Track every AI engine.
Not just ChatGPT.

You are the person who currently pastes the same questions into ChatGPT, Perplexity and Gemini by hand, and you need the reading to survive you going on holiday.

• Every scan runs your whole prompt set across the project's enabled engines (any four of the seven models on paid, a fixed four on free), so the tab-hopping becomes one dashboard read.
• Tracked prompts persist between runs, so the question set you built stays the question set, results land under the same words every scan.
• A failed or stuck run restarts from the Activity page against the same job row, instead of you re-running prompts by hand.
• Free tracks four engines indefinitely, so the workflow proves itself before budget is asked for.

1. 01Create your project and pick your four engines in the project settings.
2. 02Run prompt discovery and activate the questions your buyers actually ask.
3. 03Run your first scan from the dashboard and read the per-engine breakdown.

Metrics this role tracks: Visibility score · Mention rate · Share of voice · Average position

You run retainer reviews across many brands and need one workspace to hold several clients without their data bleeding together.

• One project per client keeps scans, prompts and history separate, and the plan caps tell you the scope before you scope it.
• Scan a client competitor as a subject in its own right and read both scores against the same prompts in the monthly call.
• Share of voice broken down engine by engine gives the review something concrete when a client asks why ChatGPT favours a rival.
• Printable reports and token-gated share links turn the comparison into the deliverable without buying the client a seat.

1. 01Give each client its own project so prompts and history never blend.
2. 02Add each client’s top rivals on the Competitors page and scan them as subjects.
3. 03Send the token-gated link from the Reports page before every review call.

Metrics this role tracks: Share of voice · Visibility score · Mention rate · Sentiment score

You own how the brand is described, and a shift in tone inside an AI answer is the early signal you would otherwise see last.

• Sentiment is plotted per scan beside the positive, neutral and negative answer counts, so drift shows as a line rather than an anecdote.
• Accuracy-risk findings raise alerts on the channels you already watch, high-risk answers interrupt, the digest records them.
• Every engine's answer to the same question reads side by side in the printable report, which is where framing differences between engines become obvious.

1. 01Switch the sentiment-drop finding to immediate in Settings → Notifications.
2. 02Read every engine’s answer to your most sensitive prompt side by side in the print view.
3. 03Open the Alerts page when tone moves and work the warning-level findings top-down.

Metrics this role tracks: Sentiment score · Mention rate · Share of voice · Visibility score

You sell in more than one market and need a reading per market rather than one blended global number.

• Set a target region per project and every engine is instructed to answer for that market, catching regional differences in how they respond.
• Name a language per project and every scan asks for answers in it, so the same domain tracks separately per language market.
• Extra-market scans are read back beside the home market instead of being folded into it, so the comparison stays explicit.

1. 01Set the target region and answer language in the project settings.
2. 02Run prompt discovery in the market’s language and activate its questions.
3. 03Read the extra-market scan back beside your home-market reading after the run.

Metrics this role tracks: Mention rate · Sentiment score · Share of voice

You produce the recurring visibility report and need the whole engine set in one artefact rather than seven screenshots pasted by hand.

• The print view renders a scan as a branded document with every engine’s answer to the same question side by side.
• Token-gated report links put the same reading in front of stakeholders who will never log in.
• The scan on screen exports to CSV or JSON when the deck needs raw numbers instead of prose.
• Score, share-of-voice and position trends plot across your scan history, so the report shows direction rather than a snapshot.

1. 01Open the latest scan from Reports and print it with answers side by side.
2. 02Create a token-gated link for the stakeholders who will never log in.
3. 03Export the same scan to CSV or JSON for the numbers in your deck.

Metrics this role tracks: Visibility score · Mention rate · Share of voice · First-mention rate