Guides
12 min readHow to Build Your AI Command Center with Duet
Stop juggling scattered AI tools. Build a centralized AI command center in minutes — one place for your integrations, knowledge base, automations, and custom apps.
Duet Team
How to Cut Rekeying in Carrier Portals with AI
Rekeying is one of the biggest hidden time drains in independent insurance agencies. The same client data gets entered into your AMS, carrier portals, quote forms, emails, and internal notes. AI can reduce this by extracting, structuring, and organizing client and policy information before it ever gets retyped across systems.
Where Rekeying Costs Agencies Time
The drag does not come from one giant task. It comes from the same information being handled repeatedly at different points in the workflow.
The biggest time leaks show up in four places:
| Rekeying Area | What Happens |
|---|---|
| Intake to AMS | Client and risk details pulled from emails, PDFs, or intake forms and entered into the AMS |
| AMS to carrier portals | The same information retyped across multiple carrier portals during quoting |
| Follow-up communications | Emails or notes rewritten manually after each quote, change, or submission step |
| Correction loops | One address, VIN, effective date, or named insured field does not match across systems |
Each step may only take a few minutes on its own. But across new business, remarkets, policy changes, and renewals, the weekly total adds up fast.
The Workflow That Reduces Rekeying
You do not need to rip out your AMS or wait for every carrier to modernize its portal.
What works today is simpler: use AI as the layer that extracts, structures, and reuses information before your team starts typing it into multiple systems.
Step 1: Turn Intake Into a Structured Summary
Convert whatever came in (email notes, PDFs, ACORD forms, declarations pages, or internal intake notes) into one clean summary.
That summary should include:
- Client profile
- Risk details
- Current policies or prior carrier information
- Coverage goals
- Open questions or missing fields
The goal is to stop working from messy source material once the submission process begins. If you handle commercial lines, this step overlaps with ACORD form and quote prep workflows.
Step 2: Create One Normalized Data Packet
Turn that summary into a single "working packet" your team can reference across quoting and submission.
This packet becomes the source you use while entering information into carrier portals, preparing internal notes, or checking follow-up items. Instead of rereading the same documents each time, the team works from one cleaned-up version of the file.
Step 3: Generate Carrier-Specific Entry Checklists
Before anyone starts portal entry, use the same packet to create a short checklist for each carrier.
That checklist can call out:
- Fields that are usually required
- Details commonly missed
- Supporting documents needed
- Questions that still need clarification before submission
This reduces the back-and-forth that happens when a portal flags missing information halfway through the process.
Step 4: Draft Updates From the Same Source
Once quoting is in motion, use that same packet to draft client updates, internal notes, or follow-up emails.
Your team is not rewriting the same account summary from scratch every time there is a quote update, a carrier question, or a next-step email to send. For a deeper look at streamlining those client updates, see how to speed up client emails and renewal follow-ups.
Example Prompt You Can Use Today
Turn the following intake notes into: (1) structured client summary, (2) missing info checklist, (3) carrier-entry packet. Keep output in bullet format and flag assumptions.
Then paste your notes (with sensitive identifiers removed).
Compliance Guardrail
Data handling matters
Use anonymized or sanitized information when working in general-purpose AI tools. Avoid putting SSNs, DOBs, full policy numbers, full claim numbers, full addresses, VINs, or driver's license numbers into a public tool unless your agency has explicitly approved it with the right data-handling terms in place.
A simple rule that works for most agencies:
- Safe: Structuring information, drafting language, creating checklists, building email templates, summarizing redacted documents, writing "what questions should I ask" agendas
- Not safe: Raw sensitive PII, unredacted declarations pages, and claim-identifying details
If you want to start today, start with redacted samples and build the workflow first. Then decide what tools and settings you need to handle live client data responsibly.
What to Measure in Week 1
Do not try to measure everything at once. In week one, focus on three indicators:
- Time from intake to quote-ready packet: How long does it take to go from incoming client information to a clean packet your team can use for quoting and portal entry?
- Number of rework loops per quote: How often does the team have to go back and fix missing, inconsistent, or mismatched information?
- Response time for client updates: How quickly can you send a clear client update after a quote step, carrier response, or request for more information?
If those numbers start moving in the right direction, the workflow is doing its job.
Related Reading
If you are building out AI workflows for your agency, these guides cover adjacent processes:
- Client Emails and Renewal Follow-Ups Faster with AI - Draft client updates, renewal reminders, and follow-ups from a single system
- ACORD Forms and Commercial Quote Prep with AI - Turn scattered intake documents into clean, submission-ready packets
- How to Use AI to Write Sales Emails That Get Replies - AI-assisted email drafting patterns that apply beyond sales
- How to Set Up a 24/7 AI Agent - Scheduling and automation patterns for always-on workflows
- How to Use AI to Run Startup Operations with a 3-Person Team - Operational automation for small teams
Want this running in your own workspace?
Start free in Duet to run persistent agents across your team.
