When a shopper's agent is choosing, how often is it you?
Early accessRecommendation Monitoring
How often are you the agent's pick, and why do you lose when you lose?
A weekly measure of how often ChatGPT, Perplexity, Gemini and Claude recommend you for buy-intent prompts in your category, with the stated reason and the competitor that won instead.
Early access. We onboard a small number of accounts at a time so every prompt bank is reviewed by hand.
Pricing
Monitoring
£149/month
50 buy-intent prompts, four engines, weekly runs, alerts, and the monthly re-scan included.
Agencies
On request
Multi-client dashboards at a per-domain rate. Ask for the agency sheet.
Prices exclude VAT and sales tax where applicable. Monthly plans are billed monthly and can be cancelled any time. Refunds and cancellation.
Is this for you?
- You want to know your share of recommendation on prompts that end in a purchase, not on "what is the best" questions.
- You need an early warning when a competitor overtakes you in agent answers, with a probable cause attached.
- You have fixed your score and want to see whether the fixes moved the needle.
Probably not for you if…
- You only want mention and citation counts across informational prompts. Several visibility tools do that well and cost less.
What it measures
- Share of recommendation
- The share of buy-intent prompts, per engine, where you are the recommendation or the transaction target.
- Stated reasons
- The grounds the assistant gives for its pick, parsed into price, delivery, reviews, availability, returns and brand.
- Who wins instead
- The competitor named on each lost prompt, and the reason it won.
- Probable cause
- Each reading is stored against that week's Agent Conversion Score, so a drop in share arrives with the change that most likely caused it.
How it works
- A prompt bank built from your catalogue
50 buy-intent prompts drawn from your categories, price points and delivery promises, reviewed once by you.
- Weekly runs through four engines
OpenAI, Perplexity, Google and Anthropic APIs with search or shopping tools enabled. Each answer is parsed for brand picked, alternatives named, reason, link destination and price quoted.
- Readings tied to your score
Every reading is stored against that week's readiness score and agent funnel, so changes correlate.
- Alerts that name a cause
When share moves more than your threshold week on week, the alert names the most likely cause from the readiness data.
What you get
- Share of recommendation over time, by engine and by prompt
- The reasons you win and the reasons you lose
- The competitors who win instead, ranked by how often
- Alerts with a probable cause when share moves
- The monthly re-scan of your Agent Conversion Score, included
What it does not do
Questions
Why buy-intent prompts only?
Because that is where the sale is decided. A prompt like "buy me waterproof running shoes under $150, delivered by Friday" produces a pick; "what are the best running shoes" produces a list. We track the pick.
How is this different from Profound, Peec or Athena?
Those tools measure visibility across broad prompts and do it well. This product measures selection on purchase prompts, records the stated reason, and links each reading to your readiness score so you know what to change. It is sold alongside the score, not instead of a visibility tool.
Can we add engines or prompts?
Yes. The standard plan covers 50 prompts and four engines. Larger prompt banks and additional engines are priced per block.
Next in the suite
Hosted Agent Layer — What if the fix were infrastructure, not a project?
Join the early access list.
Tell us your site and platform. We reply within two working days with next steps and a start date.