AI agents are beginning to change how insurance customers discover products, evaluate coverage, and compare prices. For insurers, brokers, and aggregators, the shift raises questions about visibility, the economics of distribution, customer loyalty, and the role of human advice.
Waniwani, a US-based insurtech founded in early 2026, quickly gained prominence for providing the underlying infrastructure for one of the first insurer-developed apps integrated into ChatGPT. Launched in February 2026 by Tuio, a Spain-based digital insurer, the app has enabled consumers to obtain personalized, indicative home insurance quotes directly within ChatGPT conversations before redirecting them to its website to complete the purchase. The launch triggered a sharp sell-off in insurance brokerage stocks, reflecting investor concerns that AI-powered distribution could disrupt traditional insurance intermediaries.
At ITC Vegas, Waniwani cofounder and CEO Robin Diligent spoke with Sarah Blatt, an associate partner from McKinsey’s growth and insurance joint venture, about building distribution channels for AI agents, the changing economics of insurance, and the capabilities insurers will need to compete.
This interview has been edited for clarity and length.
Why insurance distribution is changing
Sarah Blatt, McKinsey: Investors have reacted sharply to the prospect of AI disrupting insurance distribution. What do you think they understood correctly, and what’s still unclear?
Robin Diligent, Waniwani: They understood the scale of change. Insurance distribution, traditionally, has been built on friction and information asymmetry between buyers and insurers. If I’m a buyer, it’s very hard for me to get a quote and compare it. I need to give a lot of personal information to get one quote, and then I need to go through the same friction again and again in order to compare quotes. AI transforms this process, making it much easier to obtain and compare quotes. The friction within distribution is disappearing, and that will massively transform the way insurance is distributed. I think that’s the underlying fundamental of this market that is shifting, and that’s what the market reacted to that day in February.
The friction within distribution is disappearing, and that will massively transform the way insurance is distributed.
What I think some observers have missed is that this doesn’t mean that insurance brokers are disappearing. We’ve seen similar changes before: Price comparison websites made the UK market more competitive, but that transformation took years. AI could move faster, but the transition will still take time. Will it take 15 years? Maybe not. Maybe it will be faster, but it won’t happen overnight.
Sarah Blatt: You’ve said every insurer will be affected by this transformation eventually, whether it builds an AI app or not. Why is opting out just not an option?
Robin Diligent: It’s not an option because consumers and businesses will use AI to search for and compare insurance products, and no one can prevent that. In the end, the question is not about being there or not. If you choose not to be present in these AI agents, you’re voluntarily closing yourself off from this market.
If you choose not to be present in these AI agents, you’re voluntarily closing yourself off from this market.
The more important question is how insurers can control what is happening. What information are you providing to AI agents? How can you control the information that you provide to these agents, making sure that you limit the risk of hallucinations, that you control the compliance and the accuracy of the information being given back to the user? Some companies may restrict agent access while they work out their strategies, but I do not see that as a long-term solution.
Building an agent-ready channel
Sarah Blatt: What opportunity led you to start Waniwani?
Robin Diligent: We saw agentic commerce developing rapidly in retail and believed it was only a matter of time before the same shift reached services such as insurance.
We identified two gaps. First, LLMs [large language models] are a black box from the insurance vendor’s perspective. Insurers don’t have access to the prompts used by consumers to search, compare, and buy. It’s very hard to act on something you don't see or even track. Second, existing insurance distribution infrastructure, like websites, general agents, and other channels, was designed primarily for people, not AI agents. This is suboptimal in a world where most consumers will use AI to try to search for these products.
Waniwani’s mission is to help insurers regain control of the distribution, to understand this new channel, and adapt their distribution infrastructure so that it works for both humans and agents. We’re not saying that all insurance will be bought by AI agents. Traditional channels will remain important, but AI is becoming an important acquisition channel and a fast-growing one, so insurers need to adjust.
Sarah Blatt: How does Waniwani help insurers operate in that environment?
Robin Diligent: We help insurers do two things. First, make their offerings available directly through AI agents like ChatGPT, Claude, Gemini, Muse, and others. Second, maximizing the premiums and policies sold through these channels while ensuring regulatory compliance and AI security.
To take a concrete example, if you’re selling home insurance, Waniwani helps you distribute products directly through these AI agents. Then, we help you understand and optimize how customers discover, evaluate, and ultimately purchase your products.
We do this by capturing the conversations users have with the AI agents and simulating buying behaviors at scale. Starting from different personas, we create synthetic buyers who simulate these conversations. We then analyze the simulations to understand where you’re likely to win, where users might drop off, and why. Based on these insights, we work to optimize the offering, messaging, UX/UI [user experience and user interface], and the funnel to maximize conversion.
The final piece is ensuring that these agents remain compliant, both when giving advice and when handling customer data. We put the right guardrails in place to make sure they comply with regulatory requirements while also respecting the specific business rules that each insurer has around selling its products. We also work with you to mitigate the security risks surrounding these agents, notably competitors trying to scrape data and retro-engineer pricing.
Sarah Blatt: What determines whether an insurer appears in an AI-generated initial consideration set inside an LLM, and what do insurers need to understand about this new market dynamic?
Robin Diligent: This is a question we’re getting from our clients. I think about GEO [generative-engine optimization] in three parts. The first part is identifying the right customer prompts in which you want to be visible. If you monitor the wrong questions, the rest of the analysis will not tell you much. Real-world evidence of what customers ask is especially valuable; guessing prompts or simply extrapolating search engine keywords has limitations.
