FAQ
GMB only works on Google Search. VisibilityAI checks how the live AI engines answer for your business (the full list is on our homepage) and, separately and clearly labelled, asks models such as DeepSeek and Meta Llama what they know. AI engines completely ignore GMB listings.
Your VisibilityAI directory listing is available after you register. That is not the same as appearing in ChatGPT, Gemini, or another AI engine. Those engines crawl and cite independently - VisibilityAI does not inject listings into ChatGPT and does not guarantee a recommendation.
Yes - fully global. Any business, any city, any country, any language. The platform supports GST, VAT, EIN, ABN, CRN, and other local registration IDs for trust verification worldwide.
AI engines decide for themselves what their answers contain, so we do not promise that your business will appear, be mentioned or be ranked in any AI answer, or by when. What we do: we check the prompts you track on the engines your plan covers and show you every result, including when your business was not mentioned.
Traditional SEO optimizes for Google's link-based algorithm. GEO (Generative Engine Optimization) structures your content so AI models can extract, trust, and cite your business in conversational responses. Different signals, different platforms, different strategy.
Writing AI-friendly content on your own site is table stakes - anyone can do that, the same way anyone could hand-code meta tags for SEO. What you can't build yourself is the rest of the stack: (1) placement on an independent, high-authority directory and comparison network that AI models already treat as a trusted third-party source - not your own website, (2) continuous monitoring of up to 8 AI engines per prompt on Premium (3 on Starter) - provider APIs and measurement checks, not unlimited consumer apps - that tells you the moment a competitor overtakes you, and (3) protocol-level access for AI agents via MCP, agent.json, and our agentic gateway. Agencies don't build their own rank tracker for SEO - they subscribe to one. VisibilityAI is that layer for GEO.
Yes - a full REST + GraphQL API for structured business entity data (schema.org JSON-LD, co-citation knowledge graph, provenance-tracked claims), plus an MCP server so Claude and other AI agents can query real business data directly. See the Developers & API section for endpoints and pricing.
Yes. Agencies get a dedicated dashboard to manage multiple client sub-accounts, white-label reporting, and a commission structure on referred plans. See the For Agencies section for the partner program details.
50+ categories out of the box - professional services (CAs, lawyers, doctors), local services (restaurants, gyms, salons), B2B (IT companies, manufacturing, logistics), and hyperlocal businesses. If your category isn't listed, you can still register - it's not a restricted list.
No. The dashboard handles Schema markup, llms.txt generation, citation seeding, directory & comparison page placement, and cross-LLM monitoring automatically - you fill in your business details once and the system does the rest. The API/MCP surface is only needed if you're a developer building on top of VisibilityAI's data.
It asks a real LLM (Gemini) the same kind of question a customer would - "best [your category] in [your city]" - and reports whether your business is mentioned, at what position, and against a named competitor if you add one. It's a live call, not a simulated score.
Both, and we label the difference. Core facts are self-reported at registration; GST/MCA-backed claims go through an admin verification step and are marked accordingly in the API and on your public profile - so downstream consumers (including AI agents) can tell verified from unverified.
Prompt Intelligence ranks the prompts you track by how winnable they look, using only your own checks: how often your business is named, which other businesses the answers name, and which sources the engines cite. It needs at least 3 usable answers for a prompt before ranking it. No AI engine publishes how often a prompt is asked, so we never show an AI prompt volume; search interest appears only where your Search Console or a keyword-volume source supplied it, and is labelled as such. The ranking is modeled from your checks, not a forecast. It is part of the Plus and Premium plans.
We do not claim visibility in the DeepSeek app or Meta AI, and we do not count these models as live AI engines. We run two separate, labelled checks. "Model knowledge": The model answers from what it learned in training, with no search. This is not the DeepSeek or Meta AI app. "Model + our search": We run our own web search for your prompt and give the results to the model, which answers from them. This is not the DeepSeek or Meta AI app, and the model did not search. Each result shows the method, when it was last checked, and says "temporarily unavailable" when a check could not run. Model-based checks are shown on their own and are never added to your live engine results or counts.
Content templates (9 of them: comparison and alternatives, faq page, how-to guide, local landing page, buyer's guide: choosing a provider in your city, service page, case-study draft, llms.txt block, schema (json-ld) block) start a draft from one of your recommendations and the evidence behind it. Every draft is checked against your own facts: a claim that is not in your profile or that evidence cannot be approved until you remove it or add the fact. Nothing is published without your approval. A page goes live on your profile; the llms.txt and structured-data blocks are text you install on your own site, which we then read back to verify. After it is live we re-check the prompt and show what was observed. We do not promise any engine will cite or recommend the page.
On the Plus and Premium plans, yes. You set a weekly limit (Plus up to 7, Premium up to 14 pages). You still approve every page first; pages beyond the limit wait in an ordered queue with their due times, and you can publish one earlier or cancel it. Every change is recorded. The limit is a way to keep a pace you can review. It does not change how any AI engine treats your pages.
On the Plus and Premium plans, yes. For your approved English prompts we write a Hindi and a Hinglish version, you approve each one, and then the same engines on your plan check it. The result is shown beside the English check of the same prompt: the latest answer per engine, with unavailable engines left out and no percentage until enough paired checks exist. It reports what each engine answered; it does not tell you how many people search in a language and it does not predict any result. Starter shows a locked preview.
Yes - no long-term contract, cancel anytime from Dashboard → Privacy & Data. If you cancel mid-cycle you keep access until the end of the paid period.
