More buyers now start their insurance research inside ChatGPT, Perplexity, and Google’s AI answers. The carriers and brokers those tools name are winning demand before a rival’s website ever loads, and a handful of firms have started treating that visibility as a growth channel with revenue attached.
Ask a small business owner how they found their last commercial policy and a growing number will describe the same thing. They did not open ten tabs of blue links. They typed a question into an AI assistant, read the answer it wrote back, and called one of the two or three names it mentioned. The research, the shortlist, and the first hit of trust all happened inside a single answer box.
That is a quiet change in how insurance gets bought, and it carries a loud commercial consequence. When an AI engine names a few carriers or brokers in response to a coverage question, everyone left out of that answer is out of the running before a prospect ever reaches their site. The old contest was for a ranking on a results page. The new one is for a mention inside the answer that replaces the results page.
The discipline built around winning those mentions has a name: generative engine optimization, or GEO. Its close relative, answer engine optimization (AEO), covers the same ground. Strip away the acronyms and the goal is simple. Get your firm surfaced, cited, and recommended inside the answers that AI assistants generate, not only listed in the links underneath them.
From marketing line item to revenue line
For most of the last two decades, search visibility sat inside a marketing budget and got judged on traffic. GEO is starting to be judged on something harder: pipeline.
The reason is where the AI answer sits in the buying journey. A traditional search result dropped a visitor on a website and left the site to do the convincing. An AI answer does the convincing first. By the time it says “brokers like X and Y are worth a call for this kind of cover,” it has already compared options and handed its trust to a named few. The prospect who then reaches out is not browsing. They arrive pre-sold on the idea that the firm is a credible choice, which is why the practitioners working this channel report leads that close faster and cost less to acquire.
That is the case for treating AI search as a revenue stream rather than a branding exercise. It sits at the top of the funnel, but it delivers bottom-of-funnel intent.
Incumbents are hoarding the answers, for now
Here is the part that should get a growth leader’s attention. The answers are not evenly distributed, and right now they favor the names that were already big.
Geology, an organic growth studio that works on AI search for insurers and brokers, reports that between 66 and 82 percent of AI mentions in insurance answers go to incumbent carriers, leaving challengers with roughly 13 to 16 percent. Read one way, that is discouraging. Read another, it is the whole opportunity. The share is soft. An answer engine is not trying to reward the biggest advertising budget. It is trying to give the user the clearest, best-supported answer to one specific question, and that is a contest a focused, well-organized firm can win against a slower incumbent.
Channels behave like this early in their life. The cost of showing up is low, the competition is thin, and the firms that move while everyone else is still debating whether AI search matters capture a share that gets expensive to dislodge later. The window is open now precisely because most of the market has not noticed it is open.
Why insurance feels this before other industries
Insurance is close to a worst case for the old model and a best case for the new one, for the same reason: trust.
Nobody buys a policy on impulse. They research, compare, and lean hard on whoever they decide to believe. Google formally classifies insurance content as YMYL, “your money or your life,” and holds it to a higher bar for expertise and trustworthiness than almost any other category. In that environment, being the option an AI assistant vouches for is worth far more than being one link among ten, because in a trust purchase the recommendation is the moment the sale actually begins.
Regulation raises the stakes further. An AI engine that misstates a premium, a coverage limit, or an exclusion is not a harmless error. State Department of Insurance advertising rules generally hold the carrier responsible for a claim made about its product regardless of who produced it, and an assistant confidently quoting the wrong number to a prospect is a version of that problem no one drafted a policy for. This is why the people working in the space draw a sharp line: being cited is not the win, being cited accurately is. Part of the job is monitoring what the engines actually say about a firm’s policies and correcting it before a customer acts on a wrong figure.
The funnel is splitting, and half of it is going dark
To see why this turns into real money and not just a visibility metric, split search into two tracks and watch what happens to each.
Informational questions, the “what is professional indemnity” or “do I need business interruption cover” queries, used to be dependable top-of-funnel traffic. AI Overviews and assistants now answer them on the spot and keep the user in the conversation. The firm whose content fed that answer earns the impression and loses the click. The awareness stage still happens, but it happens somewhere the carrier cannot see and did not get a visit from.
High-intent questions still behave the old way. “Insurance broker near me,” “best commercial cover for a dental practice,” “get a quote,” these still produce calls and form fills, because acting on them means leaving the search box.
The trap is what this does to measurement. A carrier watching only its website analytics sees top-of-funnel traffic shrink and may read it as demand falling, when demand has moved into answers it never measured. The first-party search and click data that used to reveal the awareness stage is thinning out. Judged by sessions alone, a firm can be losing deals inside AI answers it never knew it was being compared in, and then cut exactly the wrong spend in response.
The early numbers
The clearest evidence that this is a channel and not a talking point comes from the results firms are starting to publish.
Geology documents an insurance case in which a California broker went from effectively invisible, nothing ranking and no mention on any AI platform, to more than 20 qualified leads a week within three months, with search impressions up roughly tenfold and organic clicks around fivefold over the same quarter. The firm did not outspend its established competitors. It narrowed its focus instead of widening it, answered the specific questions its buyers were asking, and made sure it was described consistently everywhere an assistant might read about it. Similar patterns, organic growth paired with lower customer acquisition cost, show up across the firm’s other published case studies in adjacent sectors.
Numbers from a single agency are not an industry benchmark, and they should be read as what they are: early, first-party, and specific to one book of business. But they point at the shape of the opportunity. A modest, focused program moved a small broker from zero AI visibility to a steady weekly lead flow in a quarter, on surfaces a standard rankings report never shows.
The answers are not built only from your website
One finding tends to surprise insurance executives who assume this is just SEO with a new label. AI engines assemble an answer from many sources at once, and a lot of those sources are not corporate websites.
By Geology’s count, insurance answers pull from an average of seven to sixteen cited sources and name a carrier in nearly every one. The single most-cited domain in Perplexity’s insurance answers, in its sampling, was Reddit, which appeared 75 times across 100 prompts. That means a meaningful share of the trust an AI assistant transfers to a named carrier is being shaped in community threads, comparison hubs, and review platforms, not on the carrier’s own pages. Showing up in the answer is partly about a site being crawlable and authoritative, and partly about how the firm is described and corroborated everywhere else the model reads.
What a carrier can do this quarter
The first move costs nothing and takes an afternoon. Open ChatGPT, Perplexity, and Google’s AI answers, and ask the questions customers ask, in the firm’s own product categories and regions. Note honestly where the firm appears, where a competitor appears instead, and whether anything the engine says about its coverage or pricing is simply wrong. For most carriers and brokers, that single exercise is the moment the abstract becomes urgent.
From there the work is recognizable but reorganized. The technical SEO fundamentals still matter, because AI engines retrieve from search indexes before they generate an answer, so crawlable, well-structured, server-rendered pages remain the price of entry. On top of that sits the newer layer: content built to be quoted, consistent descriptions of the firm across the web, third-party corroboration, and active monitoring of what the engines say so errors get caught before customers do. Specialist firms have started to package this specifically for regulated insurance work, where compliance review is part of the pipeline rather than an afterthought.
The strategic point is the one worth ending on. AI search is not a new place to advertise. It is becoming the place where a growing share of insurance buyers form their shortlist and their first judgment of trust, which makes the mentions inside those answers a channel with revenue attached. The carriers and brokers treating it that way now, while the answers are still cheap to win, are building a lead in a channel their competitors have not started to measure.
