AI agents now shop for your customers, and they skip whatever they can't read. Emtor helps e-commerce brands get picked.
People now ask an AI agent, and it compares and buys. So your store's real audience is a machine that reads.
Adobe Analytics, via TechCrunch, Apr 2026. US retail only.
AI traffic to US retail sites, Q1 2026, year over year.
You see a serum. The agent sees a form with holes in it, and it only buys what it can fill in. Write in the blanks and the end of the strip changes.
You see a serum.
ExampleIt sees a form.
Halden No. 4 Night Serum
It drops what's blank.
It buys from someone else.
Four of them, in this order. We get your product ready to answer every one, and then we check how it did.
What are you, exactly?
ReadSays who?
TrustFriday? Returns?
BuyWho got picked?
TrackSame question, same product, different reader. Each agent reads your store a little differently, and Emtor covers them all.
“A gentle night serum for sensitive skin, shipped to Germany?”
ChatGPTHalden No. 4. It lists what's in it.
Gemini and Google AI ModeA few fit. Halden No. 4 is one.
PerplexityHalden No. 4, says its product page.
Microsoft CopilotCompared a few. Halden No. 4 fits.
ClaudeHalden No. 4. It states its skin type.
Amazon RufusHalden No. 4 delivers to Germany.
You can't be picked from a list you aren't on. SEO puts you there.
Emtor is the next generation of SEO. It builds on what your SEO already does, and takes you from the list to the pick.
An agent doesn't hand over a list to browse. It names a pick.
| SEO | Emtor | |
|---|---|---|
| What's optimised | Content SEO tunes pages for people. Your pages still matter, because people still read them. | Data Emtor tunes data for agents. An agent reads facts it can check. |
| What goes in | Keywords Your keyword research still tells you what people ask for. | Questions People ask an agent in plain language, so we start from their questions. |
| The win | Ranked You can't be picked from a list you aren't on. SEO puts you there. | Picked An agent doesn't hand over a list to browse. It names a pick. |
| The moment | Clicked Some buyers will still click through. Keep the site ready for them. | Decided The purchase can get decided in the chat, before anyone opens your site. |
| What you measure | Traffic Keep counting traffic. It still tells you who came. | Named Count recommendations next to it. That's the new number. |
Scroll down the aisle. Every sign that rolls in does something to the product beside it.
Doesn't sag, not even full.
Waterproof, says the blog.
Straps hold up.
Picked Asked again Also named: 2We ask the agents real buying questions and see where you are. Then we read your store the way they do.
Then we fix what tripped them up: product data, pages and feeds.
Reviews agents can read, and mentions on the sources they already cite.
We keep asking and watching what gets picked. Agents change, so this stage never closes.
buying questions name Halden. First pick: 0 of 50.Example
Halden is a made-up brand and every number here is an example, not a real result. Real results go here once we have them.
I can't confirm what's in it.Example
Preview with sample data. The real check is free.
Can it read you?
Can it trust you?
Can it buy from you?
How does it end?
That was a sample.
Eight questions, straight from the shelf. Open a bubble and you get a straight answer back.
It is the work of making your store easy for AI shopping agents to read, trust and buy from. When someone asks ChatGPT or Gemini to find them a product, the agent does the comparing. We fix what it checks, so your products stay in the running.
Keep your SEO. It gets a page in front of a person, who then decides. An agent reads for that person and often decides what makes the shortlist, so it needs complete product data, clear policies and a way to buy. The two work together, and neither replaces the other.
ChatGPT, Gemini and other assistants that recommend or buy products. They don't all read stores the same way and they change often, so we start with the ones your customers are likely to use, and we tell you when that list changes.
We track it. We run the kinds of requests your customers make, then record whether your products appear, how they are described and who gets named instead. We also look at AI traffic and orders in your analytics. Not every agent purchase can be traced, and we'll say so when one can't. Every step is on our How we measure page.
It depends on your catalogue and your starting point. A small, tidy store and a large one with patchy data are different jobs, so any number we gave you now would be a guess. We'll tell you after the audit, once we have seen what's there.
Often you don't, at least to begin with. Much of the work is product data, content and policies, and on Shopify a lot of that can be changed without code. Some fixes do need a developer. The audit shows which ones, before anyone starts.
Access to your catalogue or product feed, a look at your analytics, and someone who can answer questions about your products. Some facts, like ingredients, materials or where you ship, only you can confirm. We won't guess them.
Yes. Ads buy a placement for as long as you pay. This work is on what an agent reads when it chooses for itself: your data, your policies, your checkout. If AI platforms offer ads you want to run, that is a separate decision.
We look at your store the way an agent does, and show you what it skips.
Your turn.