Product concept
What you would see, and what you would do next
A customer says an assistant recommended someone else. You have a hunch about why and no way to check it. Win Answer is a planned product for that situation. Here are the three screens a commercial team would use, drawn with sample data so you can judge the idea and tell us where it’s wrong.
What a buyer is told, and what to do about it
Most teams meet this problem as an anecdote. A sales call opens with “I asked ChatGPT and it suggested someone else.” Nobody can say how often that happens, for which questions, or whether it is your product page, your reviews or plain bad luck.
The product we’re exploring would answer three things, in this order: what assistants say when a buyer asks, why a competitor gets named where you don’t, and the shortest list of changes your team could make this month.
Each screen below takes one of those. Each also says what it can’t tell you, since a tool that hides its limits is worse than having none. The long version of those limits is on the methodology page.
US · English · 10 sampled runs
What’s a long-lasting perfume under $100?
Your brand was named in
3 of 10 runs
- Competitor A8/10
- Competitor B6/10
- Competitor C4/10
- Your brand3/10
Sample set 1 of 3
Pages the answers leaned on
- Review and editorial sites 7/10
- Retailer pages 5/10
- Brand sites 2/10
Sample figures written by hand to show the layout. No AI system was queried and the brands are not real.
Screen one: Answers
See what a buyer is told
You pick a market, a language and a set of buyer questions. The screen shows what assistants said when each question was asked repeatedly: which brands were named, how often yours was among them, and which kinds of pages the answers leaned on.
- Appearance across repeated runs, shown as a range, never as one rank.
- Who else was named, and how often, for the same question.
- The sources behind the answers, grouped by type: review sites, retailers, brand sites, directories.
- Market and language filters, so a German question in Berlin is not blended with an English one in London.
Facts the sampled answers drew on
2 gaps in this sample
Shaded rows mark where only competitors have the item. The brand is invented. In a built product each cell would link to its source page.
Screen two: Competitor gaps
Understand why they were named and you were not
A competitor showing up in an answer is a symptom. The cause is often mundane. Their page gives the size, the notes and the price; yours says “a rich, luxurious blend.” Theirs is on the comparison page buyers read, and yours isn’t.
This screen lines up the facts the sampled answers drew on, brand by brand, and marks where only competitors have an item. Each cell would link to the page the fact came from, so a colleague can check it in a minute.
Proposed changes, ranked by evidence and effort
Check that search crawlers can reach the product pages
Evidence: A technical check, not taken from answers: the robots rules look stricter than intended.
- Owner
- Web
- Effort
- Small
- Confidence
- High
State notes, size and longevity on every product page
Evidence: Competitors’ pages stated these in 6 of 10 sampled answers. Yours did not.
- Owner
- Content
- Effort
- Small
- Confidence
- Moderate
Publish a fair comparison against the two products named most
Evidence: Comparison pages were among the sources in 4 of 10 runs.
- Owner
- Content
- Effort
- Medium
- Confidence
- Low
After a change, the same questions are asked again. Differences smaller than run-to-run variation are reported as noise.
Sample data written by hand. The list shows the format, not a real finding.
Screen three: Opportunity actions
Turn evidence into a short list your team can act on
Any tool can produce a list of findings. A team needs to know what to do on Monday. This screen ranks possible changes by strength of evidence and amount of work, and tags each with whoever would do it: content, catalog, web or reputation.
Every item carries its evidence, a confidence level and a recheck: the same questions asked again after the change. If a change moves the numbers by less than normal run-to-run variation, the screen would say it is within noise.
What exists today, claim by claim
Everything on this page is planned except the preview. Each claim below carries the status we can back up.
| Claim | Status | What it means |
|---|---|---|
| A clickable concept preview with sample data | Demo now | Written by hand. No AI system is queried and no real brand is measured. |
| Question sets for five industries | Planned | Drafted as examples. Real sets would be built with partners, from their own buyer conversations. |
| Repeated sampling of assistant answers | Planned | Not built. The method we would hold ourselves to is written down on the methodology page. |
| Competitor gap matrix with source links | Planned | Not built. It depends on the sampling above. |
| Ranked opportunity list | Planned | Not built. The ranking rules would be shown to you. |
| Question sets in each language | Needs evidence | We believe questions are phrased differently across languages. We have not tested whether the answers differ in a way that matters. |
| Changes you make move what assistants say | Needs evidence | Google says its AI features rely on ordinary search quality, with no special markup needed. For other assistants we do not know, and would test. |
| Linking changes to traffic or inquiries | Needs evidence | Attribution is hard here, and we don’t yet know how far it can be taken. |
| A single visibility score | Needs evidence | We don’t have a method we could defend, so there won’t be one until we do. |
The same three screens, different questions
What goes on each screen depends on how your buyers decide. Start from your industry.
- Beauty and fragranceNotes, size, longevity and where to buy decide the answer.
- Consumer electronicsSpecifications, price and availability, compared across models.
- B2B softwareUse cases, integrations and who else is on the shortlist.
- Hotels and hospitalityLocation, atmosphere and whether answers point to your own site.
- Dental and aesthetic clinicsAccurate services and locations, reviewed by clinicians.
Questions people ask
Can I try the product?
Not yet, because it is not built. You can click through the concept preview, which runs on sample data. A conversation with us is the only way to influence what gets built first.
Which assistants would it cover?
We have not chosen any. The list would depend on what can be sampled repeatably and within each provider’s terms, and we would publish it along with how each one is reached. The methodology page explains why the route matters.
Will it fix my pages for me?
No. It would point to specific gaps and suggest changes, and your team or agency would make them. Nothing in the concept touches your site.
What will it cost?
There is no price because there is no product yet. If you want to be involved early, the route is to discuss design-partner access.
How do I know whether I need this?
You may not. Read the buyer guide first and run the hand check in it. If assistants name you correctly, you have your answer for free.
Tell us which of the three you would use first.
A call about your buyers and the questions they ask. We’ll show you the screens for your industry and ask what you’d change. No automated audit and no promised rankings.
Talk to us about your buyers