI CDP solution to identify HCPs who are most likely to adopt a new therapy based on their prescribing patterns and patient demographics. Armed with these insights, targeted marketing campaigns could be developed to reach these HCPs effectively. On the patient front, the Gen AI CDP solution could be used to analyze patient journeys to pinpoint areas where support or education is most needed. This enables the creation of personalized patient programs, such as disease management tools or adherence support, ultimately improving patient outcomes and increasing treatment satisfaction.
Salesforce’s pioneering work demonstrates the feasibility of this approach. By adopting a similar strategy, pharma organizations can create a scalable and secure framework that empowers a broader range of professionals to leverage the power of customer data. This, in turn, fosters a culture of data-driven decision-making, leading to increased efficiency, reduced costs, and ultimately, a greater impact on the pharma industry by realizing the promise of precision marketing in patient-centric care. Are you looking to better leverage your CDP? Write to usfor a strategic consultation today!
Share this blog
Get exclusive pharma
insights delivered to your inbox
Latest
Latest
