Launching today
GetMint
Turn AI search into a growth channel
235 followers
Turn AI search into a growth channel
235 followers
Search is moving from Google to AI. When people ask ChatGPT, Gemini or Perplexity for a recommendation, is your brand in the answer? GetMint runs the full loop: track how AI models mention and recommend your brand, get prioritized actions from an AI agent and our experts, then publish the content that gets you mentioned. Used by L'Oréal, Accor, Lacoste and 200+ brands. Start free: one brand, 10 prompts, no credit card.
Testé ce matin, zéro prise de tête, pas de carte bancaire, tu rentres une marque et t'as la réponse en 2 min. Ça change des outils où faut réserver une démo avant de voir quoi que ce soit. Beau taf!
@philippine_dumas Merci beaucoup, ça fait super plaisir à lire 🙌 C'était exactement le but : tu dois pouvoir voir ta visibilité AI tout de suite, sans carte et sans caler une démo avant. On pense que la valeur doit se montrer en 2 min, pas après un call de vente. Et si tu veux creuser, on adore les retours, alors dis-nous ce qui te manque !
Finally tried GetMint and the prompt tracking view is genuinely useful. Seeing how ChatGPT actually describes my brand versus how I think it does is humbling.
@melihabykm5prc Thank you, so glad the prompt tracking clicked for you ! That gap between how you think you come across and how ChatGPT actually describes you humbles almost everyone. But it is the most useful thing to see, because you cannot fix a story you have never read. Now you get to shape it!
Saw our brand get cited by ChatGPT on three of the ten prompts we ran, and the action items already gave our content team a clear punch list for the week.
@mira146251 That is exactly the loop we hoped people would hit 🙌 3 out of 10 is a great starting point, and now you can see the other 7 as a to-do list instead of a mystery. That is the whole idea: turn "why am I not showing up" into a clear plan for the week. Go grab those other prompts, and thanks a lot for sharing this!
CMO here, and honestly the first user of this product ! 👋
I've spent my career doing SEO and demand gen. Over the last year I watched something shift: our best content was getting summarized by ChatGPT and Perplexity before anyone clicked through. Traffic wasn't dropping because we got worse, it was dropping because the answer now lives inside the AI.
The first time I ran our own brand through GetMint, it stung a little. We were invisible on prompts we absolutely should have owned, and a competitor was getting cited instead. But for the first time I could actually see it AND fix it. Within weeks we changed which content we prioritized and started showing up in answers. Seeing the product work for me, I had to join the GetMint team.
If you own marketing and you've felt that "where did my traffic go?" feeling, this is the tool I wish existed a year ago. Try it on your own brand ;)
This looks really really cool! how soon do you see the impact on your website ?
@michael_hodara1 Thanks! 🙌 For search-grounded engines you can see movement within days-to-weeks of publishing; the model's baked-in knowledge is slower. We re-run your prompts daily so you catch it in the dashboard as soon as it shifts.
👋 Hi hunters, CTO here !
Measuring visibility in AI search sounds simple until you realize LLM outputs are non-deterministic. Ask the same question twice and you can get two different answers, different brands cited, different ordering.
Our approach: we run each prompt at scale across models (ChatGPT, Gemini, Perplexity, and more), then treat the results statistically, measuring how consistently a brand shows up, its share of voice, and how that drifts over time, rather than one lucky or unlucky sample. On top of that we parse and normalize citations so you can see why a model mentions a competitor and not you.
The hard problems are sampling, model coverage, and keeping the data stable enough to trust week over week. That's most of what we've been building.
Happy to give more detail in the comments, ask me anything about the stack or the methodology.
the tracking half seems like the harder problem honestly, not the publishing half. these models don't give consistent answers to the same prompt run twice, so how many times do you actually query each prompt before you call a mention rate reliable? worried that with brands this size you'd need a pretty large sample per prompt per model just to avoid chasing noise.