Real-world evidence of what customers ask is especially valuable; guessing prompts or simply extrapolating search engine keywords has limitations.
The second part is tracking and analytics: How often does your brand appear in response to relevant prompts, and where does it rank? The third part is optimization. That can involve creating GEO-optimized content on your website that AI systems can understand, influencing content from third-party sources, and making information on your website easier for agents to retrieve. There is a technical dimension as well as a content dimension.
Sarah Blatt: What does becoming agent-ready look like in practice, acknowledging that different players have different starting points?
Robin Diligent: Start by understanding the traffic coming to your digital channels. Is someone requesting a quote directly, or is an AI agent acting on that person’s behalf? Then consider whether the experience is appropriate for that visitor.
Today, AI agents may have to navigate websites designed for humans, clicking buttons and entering information into forms. That can be inefficient. Instead, insurers can identify whether a website user is human or an AI agent and create a specific agentic experience for the agents. There are many ways to do this. One way would be to expose the insurer’s quoting tools directly from the website to the AI agents, meaning the insurer can have more control over what information is shared with the agents. From there, treat agent interactions as a real acquisition channel, with analytics on what the agents were looking for, how many converted, and how to optimize the process.
Competition, retention, and the role of intermediaries
Sarah Blatt: As switching friction decreases in a more agent-ready world, does loyalty matter? How do you optimize retention, and what’s changing?
Robin Diligent: Some common strategies, such as charging long-standing customers more than new ones, will be very hard to keep. If an agent can continuously check prices and coverage, a customer might switch before an insurer earns the return it expected from acquiring that customer.
We still need to understand exactly how AI agents will weigh customer satisfaction, brand reputation, and service. But the basic principle is clear: having the ability to offer the best product at the best price at all times will determine whether an insurer can keep its customers and make long-term revenue.
Insurance is built on trust. For a technology company like ours, working with established participants helps us understand the market deeply and build credibility.
That’s what’s so interesting in the end. The focus switches to the core of insurance, which is understanding risk and underwriting strategy, offering the best product at the best price. We’re talking a lot about the threats, but there’s a big opportunity here for insurance companies, too, and that’s the exciting part.
Sarah Blatt: Looking across the insurance ecosystem, including carriers, brokers, aggregators, managing general agents, and insurtechs like Waniwani, which players will have to adapt? And what does being ready look like?
Robin Diligent: We’re seeing that everyone needs to adapt. The first players that moved to ChatGPT, for instance, were the price comparison platforms: Insurify, MoneySuperMarket, et cetera. They were followed by large carriers such as Aviva, Liberty Mutual, and Allstate. Every player is adjusting to this.
AI puts pressure on intermediaries whose primary role is simply moving information from one system to another. But it also raises the value of differentiated advice, service, and products.
The cost of obtaining quotes and switching insurers is falling significantly. As quoting becomes easier, the number of quotes requested may rise without a corresponding increase in purchases. That could change the economics. Insurers need better data on how customers discover products, what influences their decisions, and how those insights feed into underwriting and customer service strategies.
Sarah Blatt: Waniwani has formed partnerships with established players across the insurance value chain. Why are these partnerships important to your strategy?
Robin Diligent: We have partnerships with a range of players in the insurance value chain. What we’re seeing across all of them is that AI-powered distribution has become a top-of-mind priority. Whether you’re a price comparison website or an insurance carrier, this is a critical challenge that needs to be addressed. Technology is evolving faster than many insurance organizations can adopt it.
Insurance is built on trust. For a technology company like ours, working with established participants helps us understand the market deeply and build credibility. You cannot transform an industry effectively if you remain disconnected from how it operates. It’s very important to have a deep understanding of the market, and that’s what those partnerships allow us to get.
How far will agentic distribution go?
Sarah Blatt: Will agentic distribution remain concentrated in relatively simple personal lines, or will it move into more complex insurance?
Robin Diligent: The impact is already apparent in personal lines such as auto, home, and pet insurance. More complex products, including commercial and cyber insurance, are different because trust and expert advice play a larger role.
But AI can still participate in those decisions. A business might consult a broker for a quote and then use AI to challenge the price or assess whether the coverage is appropriate. AI is already there for more complex products; it’s just that it’s much harder to see and to track. My view is that this will continue to expand, although the impact will vary by line of insurance, depending on its complexity.
Sarah Blatt: What excites you most about the opportunity ahead?
Robin Diligent: The shift of focus. We’re talking a lot about distribution, but the agentic shift could bring insurance back to its fundamentals: understanding risk and developing strong underwriting strategies. Lower quoting and switching costs create competitive pressure, but new customer interactions can also produce richer information. Used well, that information can help insurers improve their offerings and, where regulations permit, make pricing more responsive.
Sarah Blatt: What about talent? How will the skill sets in the insurance industry need to change?
Robin Diligent: Data collection, analysis, and the ability to react quickly will become increasingly important. Insurance already has highly analytical underwriting teams, but I expect demand for additional technical profiles focused on understanding buying behavior across channels.
These teams will ask how customers behave when they use AI compared with traditional channels, what the differences mean, and how the insurer should respond. Underwriting teams will also need to work with larger and richer data sets. The ability to translate those insights into better products and decisions will be essential.
