How a Top US Pharma Company Used Agents for Analytics to Cut Sales Reporting for its Blockbuster Drug from 10 Days to 4 Hours
Sales Analytics
Agentic AI
Databricks, OpenAI, Agentic AI, mlflow
The client wanted AI agents to run pre-defined business hypotheses on the sales data of its blockbuster drug and build the sales analytics reports automatically
Key Challenges
1
Strict Architecture Governance
The AI architecture had to be approved by the client's Architecture Review Board and comply with its sophisticated guidelines and guardrails.
2
No Off-the-Shelf Frameworks
The guardrails ruled out known agent frameworks, so every agent had to be written from scratch as a custom atomic agent.
3
Zero Data Visibility for AI
Sales data had to be pulled from Databricks, but the AI could never view it. Agents were only allowed to generate SQL.
Solution accelerators
Accelerated by our solutions
a21.GenBI
Agentic Analytics — ask your data anything, get answers in seconds.