
A US healthcare-support company helps Medicare-age clients get brand-name medications through manufacturer assistance programs. Every case involves the client, a broker, the doctor's office, the drug maker's program and an internal team.
The starting point
Follow-ups ran on memory, spreadsheets and phone calls
Client history was split across Salesforce, DocuSign, inboxes and phones
Brokers called just to ask for a status
What was built
One custom CRM and data model with five portals: brokers, clients, customer service, accounting and owners
Program forms generated from CRM data, pre-filled, sent by email, SMS or fax and tracked until signed
An AI layer (LLM with retrieval) that turns every call, text and email into a structured case note and routes the case, with people reviewing edge cases
Eval sets from real cases and a regression gate before every release
How it was run
Discovery with every party to map how a case really moves
One prioritised roadmap, user stories and acceptance criteria
Staged rollout, one tool replaced at a time, with training per team; HIPAA rules designed in from day one
Role: AI product owner and delivery lead, August 2022 to present. Full case study: hireian.today/case-studies/medicare-medication-access-crm/
0 likes · 0 views · Nov 10, 2026
A US Medicare operation moved to one CRM with five portals and an AI layer that writes case notes.