AI visibility for agencies: numbers your clients can check
Clients ask why ChatGPT recommends their competitor. The honest answer is a count from repeated runs with the raw answers attached, not a single score, and a list of the pages to get onto next.
Does ChatGPT recommend you, or your competitor?
Paste your URL. See what ChatGPT, Perplexity and Google AI answer when your buyers ask, who they name instead of you, and which pages they read. Then we do the work to get you onto those pages.
Free scan, no signup, no card.
Who this is for
- SEO and content agencies adding AI visibility to an existing retainer.
- Freelance consultants who need evidence a client can read without trusting a black-box score.
- Agencies whose clients are small SaaS, where the starting point is usually zero.
Which questions to track
Agree the questions with the client up front, in their buyers' words, and keep them fixed so week-over-week numbers mean something.
- Best [category] for [client's buyer]
- [Client's main competitor] alternatives
- [Category] tools compared
- Is [client] any good? (branded, as a control)
What to fix
- 01
Show every answer, not a score
Each check keeps the full text of every AI answer. A client who disputes a number can read what ChatGPT actually said.
- 02
Report runs, not single answers
In our test only 28 to 29% of brands appeared in every run of the same question. Reporting "named in 3 of 4 runs" stops a client reading noise as a win or a loss.
- 03
Turn cited pages into the work plan
The pages AI cites for the client's questions, and which of them name rivals, become the outreach and listing plan for the month.
- 04
Fix the technical blockers first
robots.txt, firewall and rendering problems come with the exact code or setting to change, which is an easy first-week deliverable.
Each engine cites different places: ChatGPT, Perplexity, Gemini, Google AI Mode.
What welaunch does not do
- One product per account today. For several clients, run one account per client. There is no multi-client dashboard or white-label report yet.
- Agents draft; someone presses submit from the client's own accounts. We do not post on anyone's behalf.
Pricing
Free
$0
- 3 buyer questions asked on 3 AI engines, twice each
- The technical audit: can AI read your site
- The top 6 pages AI cites for your category
Tracker
$49 a month
- A full re-check every week: every question, every engine, 4 runs each
- Every AI answer in full, not a score
- The full map of pages AI cites for your buyers' questions
- 3 competitors tracked next to you
Growth
$99 a month
- Everything in Tracker
- 300 credits a month for the agents that do the fixes
- A work queue of the cited pages you are not on yet
- The promise: if AI does not name you by day 90, the months after are free until it does, up to 3
Not named by day 90? The following months are free until you are, up to 3. No overage charges. If you run out of credits, work waits for next month or a top-up you choose. Cancel any time. Full pricing.
Questions
How do agencies measure AI visibility?
Ask each engine the client's buyer questions several times, count the answers that name the client, and keep the answers. Tracker does this weekly across five engines, 4 runs per question (2 on Claude).
Can I share results with a client?
Yes. Every answer is kept in full, so you can show the client the exact text, and a product can opt in to a public proof page with its named counts over time.
Can I manage several clients in one account?
Not yet. Each account tracks one product, so agencies use one account per client today.
Is there a white-label report?
No. What exists is the raw answers, the run counts, the cited pages and the audit, all readable in the dashboard.
Which engines are tracked?
ChatGPT, Perplexity, Gemini, Google AI Mode and Claude, every week.
Why not report a single visibility score?
Because it moves on its own. The same prompt asked 4 times named a different set of brands most of the time in our test. A score from one run invites the question "is this real?", and the answer is often no.
Can the fixes be done inside welaunch?
Agents draft directory listings, replies in cited threads, outreach emails and comparison pages on credits. Growth includes credits each month; on Tracker you buy packs if you want them.
Can I pull the data into my own tools?
Yes, through the MCP server. See the MCP page for the tools it exposes.
Also
Does ChatGPT recommend you, or your competitor?
Paste your URL. See what ChatGPT, Perplexity and Google AI answer when your buyers ask, who they name instead of you, and which pages they read. Then we do the work to get you onto those pages.
Free scan, no signup, no card.
welaunch started as a launch planner: where to post, which directories to submit to, what to write. Then I asked ChatGPT, Perplexity, Gemini and Google AI Mode the questions our buyers ask. Our own site was named in 0 of 203 answers.
If a launch does not get you into those answers, people asking AI for a tool like yours never hear of you. So welaunch now measures where AI names you, and does the work to change it. We run it on ourselves first, and publish our own number.
Baris, founder of welaunch.sh