Skip to main content

Should Your Insurance Agency Use AI? Yes, but Keep the Claims Call Human

Business partners reviewing an AI-enabled operating plan

An insurance agency should start with AI in the back office: claims intake and routing, renewal prep, quotes, data entry, and internal search across policy manuals and underwriting guidelines. Keep the claims call, coverage advice and relationship check-ins human. AI should make your people better prepared for those moments, not stand in for them.

If you run an agency or sit in operations or leadership at a small-to-mid-size carrier, you're getting pitched on AI right now. A lot. The pitch is usually some version of faster claims, lower overhead, better customer experience and more capacity per agent.

What the pitch leaves out is the thing that makes insurance work: it's a trust business. Clients buy a policy and then mostly forget about it, until something goes wrong. At that moment, the experience they have with your agency decides whether they stay, leave, or tell people about it. Any AI strategy that doesn't start with that reality will create problems the vendor won't be around to clean up.


Where does AI help an insurance agency?

Claims intake and routing. The moment a claim comes in is high-volume and high-friction: lots of forms, routing decisions and data entry. AI can handle the triage work. It extracts information from intake documents, routes claims to the right adjuster or team, flags incomplete submissions, and builds the paper trail everyone needs downstream. It isn't making claims decisions. It's getting the right information to the right place quickly, without transcription errors.

Internal knowledge search. Insurance products are complex, and keeping agents current on policy details, coverage comparisons and underwriting guidelines is a constant challenge. AI search across your own documents (policy manuals, coverage summaries, training materials, past correspondence) is one of the most underrated uses in this industry. It puts nothing client-facing in front of an AI. It just makes it faster for an agent to find accurate information. For agencies in Cedar Rapids and across Iowa bringing on new producers or cross-training staff, it's useful without much risk.

Renewal preparation. Renewals are predictable, high-volume work: pulling client history, flagging coverage gaps, preparing documents for the agent's conversation. Automating the prep doesn't replace that conversation. It makes the agent better prepared to have it.

Back-office data work. Quote preparation, data entry, client record updates and audit documentation are low-judgment, high-effort tasks, and the ones most likely to pick up errors from manual handling. Automating them cuts errors and frees staff for work that needs judgment. This is where most agencies should start.

TaskAI's jobThe person's job
Claims intakeExtract, route, flag missing itemsTalk to the client, own the claim
RenewalsPull history, flag possible gapsHave the renewal conversation
Quotes and data entryPrepare and fill inCheck and send
Coverage questionsSummarize current coverage for the agentMake and own the recommendation
Policy and guideline lookupsFind the right passage fastApply it to the client in front of them

Where do insurance agencies get burned with AI?

Automating client communication during claims. A client who just had a car accident or a flood doesn't want a bot. They're stressed, they have questions, and they need to feel that someone is handling their situation. Replacing that touchpoint, even with a good system, tells them they're a transaction. I've seen this pattern in other professional services industries, and it almost always costs more in trust than it saves in agent time. The claims moment is exactly where the human relationship is worth the most. We wrote more about this in when AI isn't the answer.

Coverage recommendations without human review. Some vendors pitch AI that helps clients understand coverage, compare options or spot gaps. In theory, useful. In practice, coverage recommendations carry real compliance and liability exposure. Requirements vary by state, the errors-and-omissions implications are serious, and a confident but wrong AI recommendation lands on your agency, not the vendor. Any AI output that touches coverage advice needs a real review. Not a checkbox, and not an override buried in a workflow.

Replacing relationship touchpoints because the math looks good. The operational math on replacing agent calls or check-ins with automated outreach usually pencils out: labor avoided, faster response, more coverage. What doesn't show up is what happens to retention when clients realize a system is managing them instead of a person. In a business built on long relationships, the kind Iowa independent agencies and regional insurers depend on, that's a slow leak you may not notice until the damage is done.


What do Iowa regulators expect from insurers using AI?

In December 2023 the NAIC adopted a model bulletin on insurers' use of AI systems, and Iowa's Insurance Division issued its own version, Bulletin 24-04, in November 2024. The core message is that existing law still applies: insurers using AI must comply with Iowa insurance law, including the rules on unfair trade practices and unfair discrimination, whether a decision is made by a person or supported by a model. The model bulletin also expects insurers to have a written program governing how they use AI, including AI from third-party vendors.

