About Mirr
Your next customer won't Google you. They'll ask.
Every day, millions of people ask ChatGPT, Perplexity, Claude and Gemini about brands, products and services. The answers shape buying decisions at scale. Most brands have no idea what AI is saying about them.
01Why now
Why this matters now
Search gave you a results page: ten links, ten chances to be seen. Assistants collapse that page into a single answer. One or two brands get named. Everyone else is simply absent, and there is no page two.
Ten results
One answer
Ten blue links gave everyone a chance to be found. One synthesized answer doesn't.
02The audit
What Mirr actually checks
03GEO
SEO got you ranked. It won't get you recommended.
SEO optimizes a results page: rank higher, win more clicks. That discipline still matters, but it stops at the edge of the search box.
Assistants don't serve pages. They synthesize one answer from training data and live retrieval, and they decide which brands belong in it. Ranking fourth on Google means nothing to a model that only names two companies.
The levers are different: entity presence (Wikipedia, structured data), consistent positioning language across third-party sources, and corroboration on review platforms. That discipline is GEO, Generative Engine Optimisation, and the Mirr audit is its measurement layer.
The SEO lever
The GEO levers
SEO pulls one page up a ranking. GEO makes you present, consistent and citable on the sources AI actually reads.
The studio
Built by Jack & AI
Mirr is a product by Jack & AI, an AI agency in Amsterdam. The idea came from auditing our own clients: when we asked what ChatGPT was saying about their brand, nobody could answer. Not the brand team, not the agency, not the CMO.
We work with brands like Action, Quooker and Jägermeister to integrate AI into their marketing, operations and brand strategy. Mirr turns the blind spot we kept finding into something you can measure, fix and track.
04Verifiable
Built to be checked
A score you can't verify is just an opinion. Four reasons you can take this one seriously:
- 01The web-presence stage is deterministic: 9 signals, checked yes or no. No AI inference involved.
- 02Every report is validated against a versioned contract before delivery: always 6 identity-gap dimensions, always exactly 7 actions. Not 6, not 15.
- 03Every prompt and every formula is published on the methodology page. You can re-run our questions yourself.
- 04LLMs are non-deterministic: the same question can get a different answer tomorrow. That's why Mirr works in quarterly cycles, not one-off snapshots.
One audit is a snapshot. The cycle shows whether your actions actually moved how AI describes you.
Curious what AI says about your brand?
The preview is free and takes about 20 seconds.