Workflow mapping

Where can AI lighten the load?

Before designing screens, I mapped how a deal actually moves through arqu, then redrew it around the agent. The goal was concrete: cut the broker’s touchpoints in a deal from 18 to 5.

Target state

End-to-end deal lifecycle.

This is the lifecycle of a deal: kickoff & strategy, clearing, submission & quoting, binding. The product thesis comes through clearly: AI does the heavy-lifting, the human keeps the relationships and the final say.

AI carries the effort: classifications, extractions, creating objects and assets.

The broker (RS) sits at the review and approval gates.

Retailer actions remain untouched.

Carrier actions remain untouched.

New Flow with AI Existing flow — work done manually by broker

To-do List

Broker workflows: the new reality.

To translate the above into an actionable plan, I documented every task a broker should be able to do in our platform, and worked out the flow for each — including the happy path where “AI did its job” as well as “AI missed this” and “I want to do this myself”. This served as our quasi-requirements doc throughout the project.