Can You Sell on ChatGPT? What Works for Shops and Service Businesses Now

The short answer, before the details change again

For product sellers, the realistic version of "selling on ChatGPT" today is mostly about being found and recommended inside a chat, then handing the customer off to complete the purchase — either on your own site or, in some cases, through a checkout flow that lives closer to the platform you already sell on. Whether that handoff happens entirely inside the chat window or redirects the shopper out to a webpage has been the single most-changed detail in this space, so treat any specific claim about "buy without leaving ChatGPT" as something to double-check against current platform documentation before you plan around it.

For service businesses — contractors, salons, clinics, agencies, anyone who sells an appointment or a quote rather than a shippable item — the more accurate and more useful framing is different: an AI assistant can describe what you offer, quote a real price if it has access to one, and pass along a request. That request is best understood as a booking or lead landing in your inbox for a human to confirm, not an autonomous payment the AI completes on your behalf. Tools built for exactly this, like Tailwind's "Add Your Business to AI," are designed to route orders from AI assistants straight to your inbox rather than hand over payment authority, which is both more accurate about what the technology does today and a safer thing to promise your own customers. That's not a lesser version of the technology — it's the version that actually matches how most service purchases work anyway, since a haircut or a roof quote isn't a one-click cart checkout even on a normal website.

What happened to in-chat checkout

Earlier in ChatGPT's push into commerce, OpenAI tested a feature that let shoppers complete a purchase without leaving the chat window, aimed mainly at ecommerce retailers. Reporting through 2026 suggests that approach didn't scale the way it was pitched — merchant adoption stayed limited, and OpenAI shifted its emphasis toward a model where ChatGPT handles discovery (showing products, prices, and comparisons pulled from a merchant's real catalog) and then sends the shopper to the retailer's own site to actually pay. Large retailers reportedly kept feeding live product data into ChatGPT under this model even as the pure in-chat checkout feature lost priority.

Part of why the pure in-chat version struggled seems to be mundane rather than dramatic: fewer merchants actually turned it on than the early coverage implied, and the shoppers who did try it often still wanted the familiar steps of a retailer's own checkout — saved payment methods, loyalty points, order history, return policies — none of which an isolated in-chat flow replicated well on day one. That's a solvable problem over time, but it's a reminder that "technically possible" and "actually adopted at scale" are two different milestones, and this category is still closer to the first one than the second.

The practical lesson isn't "in-chat checkout is dead" or "in-chat checkout is the future" — it's that the specific mechanic is unsettled enough that building your sales strategy around one exact implementation is risky. What hasn't changed, and is unlikely to change, is the underlying requirement: an AI assistant can only recommend and quote your business accurately if it has access to a real, structured, current list of what you sell and what it costs. That part is worth getting right regardless of which checkout mechanic wins out this quarter.

What actually works today if you sell products

If you run a product business, the version of this worth planning around right now is discovery plus handoff, not a guaranteed in-chat sale. A shopper asks ChatGPT (or Claude, or another assistant) for a recommendation, the assistant pulls from product data merchants have made available to it, and the shopper either taps through to buy on the retailer's site or, where a given platform's checkout integration exists and applies to your store, possibly completes the purchase closer to the chat itself. Shopify's merchants, for example, have been folded into a broader "agentic storefront" push that makes participating stores discoverable across several AI assistants by default, which is a different mechanism than a business manually listing itself — so what applies to you can depend on which ecommerce platform you're already running on.

If you run a WooCommerce store, the setup looks different again, and it's worth treating as its own project rather than assuming it mirrors Shopify's approach. We've written a full walkthrough of connecting a WooCommerce store to ChatGPT and Claude if that's your platform — the short version is that WooCommerce has its own path into AI assistants, it just isn't automatic in the way some competing platforms now are.

What actually works today if you sell a service

Service businesses should stop waiting for "checkout inside ChatGPT" to apply to them, because it mostly doesn't — and arguably shouldn't. Nobody wants an AI assistant autonomously charging a card for a four-hour electrical job or a custom cake order before a human has confirmed availability, scope, and price. The useful version of AI-assisted selling for a service business is narrower and, frankly, safer to promise customers: an assistant can answer "how much does X cost" with your real pricing instead of a guess, and it can turn interest into a lead or booking request that reaches you directly.

