Article
AI in insurance: Use cases, risks, and governance priorities
June 17, 2026 · Authored by Chris Wagner and John Romano
Key Takeaways
- AI is reshaping underwriting, claims, customer service, and back-office insurance operations, making governance urgent for insurers that want practical value without unnecessary risk.
- Data quality, explainability, bias, discrimination, and third-party risk can affect policyholders, regulators, boards, and management teams, so insurers need a clear view of where AI is used.
- Strong AI governance can help insurance leaders balance innovation with oversight, especially as state regulatory expectations continue to develop.
AI is already changing how insurers work, but the organizations that benefit most will pair new capabilities with strong data, clear governance, and human oversight.
The insurance industry is rich with data. Policies, claims, customer interactions, underwriting files, third-party inputs, and operational records all create information that can support faster analysis and better-informed decisions.
As AI embeds in insurance workflows, the conversation around it is shifting. Now the question is no longer whether but how it fits. Companies are finding ways to use AI to improve efficiency, strengthen decision-making, protect consumers, and meet growing oversight expectations.
Here’s what to know:
- How is AI changing insurance?
- Where are insurers using AI?
- Why does human oversight of AI matter?
- What AI risks do insurers need to manage?
- How can insurers build practical AI governance?
- Where can insurers start using AI?
How is AI changing insurance?
AI is gaining momentum in insurance because the industry runs on information.
Life insurance, auto insurance, homeowners insurance, health insurance, and property and casualty insurance all depend on large volumes of data. That creates practical opportunities for AI to help teams review information, identify patterns, summarize complex files, and support decision-making.
For insurers, AI can improve both speed and quality. It can help employees move through large amounts of information more efficiently, but it can also help them spot trends or risks that may be difficult to identify through manual review alone.
The opportunity isn’t limited to automation. AI can also change how work gets done by helping people focus less on repetitive review and more on judgment, analysis, and customer impact.
Where are insurers using AI?
AI is already showing up across core insurance operations. In many cases, it supports employees rather than replacing them.
Underwriting is one of the clearest examples. AI can assist with policy-related information, risk review, and business intake. Property and casualty companies, along with managing general agents, may use AI-supported processes to help move business through the underwriting chain.
AI can help review large claim files, identify trends, support fraud detection, and assist with litigation analysis.
These use cases can help claims professionals work through information faster while keeping people involved in decisions that require experience and judgment.
Customer service also has strong potential. A representative who’s new to a product, policy, or customer relationship can use AI to quickly understand relevant information and respond with more confidence.
AI can support several insurance functions:
- Underwriting review
- Claims file analysis
- Fraud detection
- Litigation analysis
- Customer service support
- Back-office operations
- Third-party and vendor-enabled workflows
AI supports decision-making. It can help organize and analyze information, but insurance leaders still need to decide where people review outputs, approve actions, and remain accountable.
These use cases can help claims professionals work through information faster while keeping people involved in decisions that require experience and judgment.
Why does human oversight of AI matter?
AI can make insurance professionals more efficient, but it doesn’t remove the need for human judgment.
That matters because insurance decisions can directly affect consumers. Underwriting, claims, pricing, coverage, and service interactions all carry business, regulatory, and reputational implications. When AI influences those areas, insurers need clear oversight.
Human oversight helps organizations review AI outputs for sense, reliable data, and alignment with expectations. It also supports explainability, especially when regulators, customers, or internal leaders need to understand how AI-informed decisions were made.
This human-in-the-loop approach can help insurers use AI as an enabler. Employees can spend less time sorting through repetitive information and more time applying expertise where it matters most.
What AI risks do insurers need to manage?
AI risk often starts with data.
If an insurer’s data is incomplete, inconsistent, or poorly governed, AI may magnify those issues. The garbage-in-garbage-out maxim applies to insurance AI, too.
Regulation adds another layer. Insurers may face state-level expectations around where AI can be used, how decisions can be explained, and how consumers are protected. That can include concerns related to bias, discrimination, explainability, confidentiality, and disclosure.
Consider third-party risk, too. Insurers often rely on vendors, managing general agents, and other partners. If those third parties use AI, the insurer may still need visibility into how AI affects the broader operating model.
Key risk areas to review
- Data quality and completeness
- Data governance and access controls
- Confidentiality and sensitive information
- Bias and discrimination
- Explainability gaps
- Regulatory expectations
- Third-party and vendor risk
- Model drift over time
Pay special attention to model drift. AI models can change as data, usage patterns, and assumptions evolve. Outputs that appear reasonable today may shift over time, which makes monitoring a core part of responsible AI adoption.
Pay special attention to model drift.
How can insurers build practical AI governance?
AI governance provides insurers with a framework for moving from experimentation to responsible adoption.
At its core, governance helps answer practical questions:
- Who owns AI strategy?
- Where is AI already being used?
- What data supports those tools?
- Which use cases involve consumers?
- What vendors are involved?
- How are outputs monitored?
Frameworks such as the National Institute of Standards and Technology Artificial Intelligence Risk Management Framework (NIST AI RMF) and International Organization for Standardization (ISO) 42001 can help shape that thinking, but governance also needs to work inside the organization’s day-to-day operations.
A useful AI governance program can include:
- Clear AI strategy
- Inventory of AI tools and use cases
- Defined accountability across leadership, risk, compliance, technology, and the business
- Data governance standards
- Acceptable use guidance
- Third-party risk management
- Ongoing monitoring and reporting
- Board and executive visibility
Governance doesn’t have to block innovation. It can help insurers make better decisions about where AI belongs, what guardrails are needed, and how the organization can scale AI use with more confidence.
Where can insurers start using AI?
AI adoption doesn’t have to happen all at once. A phased approach can help insurers build familiarity, manage risk, and create momentum.
Crawl: Start with lower-risk AI support
Begin with use cases that help employees work faster without shifting major decision-making to AI. Think document review, summarization, internal research, customer service support, or back-office assistance.
Walk: Expand into AI-driven insights
With mature governance and data practices, insurers can use AI to support recommendations, trend analysis, claims review, underwriting support, and risk identification. Human review remains important, especially when outputs influence customers or regulated decisions.
Run: Move toward more advanced AI-enabled workflows
With strong monitoring, defined approvals, clear accountability, and board-level visibility, insurers can move to more advanced AI adoption, including support in underwriting, claims, service, or operations.
Some good questions to ask:
- Where are employees already using AI?
- What data supports those use cases?
- Which vendors or partners use AI on the insurer’s behalf?
- Which AI activities touch underwriting, claims, consumers, or regulated decisions?
- What governance and reporting does leadership need?
Treat AI as a business opportunity and a risk discipline to use it well. Strong data, clear accountability, and practical governance can help insurers move AI from experimentation into more responsible day-to-day use.
