Operato AI Help Center

1.1What Operato AI does

When people ask ChatGPT, Claude, Gemini or Perplexity which product or company to choose, those assistants answer with a short list of names. Operato AI shows you whether your company is on that list, where it ranks, who is recommended instead of you, and what to do about it.

Every week, Operato AI asks the AI assistants the questions your customers ask, reads the answers, and turns them into a few clear numbers and a list of actions. You see:

• How visible you are: how often you are named, cited and recommended.
• Against whom: how you compare with the competitors you track.
• Why: which websites the assistants rely on when they answer.
• What to do next: concrete actions, ranked by how much they can help.

1.2Your first 10 minutes

1. Add your company. Enter your company name and website. Operato AI reads your site and suggests a description of what you do and a list of competitors.
2. Check the suggestions. Edit the description, remove competitors that are not relevant and add the ones you know. You can change all of this later in Question set.
3. Start your first scan. Operato AI asks the assistants your questions. The first results usually arrive within a few minutes.
4. Read your Home page. Start with your Visibility score and Rank vs competitors, then open Action items.

1.3How often your data updates

Your questions are asked again automatically once a week. The Home page shows when the last scan ran and when the next one is due. After each weekly scan you receive the weekly report by email.

You can also start a scan yourself with Sample answers now. Manual scans come from a monthly allowance (see 6.3), while weekly scans are always included.

1.4Which AI assistants are tracked

ChatGPT, Claude, Gemini and Perplexity, on every plan.

2.1The five headline metrics

The top of your Home page shows five numbers. Together they answer one question: when buyers ask an AI assistant, do you come up?

2.2Other metrics you will see

• Average rank: where you usually appear among the brands an answer names. Lower is better: 1 means you are named first.
• Sentiment: how favourably you are described when you are named, from −1 (negative) to +1 (positive).
• Recommendation rate: the share of answers that actively steer the reader toward you, not just mention you. This is the closest to "the assistant would send a customer to you".
• Citation share: the share of all sources cited in the answers that are your own website.
• Lift: the change in your visibility between the first and the latest scan in the period you selected.
• Readiness score: how easily AI crawlers can read your website, from 0 to 100 (see 3.5).

2.3Question-level metrics

On the Question intelligence page, each tracked question has its own scores:

• Stage: where the question sits in the buying journey (for example discovering a category, comparing options, or deciding).
• Intent: how close to a purchase the question is. Questions with high intent matter most.
• Gap: how far the leading brand's visibility is above yours on that question.
• Impact: the upside of winning the question. It weighs the gap, how firmly a competitor holds the question, and how commercial it is. Action items and "Where to focus next" are ranked by it.

2.4Why your numbers move from week to week

AI assistants don't give the same answer twice. The same question asked twice can return a different list of names. This is normal, and it is why Operato AI measures over many questions and many answers instead of trusting a single screenshot.

What this means for you:

• Small weekly changes are often noise. A few points up or down in one week is not a trend.
• Look at the direction over 3 to 4 weeks. A steady rise or fall is real.
• Compare yourself with your competitors. If everyone dropped the same week, the assistants changed, not your position.

2.5What your numbers do not tell you

• They describe your question set, not the whole market. Your scores reflect the questions you track. If you add or remove questions, your scores can change even if nothing else did.
• They don't count traffic or sales. Being recommended more often should bring more visitors, but assistants often hide where a visitor came from, so this link is hard to see in your analytics.

3.1Home (Overview)

Your summary. It shows the five headline metrics (see 2.1), your visibility for each assistant, and how you compare with your competitors.

Open Advanced for more detail:

• Visibility by platform: your score on ChatGPT, Claude, Gemini and Perplexity separately. Assistants often disagree, so this shows where you are strong and where you are missing.
• Top mentions: the brands named most often in the answers.
• Key topics: your visibility for each group of questions.
• Competitor comparison and Competitive standings: every tracked brand side by side, with visibility, share of voice, presence, average rank, sentiment and recommendation rate.

Copy shareable summary copies a short text with your current numbers, ready to paste in an email or on LinkedIn.

