Your customers’ questions.
Each project tracks its own list, in its customers’ language and market. Questions that already name you are counted apart from discovery questions.
AI VISIBILITY / PRO
Your brand.
In the answer.
Understand how AI talks about your business. Track the questions, compare the answers and find the details worth acting on.
Available on enabled Pro projects. Provider access and limits apply.Your audience asks.
Does your brand come up?
Does your brand come up?
ChatGPT
PerplexityYour brandThe answer, in context.
QUESTIONS → ANSWERS → EVIDENCEExample Studioexample.com
ChatGPTPerplexityREAD THE SIGNAL
Named. Recommended. Cited.
Three different ways to appear in an answer. See each one separately, with the measurement behind it.
18 discovery questions · 72 answers38%
Mentioned
Your brand is named
27 / 72 answers29%
Recommended
Presented as an option
21 / 72 answers14%
Website cited
Your site is a source
10 / 72 answersThe same questions and providers. Each successful answer counts once.
FIND THE NEXT QUESTION
See where you’re missing.
Start with questions that return answers without your brand. Open them in the workspace to compare providers and read the context.
18 discovery questions · 72 answers7 / 18questions with no brand mentions
7 without a mention11 with a mention
Discovery questions · grouped by outcome. Branded and zero-weight questions are excluded.
FOLLOW THE CONTEXT
Every signal has a source.
Go from a tracked question to the original answer, then inspect its mentions and source links. Keep the evidence in the conversation.
18 discovery questions · 72 answersYour tracked questionChatGPT and Perplexity
ChatGPT and PerplexityOriginal answerInspection path
- Read the mentionSee how the brand is described.01
- Compare the alternativesUnderstand which competitors appear.02
- Inspect the linked sourcesOpen the cited page alongside the answer.03
Example figures. In your workspace, a source that is not verified stays marked.
HOW WE MEASURE
A sample you can check.
Every figure says what was asked, which service answered and when. Read the method before the percentage.
The AI services you choose.
ChatGPT, Perplexity, Claude, Gemini, DeepSeek or Grok, asked through their APIs with web search, not through the consumer apps. A signed-in person can see a different answer, because history and location differ.
On a fixed schedule.
Every week, twice a month or once a month, on Monday morning (UTC). A question can be asked more than once in a run, because the same question can get a different answer.
Four weeks, not one answer.
Headline rates cover the successful answers of the last four weeks. Each answer counts once, so one unusual reply does not decide the result.
Every reading stays open.
Each answer is read for a mention, a recommendation and a citation of your website. Open the original answer next to the reading and mark it when it is wrong.
What a result does not claim.
A failed request is shown as missing, not as “not mentioned”. Results describe the questions you track, not every conversation, and no ranking is guaranteed.
INSIDE THE WORKSPACE
See the result.
Keep the context.
The overview brings questions, providers and coverage together. Open a question in Opmerk to inspect its answers and understand the result.
AI VisibilityExample data
Projects / Example Studio / AI VisibilityAI Visibility
ChatGPTPerplexityQuestions18Answers72Mentioned38%27 / 72Recommended29%21 / 72Website cited14%10 / 72
Discovery questionsChatGPT · Perplexity
- Interior designer for a small apartmentRecommendedMentioned
- Who designs boutique hotel interiors?MentionedNot named
- Kitchen renovation studio near meNot namedNot named
- Interior architect for a home officeRecommendedRecommended
- Studio for a calm bedroom redesignNot namedMentioned
- Who can redesign a family home layout?MentionedRecommended
- Restaurant interior designerNot namedNot named
Named in the answers
- Competitor A46%
- Example Studio38%
- Competitor B31%
- Competitor C19%
- Competitor D12%
The answer is in the evidence.
IllustrationTracked questions across ChatGPT and Perplexity. Mentions, recommendations and citations stay separate.
Example data
Ask what matters.
Track discovery questions around your service, market and audience. Keep questions that already name your brand separate.
Look at who appears.
Review competing brands mentioned in the measured answers. Read the response around the name before drawing a conclusion.
Follow the evidence.
Inspect linked pages and citation coverage. An unverified source stays unverified, so your team can see what needs checking.
FROM INSIGHT TO THE PAGE
A clearer answer.
A better next step.
Bring the finding into your website review. Inspect the relevant page, discuss the detail and give the next action an owner.
A STRAIGHTFORWARD START
From a question to a clearer decision.
Configure the project.
An administrator enables AI Visibility and configures provider access, project settings and limits.
Track meaningful prompts.
Add the questions you want to understand. Keep branded and unbranded questions distinct.
Read the evidence.
Compare the answers and citations. Check which prompts and providers completed before drawing conclusions.
Is this every answer an AI service gives?
No. The results describe the configured prompts, providers and measurement runs. They are a sample, not a view of every user’s conversation.
What do I need before a run?
An enabled Pro project, configured provider access and sufficient project quota. Availability depends on configuration.
Does a cited source make an answer correct?
No. A citation means a source was linked in that answer. Read the page and the original answer to judge whether the statement is supported.
SEE IT. UNDERSTAND IT. MAKE IT BETTER.
Your next version
starts here.
Bring every detail into focus.
Open your workspaceGet the Chrome extension