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 — ChatGPT with web search and Gemini with Google Search grounding — recommend a brand when its category gets asked, against its competitors. Unlike tools that report an opaque score, GenSig ties every finding to measured evidence from real AI answers: the exact questions, sources and competitors behind each result, plus a prioritized action plan to earn the mention. It focuses on the assistants buyers actually use rather than inflating a list of model integrations, and publishes reproducible methodology and raw research data. Founded in 2026.
One line: GenSig measures how often ChatGPT and Gemini recommend your brand, shows the evidence behind every result, and gives a prioritized plan to improve it.
Fast facts
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
- Measures the live, web-grounded surfaces real buyers use — not a bare model API.
- Evidence-backed, not a black box: every recommendation ties to real answers, sources and competitors.
- Focused on the assistants that matter, not an inflated list of model integrations.
- Reproducible: methodology and raw research data are published.
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.