Guide
How insurers, agencies and brokers use AI today, what the regulators expect, where the risks are and what changes for insurance jobs.
Updated 8 min read
The short answer
AI in insurance means using software that reads documents, spots patterns and drafts decisions to speed up underwriting, claims, sales, service and fraud checks. It helps people work faster. Regulators still hold the insurer responsible for every result.
Key points
Side by side
Typical tasks by area and who keeps the final say.
| Criterion | What AI does | Who decides |
|---|---|---|
| Underwriting | Reads files, scores risk | Underwriter |
| Claims | Reads documents, checks coverage | Adjuster |
| Agency service | Drafts replies, tracks renewals | Agent or broker |
| Fraud | Flags unusual claims | Investigator |
Insurance runs on documents, data and judgment. Policies, applications, claim forms, photos and emails all need to be read, checked and acted on.
Insurance artificial intelligence applies models to that work, in auto and home, health, and AI in life insurance.
Some models predict risk. Others read a document, write a reply or flag a claim that looks unusual.
88%
of responding auto insurers use or plan to use AI
84%
of health insurers use AI or machine learning
77%
of insurers are adopting AI, up from 61% in 2023
Adoption is already wide, and regulators track artificial intelligence in insurance sector by sector.
The National Association of Insurance Commissioners (NAIC) reports that 88% of the 193 auto insurers that responded use, plan to use or plan to explore AI and machine learning models.
In health, an NAIC survey run by 16 states found that 84% of health insurers use AI and machine learning in some capacity.
AI in health insurance is now a regulatory topic as well as a technology one.
Conning's 2024 technology survey found that 77% of respondents were in some stage of adopting AI, up from 61% in 2023.
These are surveys of insurers, so read them as a sign of direction, not as proof of results.
Underwriting means deciding whether to accept a risk and at what price. AI in insurance underwriting helps by reading applications, pulling data from several sources and scoring risk before a person looks at the file.
Machine learning in underwriting is used mostly on high-volume, standard risks. Complex cases still go to an underwriter.
Claims are where many carriers start. AI in claims processing reads the claim form, photos and policy, checks coverage and routes the file.
This is what claims automation in insurance means: routine claims move quickly, and adjusters spend their time on the complex ones. We cover this in detail on our insurance claims automation page.
Agencies use AI for the daily workload: quote intake, renewals, certificates and follow-up. Insurance agency AI works best on repetitive tasks with clear rules, with a person checking anything unusual.
AI for insurance sales covers lead replies, quote follow-up and renewal reminders.
In customer communications, AI drafts replies and answers routine questions, and hands harder cases to a person. See AI in customer service for how that works in general.
45%
say leaders believe gen AI's benefits outweigh the risks
Generative AI in insurance writes and summarizes. It drafts emails, summarizes long claim files and answers questions from policy documents.
Seventy-six percent of the 200 US insurance executives Deloitte surveyed in June 2024 had implemented gen AI in one or more business functions, yet only 45% said their leaders believe the benefits outweigh the risks.
$308.6B
yearly cost of insurance fraud in the US
10%
of P&C claims are estimated to be fraudulent
80%
of respondents use predictive modeling to detect fraud
The Coalition Against Insurance Fraud estimates that insurance fraud costs the U.S. $308.6 billion a year, according to the Insurance Information Institute. Deloitte estimates that about 10% of property and casualty claims are fraudulent.
Predictive modeling is widely used. An industry study cited by the Insurance Information Institute found that 80% of respondents use it to detect fraud, up from 55% in 2018.
Deloitte says current detection rates run 20% to 40% for soft fraud and 40% to 80% for hard fraud, so there is room to improve.
~92%
of surveyed health insurers have AI governance principles
The NAIC adopted its Model Bulletin on the Use of Artificial Intelligence by Insurance Companies in December 2023.
It sets expectations for how insurers govern and control AI, and states can adopt it or write their own rules.
NAIC's health insurer survey found that nearly 92% have AI and machine learning governance principles in place that model the NAIC AI Principles.
Check your own state's insurance department for the rules that apply to you.
The main risks are bias, decisions nobody can explain and weak data. A model trained on skewed history can repeat that skew in pricing or claims.
Explainability matters because a denied claim or a higher premium needs a reason a person can state.
Keep a person in charge of final decisions and test models on real past cases before relying on them.
−4%
projected change in underwriter jobs, 2025 to 2035
81%
of underwriting executives expect AI to create new roles
The Bureau of Labor Statistics projects underwriter jobs to decline 4 percent from 2025 to 2035, as automated underwriting software reduces the need for underwriters. It also expects technology to automate some claims adjuster tasks.
The NAIC says AI is more likely to support human workers than replace them entirely.
In an Accenture survey of 430 underwriting executives, 81% believe AI and gen AI will create new roles. The work shifts toward exceptions, judgment and customer contact.
Pick one process with clear volume, such as quote intake or claim triage. Measure the hours it takes today, then test the AI on past files before it touches live work.
We build AI solutions for insurance teams on the tools you already use, so you don't need to replace your systems.
If you run an agency, see our AI insurance services page, or read AI for business for the wider picture.
Our data
We asked ChatGPT, Gemini and Perplexity 10 buyer questions about AI for insurance agencies. They named 119 companies, and 85% appeared in only one of the three.
| Criterion | ChatGPT | Gemini | Perplexity |
|---|---|---|---|
| Companies named | 48 | 46 | 53 |
| Sites cited most | sonant.ai (3 of 10) | reddit.com and wyecliff.ai (3 of 10) | agentforthefuture.com and cloudtalk.io (3 of 10) |
Nexum measurement, October 20, 2026, via DataForSEO: 10 buyer questions to each AI, US English, 30 answers in total. 10 companies appeared in all three. Google showed an AI Overview for 9 of the 10 questions.
FAQ
Final step
You leave knowing which processes to automate first and what each one costs you today, whether we work together or not.
You can keep paying for manual work, or find out this week exactly what it costs you.
What happens on the call
Before
You answer 3 quick questions when you book, so we come prepared.
On the call
We find the manual work that costs you most and put a number on it, live.
Decide
An honest read on whether we're the right fit. If we're not, you'll hear it on the call.
Once you book, we look at your site and tools before the call, so the 30 minutes go to your numbers, not introductions.
Prefer to write first? Email us directly: contact@nexumautomations.com
