Pharma CRM Pre-Call AI Agent

I’m developing an AI-native CRM specifically for pharmaceutical and life-science companies and now need the core embedded agent that handles pre-call planning for our sales, marketing, and medical teams.

The assignment
When a user selects an HCP, the agent must instantly pull that provider’s complete history from our unified customer database, surface key insights, and generate role-specific talking points that respect every compliance rule we live under. Think of it as a smart briefing document that appears in seconds, ready to guide the next field visit, virtual detail, or MSL discussion.

What I’ll count as success
• Architecture and data-flow design showing how the agent interfaces with our existing lake/warehouse and respects HIPAA, GDPR, and 21 CFR Part 11.
• A working service (API or micro-service) that retrieves data, runs the LLM/RAG pipeline, and returns clear, evidence-backed talking points in under two seconds.
• Lightweight front-end component or sample call that renders the output inside the CRM workspace.
• Evaluation script with metrics for factual accuracy, hallucination rate, and compliance wording.

Useful context
We already house HCP interactions, content engagement, sample data, and field notes in a single Snowflake cluster and expose them through REST and GraphQL. You’re free to choose the stack—Python, Typescript, LangChain, Azure OpenAI, or similar—as long as it’s containerised and cloud-agnostic.

If you have deep experience with LLMs, retrieval-augmented generation, and building regulated-industry software, let’s talk timelines and milestones so we can move this agent from concept to production quickly.