Michael Experience Design
Case study 007 · Agentic AI in Health Insurance

Simplify insurance, keep oversight

Industry
Healthcare
Role
Experience Design Lead
Company
Publicis Sapient
Year
2025
A stethoscope over a white coat
01 / Overview

Every decision matters to someone waiting for coverage

Thousands of insurance applications pass through slow, manual checks every day. We prototyped an AI assistant that speeds the work up while people stay in charge.

02 / Challenge

Three questions shaped every design decision

A

How do you bring AI into a regulated, trust-sensitive workflow?

B

How do you keep people in control and still save time?

C

How do you make AI reasoning clear enough to trust?

03 / Process
  1. 01StrategyWorked with strategists, developers and product experts.
  2. 02Visual designA clear look and feel with strong hierarchy.
  3. 03PrototypingAn interactive prototype of the review, approve, override loop.
04 / How control works

Delegate, explain, decide, recover

Delegate

The agent does the checking

It processes applications and sorts them into a queue by team and issue.

Explain

Every recommendation shows its reasoning

Reviewers see why the AI reached its recommendation.

Decide

A person makes the call

Review, approve or override, every time.

Recover

When the AI can’t process a case, it says so

Those cases are flagged for a person.

Member administration work queue shown on a laptop
Fig. 01 — The work queue, grouped by team and issue, with a clear status for each
Prototype

Simple flow

Pick an AI-processed application, review its reasoning, then approve or deny.

View prototype (opens in a new tab)
Work queue screen from the prototype
05 / Outcome

Agentic AI can simplify insurance work while people keep oversight.