We built this for ahm.
Not a deck. A working AI member-services agent that handles the two highest-volume support topics for a health insurer — claim status and cover questions, including the retention conversations that follow. The chat reads from an ahm-shaped knowledge base and the workflows match how a thoughtful member-services rep should actually answer.
The voice version of this agent runs on the same workflows and KB. In production it answers the ahm member-services line — looks up cover by phone, walks members through waiting periods, reads claim references digit by digit.
How we built this
Honest version. Here's what's behind this chat:
Wrote ahm-shaped knowledge
Five hero articles covering how extras claims work, hospital cover tiers (Basic, Bronze, Silver, Gold), waiting periods by service, policy changes (downgrade, suspend, add partner), and the limits of what support can help with.
Built workflows around your shape
A router that splits claims questions from cover and policy questions, plus two subworkflows. The retention path acknowledges, summarises, and offers suspension or a downgrade — never pressures the member.
Connected mock ahm tools
Three mock tools — member policy lookup, recent claims, submit policy change — that mirror the real APIs we'd integrate with on day one.
Went live and ran the regression
Ten simulated conversations covering claim diagnoses, cover lookups, downgrades, retention conversations, clinical refusals and knowledge gaps. See the simulation report →
Want to build the real one?
This took about ninety minutes with one engineer. Production with your real claims engine, policy administration system, and retention playbook takes a few weeks. We've shipped agents for regulated finance and healthcare subscribers across AU.
Discuss with our experts →
Demo built with