All case studies

Case studies: insurance agencies & brokers

Every missed call is a missed policy. We help you answer them.

We set up an AI voice receptionist that answers every inbound call, knows P&C terminology, qualifies the caller, and writes the call notes into Applied Epic, EZLynx or your AMS. It works even without an API. One broker went from a 12% to a 100% answer rate.
Book a 30-min workflow reviewGet a free 90-day AI plan

Inbound calls answered - BIG Pickering

  • Call answer rate12% → 100%
    BeforeBefore: 12%
    AfterAfter: 100%
Third-party published result.

Source: Sonant AI

  • 58+ hrs

    Saved per month

    O'Connor Insurance

  • 8×

    ROI in 30 days

    O'Connor Insurance

  • 7.5 hrs

    Saved per staff per week

    Best Practices study

  • 6-12 wks

    Typical first deployment

    ATI Lab delivery

Case studies

Brokerages that fixed missed calls and manual AMS work

Broker cases first, then two carrier examples for scale. Every number links to its source.

Third-party results from other companies, linked to source. They are not ATI Lab clients. How we would build it for you is below. See ATI Lab's own client builds.

Brokerages

O'Connor Insurance Associates

P&C broker - USAThird-party result

The bottleneck

Missing over 30% of inbound calls. Every missed call is a missed policy. No scalable way to handle volume without adding headcount.

What fixed it

AI voice receptionist (Sonant AI) handling inbound calls with real-time AMS data sync. Understands P&C terminology out of the box.

Published result

  • 8× ROI

    within 30 days of deployment

  • 58+ hrs

    saved per month

  • 100%

    call answer rate achieved

Vendor-published case study.

Source: Sonant AI case study (opens source in a new tab) - cited in "100+ AI Tools for Insurance Agencies: 2026 Guide"

BIG Pickering Insurance

Insurance brokerThird-party result

The bottleneck

A 12% inbound call answer rate. Nearly 9 out of 10 potential clients hit voicemail and never called back.

What fixed it

An AI voice agent answers every inbound call, qualifies the lead, and transfers to the right producer with a summary already on screen.

Published result

  • 600%

    ROI in the first month

  • 12% → 100%(increase)

    inbound calls answered, up from 12%

Vendor-published case study.

Source: Sonant AI case study (opens source in a new tab)

Best Practices Agency (Reagan Consulting Study)

$100M+ brokerage - USAThird-party result

The bottleneck

Producers spent hours each week reading loss runs, comparing coverages and checking Statements of Value by hand.

What fixed it

AI platform deployed across 4 functions: loss run analysis, coverage analysis, SOV comparison, and an "Ask Anything" agent for instant answers.

Published result

  • 7.5 hrs

    saved per staff per week

  • 244%

    return on investment

  • $168K

    larger book size for tech-enabled producers under 35

Hours saved per staff member, in a 40-hour week

  • Working week40 hrs
  • Saved with the AI platform7.5 hrs
7.5 hours is about one working day back each week. Third-party published result.

The four functions the platform covers

  • Loss run analysis (done)
  • Coverage analysis (done)
  • SOV comparison (done)
  • "Ask Anything" agent (done)

Source: Risk & Insurance (opens source in a new tab) - "Leading Insurance Brokerages Embrace AI Revolution" (Aug 2025), citing the 2025 Best Practices Study

Carrier examples

Aviva (UK)

Carrier example - claimsEnterprise referenceThird-party result

The bottleneck

Complex motor claims taking weeks. High error rates in routing claims to the right teams. Customer complaints piling up.

What fixed it

AI models deployed across the claims domain: liability assessment, routing, and complaint reduction.

Published result

  • £60M+

    saved in 2024 ($82M USD)

  • −23 days(decrease)

    cut from liability assessment time

  • +30%(increase)

    claims routing accuracy

  • −65%(decrease)

    customer complaints

Customer complaints, indexed (before = 100)

  • Complaints65% fewer
    BeforeBefore: 100
    AfterAfter: 35
Third-party published result.

Source: McKinsey & Company (opens source in a new tab) - "The future of AI for the insurance industry" (Jul 2025), citing Aviva investor reports

NIB Health Insurance (Australia)

Carrier example - health insurerThird-party result

The bottleneck

High volume of routine customer inquiries consuming agent time. Licensed staff tied up answering repetitive questions instead of advising.

What fixed it

AI digital assistant "Nibby" deployed to handle routine inquiries autonomously, with handoff to humans for complex cases.

Published result

  • $22M

    total savings

  • −60%(decrease)

    human support need

  • −15%(decrease)

    phone calls to agents

Indexed (before = 100)

  • Human support need60% less
    BeforeBefore: 100
    AfterAfter: 40
  • Phone calls to agents15% fewer
    BeforeBefore: 100
    AfterAfter: 85
Third-party published result.

Source: FactR / DataPeak (opens source in a new tab) - referenced as industry comparison in AI cost savings analysis

How we'd build it for you

What we build for agencies

  • The pain: Calls go to voicemail and never call back

    What we build: AI voice receptionist with P&C vocabulary, lead qualification and warm transfer

    Published proof: 12%→100% answer rate, 600% first-month ROI (BIG Pickering)

  • The pain: CSRs re-key every call and quote into the AMS

    What we build: AMS write-back via API or browser automation (Epic, EZLynx)

    Published proof: 58+ hrs/month saved (O'Connor)

  • The pain: Loss runs and SOV comparisons eat producer time

    What we build: Document-analysis agent for loss runs, coverage and SOV comparison

    Published proof: 7.5 hrs/staff/week, 244% ROI (Best Practices study)

Before: a missed call

  1. Call comes in

  2. Rings out to voicemail

  3. Caller never calls back

    Lost policy

    Exception

After: what we set up

  1. Call comes in

  2. AI receptionist answers

    P&C vocabulary

  3. Qualifies the caller

  4. Warm transfer

    Producer sees the summary

    Human approval

  5. Notes written to the AMS

    Applied Epic / EZLynx (API or browser automation)

Industry benchmarks

84% of large brokerages already invest in GenAI. Here's where it pays.

Where insurer costs sit

of insurer expenses are in claims processing

Source: Deloitte research. Third-party published result.

Processing cost reduction from AI-automated workflows

  • Cost reduction30–50%
Source: Databricks / industry benchmarks. Third-party published result.

  • 84%

    of brokerages over $100M revenue have invested in GenAI

    2025 Best Practices Study via Risk & Insurance (opens source in a new tab)

  • 30-50%

    processing cost reduction from AI-automated workflows

    Databricks / industry benchmarks

  • 70%

    of insurer expenses are in claims processing alone

    Deloitte research

  • Days → Seconds

    claims processing time with AI triaging

    Databricks insurance industry analysis

Next step

Your legacy AMS doesn't have an API? That's our specialty.

Start with phones or start with the AMS. We set up call answering with AMS write-back, or automate loss-run and SOV comparisons, on Applied Epic, EZLynx or whatever you run, using browser automation where there's no API. Typical first deployment: 6-12 weeks.

From the phone line to your AMS

Calls go from the phone line into the ATI Lab automation, which writes to Applied Epic, EZLynx or your AMS through the API where there is one and browser automation where there isn't.. Phone line (Every inbound call) feed into ATI Lab automation (Answer, qualify, transfer, log), which updates Applied Epic (API or browser automation), EZLynx (API or browser automation), Your AMS (Browser automation where there's no API).

Fix my missed callsInsurance broker tasks we automate