How does AI search talk about your business?
Recurring AI-visibility audit for fine-wine merchants. We probe five LLMs across reputation, catalog, value, trust, intent, and controversies, then deliver a curated report tailored to your business and watch how the answers shift week over week.
5 LLMs · weekly cadence · curated by humans
How it works
An observatory, then an action plan.
Weekly probe matrix
Every week we query 5 leading AI models with the brand-probe matrix tuned to your business, your bottles, your competitors, your region, across raw and grounded modes.
Human curation
Our editor reads the AI-generated summary before it lands on your dashboard. Tone, accuracy, and actionable framing all get a hands-on pass.
Curated delivery
You see a tone-weighted standing per axis, the chip cloud of what AIs say about you, and a printable booklet PDF, refreshed every week, evolving over time.
What we measure
Four dimensions. No fluff.
Standing & recognition
How AIs talk about you, tone-weighted across 5 providers and 6 axes (Reputation, Catalog, Value, Trust, Intent, Controversies). Plain recognition rate as the secondary signal so coverage and quality are read together.
Sentiment
Five-level tone scale per cell, judged against your business's context. We surface the exact AI quotes that work for and against you, and the ones worth reviewing live.
Catalog accuracy
Do the producers AI cites actually match what you carry? Hits, misses, and hallucinations cross-checked against your declared catalog and the bottles you flagged at onboarding.
Site readiness
Technical audit of the AI-crawler-facing layer of your site: robots.txt, llms.txt, Schema.org, Product markup, Open Graph, editorial depth. A-F grade with prioritized fixes.
What you'll see
A real-time read of how AIs talk about your business.
Below: a sample audit's headline panels, side-by-side. Tone-weighted standing across 5 LLMs and 6 axes; binary recognition rate as the secondary signal; the chip cloud is what AIs actually say about you, ranked by frequency.
AIs talked about Domaine Sample positively across 5 providers (41% recognition rate). Tone-weighted standing index: 74 / 100.
- OpenAI+3.1
- Anthropic+2.5
- Google+2.2
- old-school cellar12
- trusted by sommeliers9
- Burgundy specialist8
- premium prices6
- limited allocations5
- responsive shipping4
- narrow whites range3
Each chip is a recurring theme judged from the AI responses; the number is how often it came up.
Synthetic data, your audit will look the same shape with your numbers.
Onboarding
Tailored to your business, priced on request.
We onboard you ourselves.
Once you submit the form, we tune the probe bank to your business, the bottles you flagged, the competitors you named, the regions you sell. Your first audit ships within the week.
Quote tailored to your scope.
We don't publish tiers, every house has a different number of bottles, regions, and competitors to track. Tell us what matters to you and we'll quote a single subscription.
No support ticketing.
Every account has a direct email channel to the team. You ask, we answer in hours, not days.
FAQ
Questions we get asked.
How do you query AI models without breaching their ToS?
Only through the editors' official APIs (OpenAI, Anthropic, Google, Mistral, Perplexity). We never scrape consumer chat interfaces, it's technically fragile and contrary to the terms of service.
Models change constantly. How can trends be trusted?
We log the exact model version (e.g. gpt-4o-2025-03-15) on every call. All charts are annotated on version switches, and we systematically segment analyses so model drift is never confused with market movement.
What's the difference with Profound, Peec AI, or Semrush AIO?
We specialise in fine wine. The probe bank is tuned to merchants and houses (regions, allocations, comparisons, vintage chatter), the producer dictionary is hand-curated, and the human curation pass at the end of every audit is the differentiator. Generalists don't go that deep.
Which models do you cover exactly?
The top five: OpenAI (GPT-4o, GPT-4o-mini, GPT-4o-search), Anthropic (Claude Opus/Sonnet), Google (Gemini Pro, Gemini grounded), Mistral (Le Chat, Mistral Large), and Perplexity (sonar-online). We'll add others as your customers actually start querying them.
How long until my first audit?
Once you submit the onboarding form, our team reviews your business's setup (1-2 working days). Your first audit then runs on the next weekly cycle and lands on your dashboard once we've curated the summary, typically the Monday after.
How is pricing structured?
Single subscription, priced on request. The quote reflects what we track for you, the breadth of bottles, competitors, and regions you flagged at onboarding, so we don't publish tiers. Get in touch via the form below for a tailored proposal.