Insurance AI's next wave: 5 opportunities hiding in plain sight

Insurance isn't short on risk. It's short on time.

AI's first win is buying back hours spent reading, re-keying and chasing documents, so humans can spend more time on judgement.

What's changed in 2026 is that AI is no longer sitting on the sidelines as a clever analysis layer. It's moving into the machinery of insurance itself. Intake, triage, drafting, routing, follow-ups. The insurers getting real lift aren't just sprinkling tools across teams. They're building agentic teams, where people and intelligent systems share the work in coordinated workflows, with clear hand-offs and decision rights.

That matters because insurance isn't a "best model wins" game. It's a "best workflow wins" game. If AI stays stuck in pilots, it stays a cost line and a distraction. If it becomes a dependable part of day-to-day work, it turns into capacity. And capacity is the gateway to growth.

Here are five opportunities where leaders are already turning AI into that capacity.

1) Turning underwriting speed into a commercial weapon

In Drova's recent Insurance Outlook Report, Nigel Fellowes-Freeman, CEO and Founder of Kanopi Cover, captures how quickly agentic workflows have moved into production.

"Historically, I would have said there's no way you'd see purely agentic workflows inside an insurer within the next five years. But we already have a tier-one insurer running an agentic workflow for cyber insurance in production today, and it has taken the underwriting process from two days down to about five minutes."

Five minutes is not a small improvement. It resets expectations. It changes how brokers prioritise markets, how quickly customers get certainty, and how underwriting time is spent.

The report flags a widening gap between insurers that are operationalising AI and those still experimenting. When underwriting speed compresses this sharply, "pilot pace" becomes a commercial disadvantage.

In practical terms, the upside is simple. Underwriting teams spend less time reading, re-keying and chasing the basics, and more time on judgement, negotiation and exceptions.

2) Making unstructured information usable, at scale

Insurance runs on PDFs, emails, attachments, and documents that arrive in whatever format the world felt like using that day.

AI's near-term value is brutally simple: convert that chaos into structured inputs and useful prompts so humans can make better calls faster.

Katherine Simmonds, CEO, Fusion Specialty Insurance, points to the shift away from "tech imposed on the business" and towards AI being built from the frontline:

"Historically, at least in our industry, tech systems were imposed on the business and people by a tech team or a software provider. AI is almost going to be the reverse; it will be created and brought from the ground up. We're actively encouraging that type of innovation."

What does that look like in practice?

"One of the things we're using AI for is an AI submission reader that takes the data - it can read up to 500 pages - and brings the key data points into our system. Another AI project will produce some of the documentation that our underwriters - professionals on professional salaries - currently create."

This is where agentic teams become a competitive advantage. They connect extraction, triage, drafting, and routing into one coordinated flow. The underwriter is not "supported by AI" in the abstract. They are supported by a system that knows what the objective is and what evidence is required to progress the work.

3) Lifting claims and contact centres from reactive to proactive

If underwriting is where you win a policy, claims is where you win the customer.

AI's opportunity in claims and service is not limited to faster triage. It is about momentum and clarity. The ability to keep customers informed without creating a new administrative burden. The ability to move from sampled quality checks to broader coverage, while reducing repetitive work.

Richard Joffe, Founder, Honey Insurance, describes the practical gains already in production:

"We use Al in many areas of our business. One example is call notes: as soon as a call is finished with someone in our contact centre, the notes from that call are automatically created, synthesised and dropped into Salesforce. That saves a huge amount of time that would otherwise be monotonous."

He describes how AI expands quality assurance and compliance coverage:

"We listen in on calls, so it's easier for us to do quality assurance and risk and compliance checks at scale, rather than relying on people listening to tiny little soundbites. We use AI to coach and manage compliance, and we use it to speed up and make our people more potent - they can spend more of each hour doing good work with customers instead of bureaucracy and paperwork."

This is the customer experience opportunity. Claims and service teams spend less time documenting what happened, more time progressing what happens next.

4) Smarter pricing for catastrophic risk

There is a version of insurance AI that is just "drafting and summarising". Useful, but limited.

The bigger upside is decision support in the hard places, especially pricing and portfolio decisions where catastrophic risk keeps reshaping the map.

Richard Joffe, Founder, Honey Insurance, points directly to this link between AI and pricing for catastrophe-linked cover:

"I also think AI will play a big role in pricing certain types of cover, particularly catastrophic risks, where you can link AI to really great weather data and models to predict what the risks might be."

The opportunity is two-fold. Better risk selection and pricing discipline, and faster iteration as conditions change. When AI helps connect external data, internal loss experience, and policy structure, teams can move from periodic re-rating to more responsive decision cycles.

This is also where insurers need to be honest about capability. It is not enough to have a model. You need a workflow that brings insight into underwriting and portfolio steering in a way people can act on, repeatedly.

5) Distribution shifts when agents start shopping for insurance

The next distribution shift may not be a new aggregator. It may be AI agents acting on behalf of customers.

Nigel Fellowes-Freeman, CEO and Founder of Kanopi Cover, describes this direction clearly:

"The last piece will be in distribution - meaning you'll use platforms you use today, whether it's Gemini, ChatGPT, or whatever large language model interface you want, and you'll ask it to find you some insurance products. Then you'll have agents go out and interact with an MCP layer that exists on top of insurance products."

"That will start to be how we distribute insurance; instead of going to a website or a comparator and asking for something, we will use an LLM and an agent to do that job - and that's going to be extraordinary and transformative."

This is the opportunity hiding in plain sight: becoming agent-ready. Clear product structures. Clean data. Fast quote and bind pathways where appropriate. Straight answers that can be verified and compared.

In a world where an agent can shortlist options in seconds, insurers that are easiest to understand and transact with will get disproportionate attention.

The multiplier: Agentic teams that turn AI into capacity

These opportunities are not separate. They compound.

When underwriting intake is automated, underwriters have more time. When claims updates are proactive, call volumes fall. When call notes and QA are scaled, quality improves without slowing teams down. When pricing and risk signals are sharper, insurers can pursue growth with more confidence. When distribution becomes agent-led, the fastest, clearest insurers get found first.

What makes this work is not "AI everywhere". It is coordinated workflows where humans remain accountable and AI does the heavy lifting.

Or, as Joffe puts it:

"For me, it's very important that AI is never a decision-maker - AI is only a supporting assistant. The most important safeguard is that AI isn't able to make decisions by itself. It should be an accelerant to humans making decisions, not a replacement for them."

That is the opportunity frame for 2026. AI is buying back time. The insurers that reinvest that time into better decisions and better customer outcomes will pull ahead.