When a shopper asks AI which product to buy, does it choose yours?
A two-week, fixed-price review of how often ChatGPT, Gemini, Claude and Perplexity choose your products at the moment of decision, on your own site and on your retailers' product pages. You see which competitors win instead, why, and what to fix. One brand, one market, US$7,500.
Mentioned is not chosen
Shoppers now ask AI assistants and shopping agents to compare products, then question the answer over several turns before they decide. In AIVO's research, 87.3% of brands named at the first turn of a buying conversation had been replaced by a competitor by the fourth (1,427 probes, ten industries, 95% CI 85.5–88.9%; working paper).
For brands sold through retailers, the decision often turns on the retailer's product page: whether an assistant can read the claims, price and availability there, and trust them. In one client program, a premium skincare brand's AI Win-Rate rose from 38 to 58 in five months, measured on its own site and its retailer product pages. Read the case →
What the review covers
| Brand and market | One brand, one country |
| Products | Up to five hero products. Results distinguish a win for the brand from a win for a single product. |
| Where shoppers decide | Your own site and one retailer's product pages, such as Ulta, Sephora, Target or Amazon |
| Buying situations | 3 to 4 agreed with you, written as real shopper briefs |
| Claims | Claims such as organic, clean, sulfate-free or cruelty-free are tested as separate claims, never as interchangeable ones |
| AI assistants | ChatGPT, Gemini, Claude and Perplexity, run as shopping agents through full multi-turn conversations |
| Runs | Enough repeated runs per situation and assistant to report 95% confidence ranges |
| Comparison set | The competitors you name, plus any brand the assistants choose instead |
What you receive
- Your AI Win-Rate™How often each assistant chooses your brand and each hero product outright, with confidence ranges, by situation and by assistant.
- Competitor displacementWhich brands win instead, how often, and at which turn of the conversation you drop out.
- Retailer page checkWhether assistants can find, read and verify your claims, price and availability on the retailer's product pages as well as your own.
- Reasons you lose, and a fix listWhat the assistants rely on, what they cannot find or trust, and a prioritized fix for each reason, split between fixes you own and fixes that need the retailer.
- Replays, board summary and readoutStep-by-step records of representative runs, a two-page board summary, and a 60-minute readout with your team.
Ten working days from scope to readout
- Days 1–2ScopeA 30-minute call; products, retailer, situations and competitors agreed and frozen in writing.
- Days 3–7MeasurementRepeated conversations across all four assistants, with model version and web-search state logged on every turn.
- Days 8–9AnalysisDisplacement, the retailer page check, reasons lost and the fix list.
- Day 10ReadoutThe 60-minute readout, board summary and prioritized fix list.
No system access, customer data or integration is needed. The review observes only what public AI assistants say and what they can read on public pages. It measures AI recommendation, not sales; it does not change a retailer's own search ranking or sell-through. AI answers vary between runs and change as models update, which is why we report ranges and record every run. Full run logs are available under a mutual NDA.
Fixed price, credited if you continue
| Scope | Price | |
|---|---|---|
| AI Decision Risk Review | One brand, one market, up to five hero products, your site and one retailer | US$7,500 |
| Extended Review | Adds a second retailer or a second market, plus an AI agent test: can an agent acting for a shopper find the right product and add it to the basket? | US$12,500 |
Review or Sprint? The Review is the board-level read: up to five products, your site and a retailer's pages, an AI Win-Rate for each product, and a fix list split between your team and the retailer. If you need quick action on one site instead, the 30-day Sprint (US$3,500) delivers three finished fixes and a re-measure.
The full fee is credited against the 90-day Win Rate program (US$9,500 per brand and market, plus a performance fee paid only on measured lift) if you start it within 60 days of the readout. No extra charges for runs, assistants or revisions within the agreed scope. Your results are confidential. Download the prospectus (PDF) →
Agree the scope
Tell us your brand, hero products, main retailer and competitors. We reply within one business day and can start within a week. For procurement and security teams, a security and procurement brief is available on request.
Get in touch tim@aivocorp.com