Press & media kit

Everything you need to write about GenSig accurately. Questions, interviews or the raw research dataset: [email protected].

What GenSig is

GenSig is an AI visibility monitor and GEO consultant in one. It measures how often AI assistants recommend a brand when its category gets asked, against its competitors, across a 33-surface catalog: ChatGPT, Gemini, Perplexity, Claude, Grok and Mistral through their official APIs with live web search, and the 27 surfaces without an API — Google AI Overviews, Meta AI on WhatsApp, Copilot, Siri and more — measured by real people in the real apps, with screenshot evidence and a chain of custody. It never scrapes. Unlike tools that report an opaque score, GenSig ties every finding to measured evidence from real AI answers, plus a prioritized action plan to earn the mention, and publishes reproducible methodology and raw research data. Founded in 2026.

One line: GenSig measures how often AI assistants recommend your brand — by official API and by real people with screenshot evidence, never scraping — and gives a prioritized plan to improve it.

Fast facts

CategoryAI visibility · GEO
Founded2026
Websitegensig.app
Surfaces measured33 — 6 by official API · 27 human-measured
PricingPro $49/mo · Agency $149/mo

The research angle

GenSig published a reproducibility study — 2,500 measured AI answers plus 30 consumer-app sessions — with these headline findings:

A bare LLM API correlated just ρ = 0.23 with what real ChatGPT web sessions recommend — it missed the brands users actually see and inflated stale ones.

Determinism isn't accuracy: temperature 0 made answers highly repeatable — of the wrong brand set.

There is no single "what the AI says": ChatGPT and Gemini disagree, so each must be measured through its own live channel.

Full study, methodology and downloadable dataset: How stable are AI brand recommendations?

What makes it different

Brand assets

Please use the name as GenSig (one word, capital G and S) and link to gensig.app.

Writing a piece? Email [email protected] — happy to share the raw dataset, walk through the method, or answer questions.