Pacific Life: AI‑Assisted Review Analysis

AI assisted underwriting

Pacific Life, a leading life insurance underwriter, who partnered with Magnifi to explore how AI could reduce the effort involved in policy reviews while keeping final decisions with people. We designed and prototyped an AI-assisted exception review experience that was demonstrated live to stakeholders during an active client pitch.

The goal was to replace a fragmented, manual process with a workflow that surfaces AI recommendations clearly, while keeping reviewers fully in control of every decision.

My Role

I was brought in to support a high-priority enterprise engagement for Pacific Life, a leading life insurance underwriter. Given the urgency and business impact of the deal, the work required rapid design refinement, interaction clarity, and a fully clickable prototype within a tight deadline.

To meet this pace, I integrated AI into my workflow, using structured prompting to quickly generate and refine high fidelity UI frames. I then connected the flows manually in Figma, shaping it into a complete, end to end prototype.

[View the prototype →]

Exception Categories

Designed to provide immediate clarity before review begins. AI status, exception count, and time savings are context upfront, while category-based grouping helps users quickly scope work and decide where to start. Progress and agent activity reinforce transparency and confidence throughout the workflow.

Review Exceptions

Clicking into an exception shows you the value pulled from each source system side by side, so you can see exactly where things don’t match. Below that, the AI lays out its recommendation with a few short reasons. You approve, modify, or reject from there, and the audit trail below tracks everything automatically.

Add New Exceptions

If something’s missing from the AI’s list, reviewers can add it manually through a three-step modal. You search for the field, enter the new value and its source, and before confirming, the system shows you exactly what the change affects — which system gets updated and whether an amendment is needed.

Review Summary

Once all exceptions have been reviewed, the summary page lays out every approved and rejected decision in one place, with a clear view of what’s changing and where. The systems to be updated are listed alongside an execution checklist, so nothing gets missed before submitting.