Is AI recommending you,
or your competitor?
Your buyers ask ChatGPT before they ever open a search page. MentionUp measures how often AI engines actually cite your brand, and helps you change it.
AI answers are the new first page.
And you can’t see them.
Buyers get a shortlist, not ten blue links
When someone asks “best tool for X”, the answer names two or three brands. If yours isn’t one of them, the deal is over before your site gets a visit.
Nothing shows up in your analytics
An AI engine recommending your competitor leaves no trace in your dashboards. You lose quietly. There is no bounce rate for an answer you never appeared in.
But it is measurable
Ask the same questions repeatedly, count who gets cited, and the invisible becomes a number you can track and move.
What a before / after actually looks like
Not a vanity score: the actual AI answer, captured before and after, with citation frequency you can verify.
“What are the best tools for [your category]?”
Baseline (14 days)
Brand absent, cited in 0 of 3 runs
After intervention (14 days)
Cited in 2 of 3 runs, listed 2nd
Same questions, same engines, fixed 3-run protocol. We report frequency (“2 of 3 runs”), never a yes/no. AI answers vary run to run.
We scanned ourselves first
Before offering measurement to anyone else, we pointed it at our own product brand: the same fixed-query, three-run protocol we sell. That baseline is the starting line for six own-brand experiments (with control queries) starting shortly; real before/after snapshots will replace the format above as they complete.
Numbers move week to week. We show the curve, up or down, with a confidence label on every figure.
Measure → Diagnose → Execute → Re-measure
1 · Measure
A fixed set of buyer questions, asked 3 times each on every engine, every week. Same questions every time, so movement means something.
2 · Diagnose
For every query you’re missing from, we label the likely cause: content gap, structure problem, no third-party mentions, or entity mismatch. Labeled as hypotheses, not verdicts.
3 · Execute
Fix what the diagnosis points at: content, structure, and earned third-party references. No link buying, no AI-generated spam networks.
4 · Re-measure
The same query set, the same protocol, week after week. You see the curve, including the weeks it dips.
Two layers, not one score
Mentions
The answer names your brand.
Citations
The answer links your page as its source.
A brand can be name-dropped but never sourced, or sourced without being recommended. Tracking both layers separately tells you which lever to pull.
Some questions can be moved by content.
Some can’t. We tell you which, before you spend.
For every query in your report, we estimate movability: whether an AI engine’s answer to it actually responds to content and entity work, or is locked by factors you can’t touch. Budget goes to the questions that can move.
Built on our own engine-by-engine research into what makes AI answers change. This labeling is the part of the report competitors don’t have.
What we don’t promise
No guarantees, on purpose
- We don’t promise rankings or placement. AI answers are non-deterministic, and anyone promising certainty is selling you the dice roll.
- We don’t buy links, run private blog networks, or mass-publish AI content. Engines penalize it; it would put your brand at risk.
Probably not a fit if…
- You need visible results in two weeks.
- Your buyers don’t ask AI engines about your category yet. We’ll tell you that in the free scan, and you should keep your money.
What to actually expect
- First movement typically takes 2 weeks to 3 months, depending on movability.
- While we work together: weekly re-measurement, every number with a confidence label, dips included.
Find out what AI actually says about you.
Three fields, no signup. The report lands in your inbox.
Run free scan →