Human-verified AI visibility, for UK agencies
AnswerLift asks your client’s buyer questions of the models behind ChatGPT, Claude, Gemini and Perplexity, several times each, and stores every answer word for word with its sources. A named reviewer signs off each record before your client sees it. No scores and no estimates: just how often the brand was named, out of how many answers, and when.
Said up front: answers come from the providers’ APIs through the OpenRouter gateway, not from the consumer apps; Gemini is asked without Google Search grounding; Google AI Overviews are checked by hand, not collected. The methodology has the detail.
The problem
A screenshot of one answer proves very little, and a percentage built on three answers proves less. What holds up in front of a client is a count with its denominator and its dates, and the stored answers behind it. Why AI answers change every time.
66%
<1 in 100
How it works
Your client's buyer questions, written without naming the brand, are put to the models behind ChatGPT, Claude, Gemini and Perplexity through the providers' APIs, reached through the OpenRouter gateway: the OpenAI, Anthropic and Perplexity models searching the web and Gemini answering from model knowledge without Google Search grounding. On a Sprint each question is asked 3 times of each engine, because answers change from one run to the next.
Every answer is stored word for word with the exact model, the search method, the sources its search cited and the time. Where the consumer product matters, a person checks the ChatGPT app, Perplexity or Google AI Overviews by hand and records what they saw. Nothing is scraped.
A named reviewer reads each stored answer, confirms or corrects how it was classified, and signs the record off. Pending records are excluded from the client report by the query itself, not by the page.
Your client gets a read-only link under your logo: how often the brand was named, out of how many answers, by engine and by question, which sites the engines cited instead, and every signed-off record behind the numbers. It expires after 30 days, and the ledger exports to CSV.
The evidence gate
Every observation, whether a person typed it or an API returned it, starts pending. A reviewer reads it against the stored answer text, puts their name to it and marks it verified or needs-review. On a Sprint the reviewer is an AnswerLift reviewer who joins your workspace as a visible member; in your own workspace it is your team. Only verified records are returned to the client report.
Ask the same question again next month and it is stored as a new version beside the old one. Nothing overwrites, nothing is inferred, and a sign-off never rewrites the answer it was made about.
Recorded
Question, engine, model, search method, cited sources, times
Counted
Named in k of n answers, by engine and question, with dates
Signed off
Named reviewer and date; only then client-visible
Pricing
The workspace software is free with no subscription. Prices are in pounds sterling and are the total payable: no VAT is added.
Free
One brand, three buyer questions that do not name it, put to the models behind ChatGPT, Claude, Gemini and Perplexity. A person checks every answer; you get what each engine said and which sources it cited.
Request a Snapshot£1,250 · one-off, by card or invoice
A smaller first measurement for one brand: enough repeated answers to see a pattern, with every record signed off. Credited in full against a Sprint bought within 30 days.
Request by invoice£2,750 · one-time
The full engagement: up to 50 questions, every one asked 3 times of every engine on the panel, hand checks where they matter, a named reviewer on every record, a prioritised backlog and one client report. Delivered within 20 business days of kickoff.
See Sprint scope£345 per brand per month, by invoice
Ongoing evidence for one brand: the same questions asked every week, counted as named in k of n answers.
Request by invoiceFor agencies
The report goes out under your name, with a reviewer and a date on every record, so it holds up when your client asks how you know.
Explore the productYour logo and accent colour on every client link, the client's own logo beside it, and a print layout that saves cleanly to PDF.
Each client brand is its own audit in your workspace, with its questions, ledger, backlog and report. Invite your team at no cost per seat.
Ten sector templates of fifteen unbranded buyer questions each, versioned, to start a prompt set from.
Every record exports to CSV with the reviewer and every timestamp, for your own reporting or archive.
Straight answers
More in the guides, including the ChatGPT API versus the app.
