--- title: "We Have No Marketing Team. Our Biggest Channel Turned Out to Be AI Assistants." url: "https://marketermagazine.co/insight/we-have-no-marketing-team-our-biggest-channel-turned-out-to-be-ai-assistants/" author: "Daniel Brinzan" published: "2026-09-21" updated: "2026-09-21" --- # We Have No Marketing Team. Our Biggest Channel Turned Out to Be AI Assistants. We are three people building a consumer finance product. None of us runs marketing. There is no growth function, no content calendar, no paid acquisition budget. For a long time I treated that as a gap we would fix later. Then I looked at where new users were actually coming from, and a real share of them said the same thing. They asked ChatGPT or Perplexity a question, our product came up in the answer, and they showed up. Nobody on our side had placed us there. The assistant did. That changed how I think about the top of the funnel. For a growing part of the market, the first touch is no longer a search results page you can optimize your way onto. It is an AI assistant deciding what to mention. And you do not buy your way into that answer. ## **Search used to be a page. Now it is a recommendation.** The old model was legible. Someone typed a query, got ten blue links, and you competed for one of them with keywords, backlinks, and content volume. You could see the board and play it. AI assistants collapse that. A person asks a question in plain language and gets one synthesized answer with a handful of things named inside it. There is no page of options to rank on. You are either in the response or you are not, and being absent is invisible. The user never learns you existed. This is the part marketers keep underrating. Losing the old game meant slipping to page two, where a few determined people still found you. Losing the new game means not being said at all. ## **You cannot keyword your way into an AI answer** The instinct is to treat this like SEO with new tactics. Teams are already spending on "AI visibility strategy" the way they once spent on keyword density. I think that is aiming at the wrong thing. An assistant names something because there is real material for it to draw on. Documentation that explains how the product actually works. People discussing it in public because they used it. Independent write-ups, forum threads, comparisons written by someone with no stake in the outcome. The model is retrieving what exists about you, not what you claim about yourself. We never wrote for this. We routed part of our product to specialized infrastructure partners, integrated with them in public, and shipped something people could actually use and talk about. The integrations were engineering decisions, not marketing ones. The visibility came out of the product being real and discussed, and it arrived as a side effect rather than a campaign. So the honest version of an "AI visibility strategy" is uncomfortable for most marketing teams: build something worth describing, then make sure accurate descriptions of it exist in public. If the only material about your product is copy you wrote about yourself, the assistant has nothing to lean on, and it will reach for a competitor who left a real trail. ## **The users who arrive this way are different** The channel changes who shows up, not just how many. People who arrive from an AI answer start with intent. They asked a real question and got pointed somewhere. They are not clicking a paid placement that interrupted them or chasing an incentive. They came looking for a solution to a problem they already had. Those users behave better once they arrive. They convert at a higher rate and they stay longer, because they were not bribed in with a discount or a token reward that evaporates the moment it stops paying out. The old playbook of buying attention and hoping it sticks produces the opposite: cheap to acquire, expensive to keep. Intent-led users invert that. For anyone doing market analysis, this is the shift worth tracking. The metric that used to matter was cost per click. The metric that matters now is presence in the answer at all, and the quality of user that answer sends you. ## **What this means for how you spend** If you run marketing, three things follow. Measure it. Ask new users, in plain words, how they found you, and watch for AI assistants showing up in the answers. Most analytics will not attribute this cleanly, so the qualitative signal is the one you have. If it is already meaningful, it will keep growing. Fund what the model can cite. Real documentation, honest public case studies, genuine third-party discussion. Not landing pages tuned for keywords. The question to ask before a spend is simple: does this create material an assistant could accurately draw on, or does it only exist to rank? Fix the product before the funnel. An assistant recommending a product that disappoints just accelerates the disappointment. When the first touch is a trusted recommendation, the product has to be good the moment someone opens it, because the referral came with borrowed credibility you did not earn and can burn fast. ## **Conclusion** The top of the funnel is quietly moving from a page you optimize to a recommendation you earn. You cannot buy your way into an AI assistant's answer, and you cannot keyword your way in either. You get named because the product is real, people use it, and accurate accounts of it exist in public for the model to find. That is inconvenient for teams built around buying attention. It is very good news for teams built around making something worth talking about. The distribution follows the substance, which is the way it was supposed to work all along. ## **Frequently Asked Questions** **Is AI search really a big enough channel to plan around yet?** For most categories it is early but climbing. Treat it as a signal to track now rather than a channel to bet the budget on. The teams that understand how it works before it matures will be positioned when it does. **Can I pay to appear in AI assistant answers?** Not in the way you buy search ads today. Assistants surface what they retrieve as relevant, which is why the durable lever is creating accurate, public material about a product that genuinely solves the problem, not ad spend. **How is this different from traditional SEO?** SEO competes for ranked positions on a results page. AI answers name a small set of options with no page to rank on. You are in the response or absent, and absence is invisible to the user, which raises the cost of being left out. **What kind of content actually helps?** Material a model can cite: clear product documentation, honest third-party reviews, real case studies, genuine discussion in public communities. Self-written marketing copy carries little weight because the model can tell it is a claim, not evidence. --- Daniel Brinzan is the Founder of [Nika Finance](https://nika.finance/), a non-custodial mobile application combining spot trading, perpetuals, staking, yield, and prediction markets in a single interface.