Home / Case study

Case study · one insurance claims firm

Real work, real hours back

One insurance claims firm, three AI systems, and the hours their team got back. The client stays private. Everything on this page is real, and nothing is rounded up.

Book an AI workflow audit See the three systems
  1. The problem

    Staff chased missing documents by email, one party at a time, on every claim.

  2. What we built

    An agent that asks for what's missing, checks what comes back and follows up. Follow one claim down the belt.

  3. The result

    10–20 hours back every week. Staff still approve every reply.

Real result · one insurance claims firm

Insurance claims: missing documents

Before, someone on the team did every one of these steps by hand.

  1. New claim (Your systems)
  2. Spots what's missing (AI does it)
  3. Emails the client (AI does it)
  4. Checks replies (AI does it)
  5. Team reviews (Your team checks)
  6. 10–20 hrsback every week, from one agent
  • AI does it
  • Your team checks
  • Your systems
How the agent was built
  • Running today10–20 hrsMissing-documents agentback a week for the teamHow it works
  • Running todayAbout 10 minAI file reviewsaved a case, on about 100 cases a monthHow it works
  • In progressAbout 3 minIntake to case filesaved per new client, about 3 a dayHow it works
An insurance policy under a magnifying glass
Every claim waits on documents from someone.

The firm and the problem

Our client is an insurance claims firm in South Florida. Every claim waits on documents from someone: the homeowner, the adjuster, the contractor, the carrier.

Their team spent a big part of each week on work around the claim, not the claim itself. They chased missing documents by email, read whole files before they could start, and retyped intake forms into their case system.

None of that needed their judgment. All of it took their time. So we started where we always start: with an audit of how the work really moved.

Three systems, one firm

Each one took a job off the team's plate. Each one still asks a person when it should.

Running today

Missing-documents agent

10–20 hoursback every week

The problem

Every open claim was missing something. Staff emailed each party one at a time, opened every reply to check it, and wrote back with what was still needed. Then they did it again a few days later.

What we built

An AI agent that knows which documents each claim needs, sorted into critical and nice-to-have. It asks every party for what's missing, checks what comes back, and replies with exactly what's still needed. Anything unusual goes to staff.

BeforeStaff chased documents by email, one party at a time, and checked every attachment by hand.
AfterThe agent asks every party, checks what arrives and follows up. Staff step in only when something's unusual.

Running today

AI file review

About 10 minutessaved a case, on about 100 cases a month. About 16 hours a month.

The problem

Before anyone could work a case, they had to read everything in it: policies, estimates, letters, photos and notes. On a busy month, that's a lot of reading before any real work starts.

What we built

AI that reads every document in a case, including scans, and writes a short summary: who's involved, what's been received, what's still open, and anything that needs a closer look.

BeforeStaff read the whole file before they could start working the case.
AfterA short summary is waiting when they open the case. They read the documents that matter.

In progress

Intake to case file

About 3 minutessaved on every new client, about 3 new clients a day

The problem

Every new client filled out an intake form. Then someone typed the same details into the case system by hand, which is slow and easy to get wrong.

What we're building

When a client sends the intake form, the case is created in the case system with the details in the right fields, and staff are told it's ready. This one is still being finished.

BeforeEach intake form was retyped into the case system.
AfterThe case is created the moment the form comes in.

How it was built

The same steps every client goes through. Each one was priced before it started.

A fountain pen resting on an open notebook marked Notes
It started with a written plan.

Audit

Where the hours went

We sat with the people doing the work, mapped how claims moved and counted the hours. The written plan ranked document chasing first.

Build

One system at a time

We built the missing-documents agent first and tested it on real files with the team. Then we added file review.

Run

Watched and improved

We watch the systems, fix what breaks and keep improving them. Intake to case file is the next piece being finished.

Keep going

The client decides

As the firm changes, the plan gets updated. The client decides what's worth building next.

What changed for the team

The numbers matter, but this is what the team noticed first.

  • More time for judgment

    Hours that went to chasing and retyping now go to the claim decisions only people can make.

  • Fewer chases

    Nobody has to remember who still owes what. The agent keeps track and follows up.

  • Every file checked the same way

    The same checks run on every claim, so less slips through on a busy day.

  • People stay in charge

    Anything unusual goes to staff. The AI handles the routine and asks when it should.

More case studies are on the way. We publish them only when clients say yes, and we never share a client's name without permission. See what's possible for your field

Where are your team's hours going?

The audit finds out, and prices what AI could take off their plate. From $500.

Book an AI workflow audit

Not ready? Get the free “Is your firm ready for AI?” checklist

Find out where AI fits in your firm

Start with a free call, or go straight to the audit. Either way, you talk to the person who builds it.

Not sure yet?

Book a free 20-minute call

Tell us what's eating your team's time. We'll tell you honestly if AI can help. No pitch, no obligation.

Book a free call

Ready?

Book your AI workflow audit, from $500

1–2 weeks, ending in a written plan ranked by hours saved, with a price for each piece.

Book your audit

Today, every call and every audit is led by our founder, Aramis M. Figueroa. Or call (305) 807-7015.

Not ready yet? Get the free “Is your firm ready for AI?” checklist

A 10-question checklist you can fill in over coffee. It shows where your team's hours go and which work AI could take on first.

One email with the checklist. No newsletter unless you ask for it. See our privacy policy.

Book a free 20-minute call