That's the model behind Tailwind's "Add Your Business to AI" tool, which publishes a structured, accurate list of your actual services and prices so an assistant can quote them correctly instead of inventing numbers, then passes real requests straight to your inbox — or into a connected WooCommerce store for product businesses — for you or your staff to act on. It doesn't hand over payment authority to an AI, and it doesn't promise a sale will close itself; it closes the much more solvable gap, which is making sure the assistant has correct information to work from in the first place.

Why "it depends on your platform" is the real rule

Across both products and services, the pattern holding steady is that what you can do with ChatGPT or Claude depends more on which commerce platform or connection method you're using than on some universal "ChatGPT selling feature." A Shopify store, a WooCommerce store, a service business with no ecommerce platform at all, and an enterprise retailer with a direct integration are all living under meaningfully different rules right now, and those rules have each changed at least once in the last year. That fragmentation is itself likely to simplify over time as the underlying protocols standardize, but "likely to simplify eventually" isn't something to build a Q1 sales plan on.

How to get your business in front of AI assistants without guessing

Rather than chasing the exact checkout mechanic of the month, the more durable move is making sure your catalog and pricing are accurate and available wherever an assistant might look for them. A few things are worth doing regardless of how the checkout layer evolves:

• Keep your real prices current and visible. An assistant that can't find accurate pricing will either skip you or guess, and a wrong quote is worse for your reputation than no quote at all.
• Make your service or product list structured, not just prose. A paragraph description on a website is harder for an AI system to parse reliably than a structured catalog built for that purpose.
• Decide where a lead or order should land before you need it. Whether that's an inbox, a booking calendar, or a connected store, know the destination before an AI assistant starts sending you real inquiries.
• Recheck platform-specific mechanics periodically. Given how often checkout and discovery features have changed across ChatGPT, Claude, and the major ecommerce platforms, treat any specific claim about what's possible "inside the chat" as time-stamped, and verify it against current documentation before promising it to customers or investors.

Common questions about selling on ChatGPT

Can you actually sell products directly inside ChatGPT right now?

As of this writing, the more reliable pattern is discovery and recommendation inside the chat with the purchase completed on the retailer's own site, though some platforms and retailers have closer integrations than others. This is exactly the kind of detail that's worth confirming against current OpenAI and platform documentation before you plan a launch around it, since it has shifted before.

Is "selling on ChatGPT" different for a service business than a product business?

Yes, and the difference matters more than most coverage of this topic admits. Product sales can plausibly move toward some form of in-chat or near-chat checkout over time. Service sales — bookings, quotes, consults — are a request-and-confirm process even in the best-case scenario, so the realistic goal is accurate quoting plus a lead landing in your inbox, not an AI completing the transaction unattended.

Do I need a Shopify or WooCommerce store to show up in ChatGPT at all?

No. Ecommerce platform integrations are one path, particularly for product catalogs, but a business without any ecommerce platform — including most service businesses — can still make its real services and pricing available to AI assistants through a dedicated connection built for that purpose, without needing a storefront at all.

What should I do if I don't want to keep tracking every platform change?

Focus on the part that doesn't change regardless of which checkout mechanism wins: keeping an accurate, structured, current catalog of what you sell and what it costs, available to whichever assistants customers are already using. That's the foundation every version of this technology has needed so far, and it's reasonable to expect it'll keep being needed even as the surface-level mechanics keep shifting.

Will this be different again in six months?

Almost certainly, at least at the level of specific mechanics — which assistants support which checkout flows, which ecommerce platforms are folded in by default, and what the submission or approval process looks like for a merchant who wants deeper integration. That's been the pattern for every quarter since this category started moving, and there's no strong reason to expect it to settle soon. What's much less likely to change is the underlying shape of the opportunity: businesses with accurate, accessible pricing and catalog data get recommended and quoted correctly; businesses without it get skipped or misquoted, no matter which specific checkout feature is in fashion that month.

Putting it together

The specifics of in-chat checkout have changed before and will likely change again, so the safest plan is the one that doesn't depend on any single version of it. For products, that means staying visible and accurate wherever assistants look, on whichever platform you sell through. For services, it means quoting correctly and letting real requests reach a human who can confirm them — which is both the more accurate way to describe this technology today and the safer way to promise it to your own customers.