The AI shopper panel
Know when AI
recommends you
mapou runs a panel of AI shoppers. Every day they ask ChatGPT, Gemini, Claude and Grok what to buy, the way your customers do, and record which brands get recommended and why. Where the answer goes to a competitor, we can run the ChatGPT ads that put you back in it.
Free during early access. No card needed.
- Where you stand
- Your share of AI answers against competitors, on every assistant, every day.
- Why the AI picked them
- The sources and reasons behind each recommendation, so you know what moves it.
- What to do next
- Page fixes to earn the answer, and a ChatGPT media plan where you can only buy it.
The latest reading
Latest edition of each study. Share of AI answers that name the brand.| How often each assistant checked the live web | ||||||
|---|---|---|---|---|---|---|
| Category | Named most | Share of answers | ChatGPT | Gemini | Claude | Grok |
| Live TVFielded October 2026 | YouTube TV | 77.6% | 82% | 90% | 97% | 97% |
| Allergy medicineFielded October 2026 | Zyrtec | 66.4% | 71% | 85% | 100% | 98% |
| Credit cardsFielded October 2026 | Chase | 50% | 78% | 82% | 88% | 98% |
| StreamingFielded October 2026 | Netflix | 41.3% | 85% | 85% | 96% | 99% |
| BeautyFielded October 2026 | SkinCeuticals | 32.5% | 70% | 82% | 99% | 93% |
- Live TV77.6%Fielded October 2026. Named most: YouTube TV. Checked the web: ChatGPT 82%, Gemini 90%, Claude 97%, Grok 97%
- Allergy medicine66.4%Fielded October 2026. Named most: Zyrtec. Checked the web: ChatGPT 71%, Gemini 85%, Claude 100%, Grok 98%
- Credit cards50%Fielded October 2026. Named most: Chase. Checked the web: ChatGPT 78%, Gemini 82%, Claude 88%, Grok 98%
- Streaming41.3%Fielded October 2026. Named most: Netflix. Checked the web: ChatGPT 85%, Gemini 85%, Claude 96%, Grok 99%
- Beauty32.5%Fielded October 2026. Named most: SkinCeuticals. Checked the web: ChatGPT 70%, Gemini 82%, Claude 99%, Grok 93%
Share of answers is how often the AI named the brand. The four engine columns are about the assistant, not the brand: how often it searched the web before answering. The same question, asked of four assistants, gets four different kinds of evidence.
Where the AI reads about credit cards
cnbc.com is the most-cited source (10.7%), ahead of any issuer's own site. Chase is named in 50% of answers, yet Chase's own sites supply 7.6% of the sources. Your brand story is being told by pages you don't own.
Read the credit card studyThe mapou workspace
Put AI shoppers to work for your brand
The public panel shows your category once a month. In the workspace, AI shoppers ask about your brand and competitors across four assistants every day. You see who gets recommended and why, ask questions of the data, and plan what to fix and what to buy.
Ask mapou
Where is our brand invisible to AI right now?
Research
Your share of answers on every engine, every day
Insight
Why it moved, which sources drove it, what to fix
Activation
Where competitors win, as a ChatGPT media plan
Simplified product preview
Numbers you can take to a CFO
The same questions, every month
Buying questions are written before each run and never swapped out after the fact. When a share moves, the AI changed, not the test.
Every assistant buyers use
ChatGPT, Gemini and Claude every month, Grok and Perplexity each quarter. Deep studies add Grok to every run. Each engine counts equally.
Whole conversations, not one question
In the deep studies, each shopper has a budget and a need and keeps talking: asks, follows up, narrows down, then buys or walks away. One question in a fresh chat only sees the opening reply.
Honest error bars
Every share comes with a 95% confidence interval, and the method page lists what the panel does not claim.
Three ways in
The workspace
Your brand and competitors, checked across four assistants every day, with answers to your questions and a plan for what to fix.
Free during early access
Start freeDeep studies
Long-form reports on how AI recommends in one category: credit cards, streaming, live TV, beauty and allergy medicine.
Free and public
Browse the studiesChatGPT ads pilot
Six weeks. We use the panel to pick the buying questions worth paying for, then run the campaign in your own ad account. Live by day 14 or the fee comes back.
$7,500 fixed, ad spend separate
How the pilot worksQuestions brands ask
What is the mapou Panel?
A panel of AI shoppers. Each shopper is an agent that asks the major AI assistants a buying question the way a real customer would, reads the answer, and records which brands were recommended, which sources the assistant relied on, and what changed since the last reading. The category panel runs ChatGPT, Gemini and Claude monthly, with Grok and Perplexity each quarter. Deep studies run ChatGPT, Gemini, Claude and Grok.
Are the shoppers real people?
No. Each shopper is an AI agent holding a scripted, multi-turn buying conversation with the assistants. The panel measures what the AI recommends, not what people end up buying, so a share here is a share of AI answers, not market share.
How is this different from an AI visibility tool?
Most AI visibility tools and indexes put one question to one or two assistants in a fresh chat and count the brands in the reply. In mapou's deep studies, shoppers hold whole buying conversations with ChatGPT, Gemini, Claude and Grok, following up until they buy or walk away, so you see the brand the answer ends on. The monthly panel fixes its questions before each run and puts a confidence interval on every share, so the numbers hold up over time. Where the answer goes to a competitor, mapou runs the ChatGPT ads that put you back in it.
What do brands buy?
The mapou workspace checks your brand and competitors across ChatGPT, Gemini, Claude and Grok every day, answers questions from that data, and plans fixes and ChatGPT campaigns. It is free during early access; paid plans start at $299 a month. Deep studies are free and public. A ChatGPT ads pilot is $7,500 for six weeks, with ad spend paid to OpenAI. A custom panel runs your own shoppers, competitors and questions on the same method.