Challenge
A regional commercial insurance brokerage had already approved AI usage in parts of the team, but the work still lived in personal habits. Producers, account managers, and operations staff were using AI privately to summarize renewal emails or draft follow-ups, yet the renewal process itself still restarted from the inbox every time an account changed hands.
Renewal packets, policy documents, endorsements, client notes, and long email threads all had to be reread by whoever touched the account next. Context was rebuilt too often, handoffs were inconsistent, and the quality of internal summaries depended more on individual discipline than on any shared operating model.
What We Implemented
We did not treat the problem as prompt quality. We treated it as workspace design.
We rebuilt the renewal flow around a shared AI workspace so the team had one repeatable way to prepare, review, and hand off account context:
- a shared workspace structure for renewal prep, account summaries, and draft client follow-up
- repeatable agent roles for renewal briefs, missing-item checks, account-history summaries, and handoff notes
- approved input boundaries around policy documents, endorsements, renewal history, internal notes, and relevant email threads
- review boundaries so client-facing drafts and coverage-sensitive language still stayed with the account team before anything went out
- a staged rollout that started with renewal-heavy accounts and then expanded once the handoff pattern held
Outcomes
The team stopped rebuilding the same account from scratch every time work changed hands. Renewal prep became easier to start, missing items surfaced earlier, and internal handoffs became more usable because the next person inherited a clearer file instead of a pile of email and memory.
The brokerage also moved AI usage out of private experimentation. Instead of each person using the tools differently, the team had a shared layer for how renewal work should be read, summarized, and passed forward.
Why It Worked
The improvement did not come from asking people to use AI more often. It came from taking the renewal mess off their plate and turning it into one clearer system.
Once the workspace held the structure, the inputs, and the review boundaries, the team no longer had to rely on personal memory to keep renewal work moving.