3.2Action items

The list of what to do, ranked by how much it can help. Each action explains what was found, shows the evidence behind it, and links to the questions or files concerned. The actions you can see include:

• A technical problem on your website, for example an AI crawler blocked from reading your site. Click Get fixed file to download the corrected file.
• Most of what you lose, you lose to one competitor: one rival wins most of the questions you don't. Study what they publish and where they are cited.
• Questions with buying intent where a competitor is recommended instead of you: the most valuable questions to win first.
• The answers are built from other people's pages, not yours: the assistants rely mostly on other websites. The fastest lever is to be mentioned on those sites (see 3.4).
• You already lead some questions: where you are ahead, so you know what to protect.

Action items are calculated from your real results. The same results always produce the same actions.

3.3Question intelligence

Every question you track, with its score and who wins it.

• Questions tracked, Questions you lead, Never mentioned: a quick count of where you stand.
• Question performance: each question with its stage, intent, your visibility and presence, and the brand that leads the answer. Click a row to read the real answers the assistants gave.
• Where to focus next: the questions where you are behind, ranked by impact, with your score next to the leader's.
• Content briefs: for any question you are losing, Write brief drafts a content plan: what a page should cover so the assistants can name and recommend you. Generate top 5 briefs writes the five most valuable at once. Briefs are drafted by AI from your real results, so review them before you publish anything.

3.4Citation sources

The websites the assistants rely on when they answer your questions.

• Citation mix: how the sources split between your own site, your competitors' sites, independent publishers, and community sites such as Reddit or YouTube.
• Publisher targets: the independent sites cited most often. These are your best opportunities: an article, a listing, a review or a mention there can bring you into the answers.
• All sources table: every cited site with how often it is cited, on how many questions and by how many assistants, plus a suggested next step. Click a site to see the exact pages that were cited.

How to use it: pick 3 to 5 publisher targets, check whether they already mention you, and work on being listed, reviewed or quoted there.

3.5Technical readiness

Whether AI crawlers can read and cite your website. Your Readiness score (0 to 100) is based on these checks:

• AI crawlers allowed (for example GPTBot, ClaudeBot, PerplexityBot and Google-Extended)
• robots.txt present
• Homepage reachable
• Server-rendered content (the text is in the page itself, not only loaded by JavaScript)
• Title and meta description
• Canonical URL
• Structured data present
• Organization schema
• FAQ schema
• llms.txt published
• XML sitemap

Each failed check is listed as a fix, most important first. A blocked AI crawler is the most serious problem: an assistant that cannot read your site cannot cite it.

3.6My toolkit

Ready-to-use files that fix what the audit found:

• robots.txt: your existing file plus an explicit "allow" rule for every AI crawler that is currently blocked.
• llms.txt: a short summary of your website for AI assistants, built from your homepage's real title, description and sitemap.

Click Get file, then ask whoever manages your website to upload it at the root of your domain (for example yoursite.com/robots.txt). Files are generated fresh each time from your latest audit.

3.7Question set (Settings)

Everything Operato AI measures comes from this page. It has four parts:

• Your company: the company name the assistants should recognise. Use Rename if customers know you under a different name.
• How do you define what you do?: short descriptions of your category, for example "custom clothing manufacturer". Questions are built from these lines.
• Who do you compete with?: your tracked competitors. Without competitors there is no comparison and no rank.
• Questions we are tracking: the questions asked to the assistants every week. You can pause or resume any question, tick suggested questions to track them, write your own, or click Ask an assistant for more for new ideas.

3.8Account and billing

Your plan, your trial status, your invoices, and the weekly report switch: who receives it, when the next one is due, and how to turn it off.

3.9Team

Invite colleagues at no extra cost: there is no seat limit on any plan. Roles:

• Owner: full access, including billing. There is always at least one.
• Admin: full access, including inviting and removing teammates.
• Analyst: can view and edit data, cannot manage the team or billing.
• Viewer: read-only access to every report.

Invited people get an email with a link to set their own password.

4.1Add or remove a competitor

Go to Question set, then Who do you compete with?. Click Add a competitor and enter its name and website. To remove one, click Stop tracking. The change is measured from the next scan.

Tip: track the 3 to 6 companies your customers actually hesitate between. Too many competitors dilute the comparison.

4.2Change the questions you track

Go to Question set, then Questions we are tracking.

• Pause a question that is not relevant. Its history is kept.
• Tick suggested questions and click Track the ticked questions.
• Add questions your customers actually ask: type them in your own words, the way a buyer would ask an assistant.

Good questions sound like a real buyer: "What is the best project management tool for a small agency?" rather than just "project management".

