Absence is invisible
A model that never names you generates no impression, no click, and no line in any dashboard you own. The loss is real and completely unlogged.
LLM visibility & brand intelligence
Find out how AI describes your brand.
Your buyers are asking assistants which product to choose, and getting one confident answer back. mentrel measures whether that answer names you, which competitors it names instead, and what it says about you when it does.
No account, no tracking pixel, no call required to see your first findings.
Why it matters
A page of ten blue links became a single answer.
Search used to give buyers a list and let them choose. An assistant gives them a recommendation — usually three or four named products, described in the model's own words. If you are not in that shortlist, you are not being compared. You are simply absent, and nothing in your analytics will tell you.
Models repeat themes
Assistants tend to reach for the same handful of characterisations about a brand. A stale complaint about pricing or support can follow you through thousands of answers.
It is measurable
You cannot control what a model says, but you can measure it, watch it move, and act on what the sources behind it are telling the model to think.
What we measure
Four questions, answered with evidence.
Every figure in a mentrel report traces back to specific logged answers. Nothing is modelled, estimated, or extrapolated from a proxy metric.
Mention rate
Across a structured set of real buying questions in your category, how often does an assistant name your brand at all — and in what position within the answer?
Competitor share of voice
Who gets named when you don't. Which rivals the model reaches for first, how consistently, and the specific framing that puts them ahead of you.
Praise & complaint themes
The recurring characterisations models attach to you, grouped and counted — what they consistently credit you for, and the criticisms they keep repeating.
Cited sources
The pages models lean on when describing your category. This is the actionable layer: the material shaping the answer is usually material you can influence.
How it works
Measured, not predicted.
We don't score your website against a checklist of best practices and guess at the outcome. We ask the assistants and record what they actually say.
-
We build your question set
A structured set of the buying questions real customers ask in your category — comparisons, recommendations, alternatives, and objections — tailored to how your market actually shops.
-
We run them and log everything
Each question is put to the assistants and the full response captured, so every number in your report has an answer sitting behind it that you can read yourself.
-
You get findings and a plan
A written report with your mention rate, competitive picture, the themes attached to you, and concrete recommendations — ordered by what would move the needle first.
The deliverable
A report you can act on Monday.
Written for an operator, not a dashboard. Findings link to the exact evidence behind them, so you can check any claim we make rather than taking it on trust.
- Mention rate and position across your full question set
- Per-competitor teardowns showing where each one beats you and why
- Praise and complaint themes, grouped, counted, and quoted
- The sources models cite, linked to the exact pages
- Prioritised recommendations tied to specific findings
- Re-run over time so you can see whether anything you changed worked
FAQ
Questions we get asked.
What is AI visibility?
How often, and how favourably, your brand appears when someone asks an AI assistant a buying question in your category. It differs from search ranking in a way that matters: an assistant returns one synthesised answer naming a handful of options, so being in that answer is worth more than ranking on a page of links nobody reads to the bottom of.
Is this just SEO with a new name?
No. Traditional SEO optimises for position among ranked links. AI visibility is about whether a model names you at all inside a generated answer, and what characterisation it attaches to you when it does. The underlying inputs overlap — both care about what the web says about you — but what gets measured, and what you do about it, are different.
How is this measured rather than guessed?
We put a structured set of real buying questions to the assistants and record the full responses. Reporting is built from those logged answers, so every figure has evidence behind it. Runs repeat over time, which turns a snapshot into a trend you can manage.
What does a teaser report include?
Real findings for your brand — your mention rate, the competitors showing up in your place, and the themes models repeat about you. It is genuinely useful on its own. The full report adds the complaint breakdown, cited sources, per-competitor teardowns, and the prioritised recommendations.
Which assistants do you cover?
Reporting today is based on ChatGPT, which is where the majority of consumer assistant usage sits. Additional engines are being added, and because every run is logged and repeatable, brands measured now keep a comparable history as coverage widens.
How do I get started?
Email kareem@mentrel.com with your brand and your main competitors. We'll run a teaser and send the findings back — no account, no call required, and nothing to install.
Start here
See what assistants say about you.
Email your brand and a few competitors. We'll run a teaser report and send you the real findings — free, and yours to keep either way.
kareem@mentrel.com