Citable - Make brands show up in Agentic Search
by•
Citable helps brands show up in AI answers. We simulate real personas by pre-warming AI accounts (ChatGPT, Perplexity, Gemini, Claude, Grok) with memory, then run thousands of prompts daily to measure share-of-voice and capture every cited source. We surface gaps by persona, market, and model—and explain why. Finally, we give prescriptive actions: publish specific articles, join targeted Reddit threads, and prioritize the highest-leverage plays for your team.
Replies
Maritime
@maria_gorskikh1 Curious how the pre-warming personas work in practice across different AI models.
@maria_gorskikh1 @masump Thanks Masum! We spin up long-lived accounts on each model and ‘pre-warm’ them with a chat history that matches a real segment (role, intent, location, interests). Then we run the same query set through those personas across ChatGPT/Perplexity/Gemini/Grok/Claude and compare how often each brand gets cited. That way you see model + persona level lift, not just one-off prompts.
Swytchcode
For a brand, showing up in AI answers is really of immense importance. Love what citable is doing
Congrats on the launch!
@chilarai Thank you, Chilarai 🙏 Exactly our thesis: if you’re invisible in AI answers, you’re invisible to a big chunk of future demand...
minimalist phone: creating folders
Cool, maybe include this in one of my next newsletter articles, cause it will be pretty relevant.
@busmark_w_nika That would be amazing, thanks Nika! Happy to share a couple of screenshots, examples, or a short GEO explainer if that’s helpful for the newsletter—DM me or email maria@citable.xyz and I’ll send over a mini pack.
minimalist phone: creating folders
@maria_gorskikh Hey Maria, could you please send me them at businessmarketingwithnika@gmail.com ?
Really interesting concept! But how accurate can simulated personas be when capturing how real users interact with models linke chatgpt or gemini?
@ludovicmartin97 Totally fair question. We don’t claim to perfectly clone every user; we build calibrated ‘segment personas’ (e.g. US dev lead, EU solo founder, etc.) based on real search behavior + customer data, then keep their history consistent across runs. The goal is directional insight by persona/market (who gets mentioned, where, and why), not pixel-perfect user simulation
Gfycat: GIFs for Gmail
Very exciting to see this launch. Given how users' habits are changing from 10 blue links, GEO will be very important as websites adapt to a new world dominated by discovery through LLMs
@richardrabbat Thanks Richard 🙏
What’s been surprising already is how different the winners are vs classic SEO. Some brands that own Google barely show up in AI answers, while smaller ones get recommended nonstop because they’re better represented in the model’s ‘memory’—docs, forums, APIs, long-tail content.
GEO for us is about mapping that trust graph and giving teams a concrete way to shape it, instead of assuming search rankings will carry over.
Citable
The dashboard is extensive with AI visibility, competitor ranking, personas and sources. Collectively help brands to understand how can they improve their AI SEO
Maritime
@yiko_li Adding one more thing here – the dashboard also shows which pages/posts are actually feeding those AI answers by persona + market. So you can see ‘these 5 URLs are doing all the work for US SMB founders on ChatGPT’ and decide what to double-down on or rewrite
The depth of analysis across persona, market, and model feels built for people who actually care about growth.
Maritime
@isabella_wandrei Thanks Bella 🧡 That’s exactly who we’re building for. We’re trying to go past vanity ‘AI visibility’ charts into stuff a growth team can act on: which topics/angles to publish next, which markets are under-served, and how model citations move after each new campaign or article.
Very interesting idea! Will your platform help increase agency recommendations when clients search for services in LLMs? Just Reddit or Medium isn’t enough for this — the formula is more complex, as far as I know.
Maritime
@mykyta_semenov_ Great question, Mykyta!
Yes, agencies are a big use case. We don’t just look at Reddit/Medium – we map all the places models seem to pull from (docs, blogs, forums, tools, etc.) and show which ones correlate with getting recommended in LLM answers. That gives agencies a more concrete playbook for where to show up and how to tune their clients’ presence
@maria_gorskikh1 @clyp1108 @yiko_li really interesting - and I guess the question becomes - what separates this from other GEO ranking platforms doing something comparable on the warm up front (and I guess how do these personnas avoid being flagged as bot like - or.. non human?) 2025 is so weird haha.
Really great idea and congrats on the launch!
Maritime
@clyp1108 @yiko_li @dzaitzow Appreciate it! We actually run an agentic simulations, not just a warmed account. The goal isn’t dodging flags—it’s capturing how memory drives model answers. Each persona/geo has a long-lived history; we control what it’s “seen” (docs, forums, tools), let it interact in realistic sessions, then rerun the same intents to see how recommendations shift. That lets us attribute changes to specific sources and explain why a brand gets cited.
@clyp1108 @yiko_li @maria_gorskikh1 interesting (and way outside of my scope) to really comprehend! haha
love what you're building here @maria_gorskikh1 . There is definitely a demand for this, brands don't have ways to find out how well they're performing in this AI driven search era. The “Google” has changed, and so should the way we measure visibility. Looking forward to seeing how this evolves!
Maritime
@new_user___3142025cf07667ae158a038 Thank you, Aditya 🙏
That’s exactly the gap we’re trying to close – there’s no ‘analytics’ layer for AI answers yet. Citable basically tracks your share-of-voice across models, personas, and markets, then ties it back to the specific pages and sources driving those mentions. Thanks for trying out our product!