4.3Track a second company

Open the company menu at the top and click + Add a company. It gets its own competitors, questions and weekly report, separate from your other companies. All companies count toward your plan's question limit (see 6.2).

4.4Read your weekly report

The report arrives by email after each weekly scan. It shows your rank and visibility, how you moved since last week, who gained and who lost, and the most important action to take. Click any link to open the details in the app.

You can turn it off from the link at the bottom of the email or in Account and billing.

4.5Fix a technical problem on your website

1. Open Technical readiness and look at the failed checks.
2. If an AI crawler is blocked or llms.txt is missing, open My toolkit and click Get file.
3. Send the file to whoever manages your website and ask them to publish it at the root of your domain.
4. The next audit confirms the fix.

4.6Get mentioned on the sites the assistants trust

1. Open Citation sources and look at Publisher targets.
2. Click a site to see which of its pages were cited.
3. For each site, find the right way in: a directory listing, a review, a guest article, a comparison page, or a correction if they describe you wrongly.
4. Watch your Presence rate and Citation rate over the following weeks.

4.7Turn a lost question into content

1. Open Question intelligence and look at Where to focus next.
2. Click Write brief on a question you are losing.
3. Use the brief to write or improve a page on your site that answers that question clearly.
4. Review the brief before publishing: it is drafted by AI from your results.

4.8Share your results

On the Home page, click Copy shareable summary, then paste it in an email, a slide or a LinkedIn post. Use Export CSV at the top of a page to download its data with the date it was taken.

4.9Ask the in-app assistant

The chat bubble in the app answers questions about your account and your numbers, for example "Why did my visibility drop?" or "Which competitor should I worry about?". It uses your real results and tells you plainly when something has not been measured yet.

5.1What makes an assistant recommend a company

AI assistants recommend what they can find, understand and trust. In practice, three things matter:

1. Being readable: AI crawlers can access your site and its content is clear.
2. Being described on other websites: reviews, directories, comparisons, press and community discussions. Assistants rely heavily on what others say about you.
3. Answering the exact question: pages that clearly answer the questions buyers ask, with facts, prices, locations and comparisons.

5.2A simple monthly routine

• Every week (5 minutes): read the weekly report and check one action item.
• Every month (1 hour): pick one publisher target to work on, one lost question to answer with content, and one technical fix if any remains.
• Every quarter: review your competitors and questions so they still match what your customers ask.

5.3How long before results show

Technical fixes can show within a few weeks, once crawlers have read your site again. Mentions on other websites and new content usually take one to three months to appear in answers. That is why the weekly trend matters more than any single week.

6.1The free trial

Your trial lasts 14 days, with full access. During the trial you can track up to 175 questions.

6.2Plans

Every plan includes the four AI assistants, the weekly scan and report, and unlimited teammates. Scale also includes API and MCP access.

6.3Manual scans

Weekly scans are always included. Scans you start yourself with Sample answers now come from a monthly allowance: roughly one extra check of each tracked question every 30 days.

6.4What happens when the trial ends

Nothing is deleted. You can still sign in and read all your results and exports. New scans and question changes pause until you choose a plan. Choose a plan in Account and billing to pick up where you left off.

6.5Change or cancel your plan

Go to Account and billing and click Manage plan. There is no commitment: you can cancel at any time and keep access until the end of the period you paid for.

Why does ChatGPT give me a different answer when I check myself?

Assistants vary their answers, and your own account, location and history can influence what you see. Operato AI asks many times in a neutral way and reports the pattern, which is more reliable than a single check.

Why am I named but never cited?

The assistant knows your name, but it doesn't use your website as a source. Improve your technical readiness, make your pages answer buyer questions directly, and get mentioned on the publisher targets.

Why is a competitor I never heard of ranked above me?

Assistants rely on what is written about each company online. A smaller competitor with many reviews, listings or comparison articles can be more visible than a bigger one. Add it to your competitors to follow it.

My visibility dropped this week. Should I worry?

Check whether your competitors dropped too and look at the trend over 3 to 4 weeks. One week alone is often noise (see 2.4).

Can Operato AI change what the assistants say for me?

No. Operato AI measures and tells you what to change. The changes themselves (your website, content, listings and mentions) are made by you or your team.

Do you track Google AI Overviews or AI Mode?

Not directly. Gemini, which is also Google, is the closest signal in your results.

Who can I contact?

Write to sarah@operato-ai.com.