AI visibility

More people now ask an AI assistant which tool to use before they search. AI visibility asks ChatGPT, Claude, and Gemini the questions your buyers ask, then reads every answer: does it name your property, does it recommend it, does it describe it correctly, and who does it suggest instead. It lives on the SEO page, one property at a time.

Write the questions buyers ask

A check asks each engine the same short list of questions, up to 10 per property. Good questions are the ones someone types before they know you exist: the problem ("how do I see Stripe revenue for several products in one place?"), the category ("what is the best uptime monitor for a small SaaS?"), or the choice between tools ("what are alternatives to a big SEO suite for a small site?"). Leave your own name out; a question that names you only tests whether the engine can repeat it.

1. Add your questions

   Type them under Buyer questions, one per line. Or press Suggest questions and LopeBase drafts a few from your site's title and description; they are added to the list for you to edit.

2. Save them

   Press Save questions. Changing a question's wording starts its history over, so keep the list steady once checks are running.

Anyone who can edit SEO can change the questions. Suggesting them uses a few AI credits and shows the cost first.

Run a check

Press Run the first check (later, Check again). LopeBase sends every question to each engine, then reads each answer. A check takes a few minutes and keeps running if you leave the page; the line under the button says which question it is on, and a toast tells you when it is done. Only an admin can run one, because it spends AI credits.

LopeBase asks each engine's model through its API, the way a developer would, without live web search. The ChatGPT, Claude, and Gemini apps can search the web, so what a person sees there may differ. Read the results as what each model knows about you, and watch how that changes over time.

Read the results

The top of the tab says how many answers recommend or mention you, then four readings: the share of answers that mention you, the share that recommend you, how many get you wrong, and how many questions you are missing from.

Each engine's verdict

A card per engine shows the model it asked, how many answers recommend and mention you, and a strip with one mark per answer: green recommends you, purple mentions you, gray leaves you out. Hover a mark to see its question.

Whether an answer names you is read from its text: your property's name or its site address has to appear in it. The rest (whether it recommends you, where you sit in its list, whether it describes you correctly) is read by an AI model and shown next to the full answer, so you can check it.

Every answer

Every answer lays each question against each engine. Click a cell to read that engine's whole answer, with your name highlighted and the other brands in bold, plus how it describes you, the brands it names, and the sites it points to.

Over time

From the second check on, Over time charts the share of each engine's answers that mention you, check by check. The questions stay the same, so a rise means the engines know more about you.

Act on the questions where you are missing

Where you are missing lists each question at least one engine answered without you, which engines left you out, and who they named instead. The usual fix is a page on your site that answers that question well. Press Add as article topic to put the question in Content's topic ideas, where Content autopilot or your coding agent can write it.

See which competitors the answers name

Competitors in AI answers lists every brand the engines named in the latest check: how many answers named it, where it usually sits in the list, which engines named it, why the answers suggest it, and a small trend across checks. Your own row is first, so you can compare. It comes from the same answers, so it costs nothing extra.

Sites the answers point to lists the websites the answers name or link, by how many answers name them. Your own site shows in green. Being listed on the others is one way into the answers.

Check every week or month

Under Questions and schedule, an admin can set the property to check on its own. Pick Weekly or Monthly, untick any engine you do not want to ask, and press Save schedule. The line under the engines shows what one check costs and about what that comes to each month. Off stops the schedule; you can still check by hand.

Open the schedule

What a check costs

A check uses AI credits for each answer and for reading the answers, so it grows with the number of questions and engines. The button shows the low and high end before you press it, and what the check used once it is done. Each check's cost is listed under Past checks, and on the AI credits page.

• AI visibility always uses LopeBase AI credits, even if your workspace has its own AI key, because one key cannot reach all three engines.
• A check only starts when your balance covers its high end, so it does not stop halfway for lack of credits.
• Fewer questions or engines make a check cheaper.

Read it from an agent

An agent connected to LopeBase's MCP server can read the latest check with the get_ai_visibility tool: each engine's verdict, the brands named instead, the sites cited, and the questions where you are missing. It never runs a new check.

Where to go next

• Use the SEO toolkit for how people find you in Google and Bing.
• Write articles with Content to answer the questions where you are missing.
• Find discussions about your product to join the threads where people ask the same questions.