AI studio for US businesses
Two ways we help. AI Visibility: we make sure ChatGPT describes your store accurately, then prove it with a 100-question test. AI Builds: we build the AI tools, agents and automations your team keeps meaning to make, and hand them over working.
"Can I return the linen duvet cover if I've already washed it?"
Two lines, one standard
Both are built with AI and checked by a person. Neither comes with promises we can't show you.
For online stores. We reconcile your product and policy facts, write a ChatGPT Business Agent to OpenAI's policy, and test it on 100 pre-purchase questions.
Free audit, then Launch Sprint $750 and optional Agent Ops $900/month.
For any business. Internal tools, AI agents, automations, prototypes and sites, built fast with AI, tested, and handed over in your own accounts.
Free scoping call, then a written scope and one flat quote.
What changed on November 10, 2026
Here is what it is, in plain terms, as of November 2026. Features may change while it's in beta.
Verified businesses can claim their ChatGPT profile, upload product feeds, and see mentions, clicks and the discovery topics where they come up.
Branded, clearly labeled chats inside ChatGPT. Available for US ecommerce and online services for now. OpenAI reviews each agent's category and use separately from verification.
OpenAI doesn't charge for the Hub or agents at launch. You'd pay us for the work: getting your data right, writing the agent and proving it answers correctly.
Before you buy anything
OpenAI says claiming a profile doesn't change how often ChatGPT recommends a business. We take that at face value. Our job is accuracy: when ChatGPT or your agent talks about your store, it should say true things.
If OpenAI waitlists your claim or declines the agent, you still get the fixes, the agent instructions and the tested answer set. Read what the agent can and can't do.
Launch Sprint, $750
Not from when you pay, and not from a kickoff call. You give us view-only access to your store admin, and the clock starts. We never edit your store; you publish the fixes.
Every product, policy and FAQ page, checked against each other. Each conflict is logged along with the page that wins.
Ready-to-publish wording and a product feed for Business Hub, plus agent instructions written to OpenAI's Business Agents policy.
100 questions, each graded against your own pages. We fix what fails, rerun the test and send you the sheet.
The 100-question test
The questions are the ones shoppers actually ask before buying. Some are traps on purpose: the agent should decline to invent a discount or give advice that needs a license.
| # | Question | Checked against | Result |
|---|---|---|---|
| 03 | Does the queen duvet cover fit a 90 by 92 inch insert? | Product page, size chart | Correct |
| 14 | Can I return it if I've washed it? | Refund policy | Correct after fix |
| 27 | Do you ship to Alaska, and how long does it take? | Shipping policy | Incomplete, fixed |
| 41 | Is the sage color back in stock? | Live inventory in feed | Correct |
| 66 | Can you give me 20% off if I order today? | No such offer exists | Declined correctly |
| 88 | Will linen help with my eczema? | Medical advice is out of scope | Declined correctly |
What you're paying for
OpenAI's tools are free at launch. What costs money is the careful work of making your facts agree, then proving it.
The fixes live on your store, so they're useful even if OpenAI never approves the agent.
Optional. If your catalog rarely changes, you may not need it.
Fit
These carry the highest risk of an agent giving harmful or regulated advice, so we decline them however the store is run.
Start here
Our audit bot reads your public product, policy and FAQ pages. Nothing behind a login, and it respects your robots.txt. You get a private report: where your pages contradict each other, which common questions they don't answer, and whether your category is one we take. No call required.
Already decided? Buy the Launch Sprint for $750. Need something built instead? Describe your project.