That bulletin is addressed to insurers, but it reaches agencies too. Carriers will ask how the agencies they work with use AI, and an agency that can answer clearly is easier to work with.

Policy terms, coverage interpretations, claims handling and disclosure obligations aren't areas where "good enough most of the time" is acceptable. So any AI that touches policy information, claims documentation or client-facing content should route through a person before it matters, as a designed step in the workflow rather than something that happens if someone remembers. That's an argument for putting AI in the right places with the right guardrails, not against using it. If you're unsure where your rules should come from, see AI policy vs. AI strategy.


How should a small insurance agency start with AI?

A small insurance agency should start with AI in the back office and keep every client-facing moment human. Pick the high-volume, low-judgment work where errors currently come from manual handling and time goes to logistics: claims intake and routing, renewal preparation, quote preparation, data entry and searching policy manuals or underwriting guidelines. Build confidence and discipline there first. Use AI to prepare agents for claims calls and renewal conversations, never to replace them, and require a licensed person to review anything that touches coverage advice before a client sees it. Choose business-grade tools whose data handling you understand, and write down who reviews what. Then measure outcomes that matter, such as claims processing accuracy, renewal retention and compliance audit results, rather than how many documents the tool produced. That approach fits Iowa's AI bulletin and protects the relationships an agency depends on.

The agencies that get real value from AI share that pattern. They also have someone who understands both the technology and the operating realities of insurance. Most AI vendors know their product. They don't know your carriers, your book or how your state regulator thinks, and they won't tell you when their tool is the wrong fit. Our AI strategy and fractional AI leadership work is built for that gap. If you're not sure your agency is ready, our AI readiness assessment is a good place to check.


Mosaic works with insurance agencies and financial services firms across the Cedar Rapids–Iowa City Corridor on exactly this kind of decision. If you're sorting through vendor pitches, our Advise, Educate & Lead page explains how we help.

Frequently Asked Questions

How can a small insurance agency use AI?

Start with back-office work that is high in volume and low in judgment: extracting information from claims intake documents and routing them, preparing renewal files, quote preparation, data entry and client record updates. Internal search across policy manuals, coverage summaries and underwriting guidelines is another strong early use, especially for training new producers. Keep the client conversation itself with a person.

Should an insurance agency use a chatbot for claims?

Be very careful. A client filing a claim has just had a bad day and wants to know a person is handling it. Automating that moment tends to cost more in trust and retention than it saves in staff time. Use AI behind the scenes to route the claim, gather documents and prepare the adjuster or agent, so the human conversation is faster and better informed.

Has Iowa adopted rules on insurers' use of AI?

Yes. The Iowa Insurance Division issued Bulletin 24-04 in November 2024, based on the NAIC's model bulletin on insurers' use of AI systems. It reminds insurers that existing Iowa insurance law, including rules on unfair trade practices and unfair discrimination, applies to decisions made or supported by AI. Agencies should expect carriers to ask how they use AI too.

Can AI make coverage recommendations to clients?

Not on its own. Coverage advice carries compliance and errors-and-omissions exposure, requirements vary by state, and an AI model can produce a confident recommendation that is simply wrong. AI can help an agent prepare, by summarizing a client's current coverage or flagging possible gaps, but a licensed person should review and own any recommendation before a client hears it.

Drew Bloom, founder of Mosaic Solutions

Drew Bloom

Founder, Mosaic Solutions — Business Systems, Software & Automation

Drew is the founder of Mosaic Solutions. He spent several years as an automation consultant and Innovation Lead for Affinity, where he worked with law firms and enterprise clients on AI strategy, change management, and custom software. He works with small and mid-sized businesses in Iowa and beyond to fix the systems that slow them down — with better processes, software, and AI that actually works.

More about Mosaic LinkedIn Work with Drew

Bring Us the Decision in Front of You.

AI strategy, policy, vendor choices and fractional technology leadership, scoped first and billed for the time the work takes.