The future of agentic distribution in Insurance - our CEO's interview with McKinsey

Last updated: 2 November 2026

AI agents are changing how people find, compare and buy insurance. At ITC Vegas, Waniwani cofounder and CEO Robin Diligent sat down with Sarah Blatt, Associate Partner in McKinsey’s Growth and Insurance Joint Venture, to talk about agentic insurance distribution: what it changes for insurers, and what they need to build to compete. McKinsey published the full interview. This post sums up the four ideas we think matter most.

The four takeaways in short

1. Why the insurance market reacted so strongly to AI

Insurance distribution was built on friction and information asymmetry between buyers and sellers. AI removes both. That is a paradigm shift, and it explains why brokerage stocks sold off sharply when the Tuio app launched in ChatGPT in February 2026, while nothing similar happened in retail, even though agentic commerce was more advanced there.

The friction is easy to describe. To get one quote, a buyer gives a lot of information. To compare quotes, they go through the same process again and again. As Robin put it in the interview:

“The friction within distribution is disappearing with AI, and that will massively transform the way insurance is distributed.”

This does not mean brokers disappear overnight. Price comparison websites made the UK market more competitive, but that change took years. AI can move faster than that, but the transition will still take time.

2. Why opting out of AI distribution is not an option

Consumers and businesses already use AI to get quotes and buy insurance. Personal assistants such as Muse and Instinct are visiting insurer websites at scale, and no one can stop this. If an insurer is not present in these assistants, it closes itself off from a growing part of the market.

Blocking agents is not a long-term answer. Some companies restrict agent access while they define a strategy, but this does not hold: Insurify first blocked agents and then changed its mind and now lets them in.

The real question is control:

• What information do you give to AI agents?
• How do you limit the risk of hallucinations about your products?
• How do you make sure that the information the agent gives back to the user is accurate and compliant?

3. Why an app in ChatGPT is not enough

In the early days of the internet, a website was not enough to sell. Vendors had to bring traffic to it, track conversion, understand why buyers converted or not, and optimize. These are classic acquisition channel problems, and a new generation of companies emerged to give web vendors the infrastructure they needed. Insurers now need the same for AI.

Two gaps make this hard today. First, large language models are a black box for insurers: they do not see the prompts that shape a buying decision, and it is very hard to act on something you cannot track. Second, insurance distribution infrastructure (websites, forms, call centers) was designed for people, not for AI agents.

An agent-ready channel has four parts:

1. Presence. Make your products available directly inside AI assistants such as ChatGPT, Claude, Gemini and Muse, so agents get quotes from your own tools, not from scraped web pages.
2. Traffic recognition. Know whether a visitor is a person or an AI agent acting for that person, and give each one the right experience. Exposing your quoting tools directly to agents gives you more control over what they receive.
3. Measurement. Treat agent interactions as a real acquisition channel: what agents were looking for, how many converted, and where users dropped off. Synthetic buyers that simulate conversations at scale show where you are likely to win and why.
4. Optimization. Use these insights to improve the offer, the messaging, the user experience and the funnel.

Visibility in AI answers (often called generative engine optimization, or GEO) is part of this. It starts with the right prompts: the questions real customers ask, not keywords copied from search. Then you track how often your brand appears and where it ranks, and you improve both your content and how easily agents can retrieve your information. For a deeper look at the channel, see what AI distribution means for financial services.

4. Why compliance is the condition for AI distribution to scale

Today, agents like Muse and Instinct quote insurance without any constraints. In a regulated market, this cannot last. The insurance market needs clear rules on what AI agents can and cannot do when they buy insurance. Without them, the market will continue to react chaotically.

For insurers, compliance applies at two levels. When an agent gives advice, it must respect regulatory requirements. When it handles customer data, it must protect it. On top of regulation, each insurer has its own business rules about how its products can be sold, and agents must respect those too. We wrote about the current regulatory gaps in AI is already distributing insurance, the regulations haven’t caught up.

What changes for retention and for intermediaries

Lower switching friction puts pressure on some common practices. Charging long-standing customers more than new ones becomes very hard to keep when an agent can check prices and coverage at any time. A customer may switch before the insurer earns back what it paid to acquire them.

The quote economics also change. As quoting gets easier, the number of quotes may rise without a matching rise in purchases. Insurers will need better data on how customers discover products and what drives their decisions.

AI puts pressure on intermediaries whose main role is to move information from one system to another. It also increases the value of differentiated advice, service and products. Everyone in the value chain is adapting: price comparison platforms such as Insurify and MoneySuperMarket moved to ChatGPT first, followed by large carriers such as Aviva, Liberty Mutual and Allstate.

How far will agentic distribution go?

The impact is already clear in personal lines such as auto, home and pet insurance. Commercial and cyber insurance are different, because trust and expert advice play a larger role. But AI is already part of those decisions too: a business can get a quote from a broker and then use AI to challenge the price or check the coverage. It is simply harder to see and track.

The opportunity goes beyond distribution. Lower quoting and switching costs bring insurance back to its fundamentals: understanding risk, building strong underwriting strategies, and offering the best product at the best price. The richer data from new customer interactions can help insurers improve their offers and, where regulation permits, make pricing more responsive. Teams that can collect, analyze and act on this data quickly will have an advantage.

FAQ

What is agentic insurance distribution?

Agentic insurance distribution is the sale of insurance through AI agents and assistants, such as ChatGPT, Claude, Gemini or Muse, that search, compare and request quotes for a customer. The agent does the work that a buyer used to do on websites and forms, which removes most of the friction in the buying process.

Why did insurance brokerage stocks fall after the Tuio app launched?

When Tuio launched one of the first insurer-built apps in ChatGPT in February 2026, investors saw that AI could remove the friction and information asymmetry that insurance distribution depends on. Brokerage stocks sold off on concerns that AI distribution could disrupt traditional intermediaries.

Can insurers block AI agents from their websites?

They can for a short time, but it is not a long-term strategy. Customers already use AI to get quotes, and an insurer that blocks agents closes itself off from that demand. Insurify first blocked agents and later let them in. The better approach is to control what information agents receive.

Is an app in ChatGPT enough to sell insurance through AI?

No. Like a website in the early internet, an AI app needs traffic, conversion tracking and continuous optimization. Insurers need infrastructure to see how agents and customers find their products, why they convert or drop off, and how to improve results while staying compliant.

Will AI agents replace insurance brokers?

Not overnight. Price comparison websites took years to reshape the UK market, and AI will also take time, even if it moves faster. AI puts pressure on intermediaries that only move information between systems, but it increases the value of brokers who give differentiated advice and service.

Which insurance lines are most affected by AI agents?

Personal lines such as auto, home and pet insurance are affected first, because the products are simpler to quote and compare. Commercial and cyber insurance depend more on trust and expert advice, but businesses already use AI to challenge broker quotes and check coverage.

Why does compliance matter for AI insurance distribution?

Insurance is a regulated market, and today some agents quote without any constraints. Insurers must make sure that agents give accurate, compliant information, protect customer data and respect each insurer’s business rules. Clear rules on what agents can do are necessary for the channel to grow in a stable way.

Read the full interview on McKinsey.com. Thanks to Sarah Blatt for the conversation, and to Kate McCarthy and Grier Tumas Dienstag for making it happen.