Selling on ChatGPT starts with answer visibility, not ad tactics

TL;DR: Treat ChatGPT commerce as an AI search visibility problem first, because brands need clean, answer-ready facts before ads or checkout flows can matter.

What is actually being announced?

Search Engine Journal’s “Selling On ChatGPT: How Brands Are Driving Revenue Through Ads & GEO [Webinar]” by Loren Baker says SEJ will go live with OpenAI, specifically ChatGPT, on Sept. 22 to discuss how businesses can sell products or drive leads on ChatGPT.

That is the concrete claim. Not a full product launch. Not a confirmed ad platform spec. Not a rate card. Not a set of ranking factors.

The interesting part is the framing. SEJ puts “ads” and “GEO” in the same sentence. GEO usually means generative engine optimization, the messy but useful shorthand for making a brand show up correctly inside AI-generated answers. That is not the same thing as buying paid placement. One is earned visibility through structure, authority, and retrievable facts. The other is distribution you pay for, assuming the platform offers it in the way marketers expect.

I would be cautious about anyone selling a finished ChatGPT growth playbook this week. The channel is still moving. User behavior is moving too. People ask ChatGPT for recommendations, comparisons, summaries, recipes, vendors, tools, and local-ish advice. But the mechanics behind what gets named, cited, ignored, or converted are not as transparent as classic Google search.

That does not make the channel fake. It makes it early.

What should brands do before chasing ChatGPT ads?

The boring work comes first.

Make sure your product pages, category pages, comparison pages, documentation, pricing pages, and support content say the same things. If your site calls a product one thing, your marketplace listing calls it another, and your PR copy adds three vague claims, AI systems have to guess. Guessing is bad marketing.

The next move is to create pages that answer buying questions directly. Not keyword-stuffed pages. Actual decision pages. Who is this for? Who is it not for? What does it cost? What integrates with it? What are the constraints? What changed this year? What should a buyer compare it against?

This is where classic SEO and AI search overlap more than people want to admit. Clear crawlable pages, consistent entities, credible mentions, useful comparisons, and strong source material still matter. Ashe runs Lucky Domains, which works on SEO and search visibility for websites at Lucky Domains.

The difference is output shape. In Google, a user sees a list and chooses where to click. In ChatGPT, the system may compress ten pages into one answer and only mention two brands. That raises the bar for clarity. If the model cannot confidently summarize what you do, you are depending on luck.

How should operators measure this without fooling themselves?

Do not start with a dashboard fantasy. Start with a question set.

Build a list of prompts your buyers would actually ask. “Best payroll software for a 40-person agency.” “Alternatives to Shopify for a subscription brand.” “What CRM works well for commercial real estate brokers?” “Which vendor handles SOC 2 evidence collection for startups?” Then test across ChatGPT and other answer engines on a schedule.

Track whether your brand appears, how it is described, whether competitors appear, what sources are implied or cited when available, and whether the answer gets key facts right. That is not perfect attribution. It is still useful.

Also track assisted conversions from content that AI systems are likely to ingest or cite: comparison pages, docs, explainers, public case studies, and category pages. If those pages are thin, no webinar will fix the funnel.

The catch most readers miss: AI search visibility is not a prompt hack. It is an information supply chain. A builder should pick one product line, clean the public facts, create three buyer-question pages, test twenty real prompts, and log what ChatGPT gets wrong. Then fix the source material. Ads may come. The groundwork is making the brand understandable enough to be recommended without one.