The Overview, top to bottom
Your configuration summary
The top of the page shows what’s being monitored: your markets, models, prompt count, and how many answers came back in the period — e.g. “234 prompts, 1,178 answers over the last week.” It’s your sanity check that the engine is running at the scale you expect.
Headline KPIs — total and per prompt tier
Total visibility, plus the split by Must-Win, Core, and Foundational prompts. The tiers mirror the purchase funnel — Must-Win is where users are closest to a decision, Foundational is broad research — so a strong total driven by Foundational prompts while Must-Win lags is a warning, not a win.
The competitive field
Visibility in isolation is abstract; the brands comparison makes it concrete. See who’s ahead, who’s behind, and who you’re actually fighting — often a two-horse race at the top with a gap to the rest.
The model breakdown
The same story per AI model. A brand can be strong in ChatGPT and nearly invisible in Google’s AI Mode — the aggregate hides this, the breakdown shows it, and a weak model is a natural target for a focused task.
Development over time
Switch the period from a week to a month or more to see how visibility has actually developed — trend direction beats any single day’s number.
Citation share
When a model answers, it runs web searches behind the scenes — typically visiting 5–25 pages — and 3RD captures which sources it used. Your citation share is the fraction of all those sources that are your pages. A share of 1.3% means roughly 1 in 100 sources is yours — a clear signal to invest in content that answers your prompts, and in crawlability so models can access it.
Brand sentiment
Every response is analyzed for positive and negative phrasing. You’ll see the key phrases AI uses about you — “innovative footwear, sustainable materials” — and can click through to the exact responses. Negative phrases can usually be traced back to a citation source, which turns a vague reputation problem into a concrete outreach target.
Recommended tasks
High-impact tasks 3RD believes will move your visibility — e.g. a content-coverage task for a prompt where you have no matching page, complete with a content brief and the fan-out queries (the keywords models actually search with) to weave into the piece. Accept a task and it lands on your Task Management board.
Topics breakdown
Visibility per topic with period-over-period change — e.g. sportswear at 60%, kidswear at 47% and down 7. Click into any topic to find where the real problem is (sometimes one market drags the whole number down). Focus on Must-Win prompts with low visibility, not the topics you’ve already won.
The Leaderboard: six views of the competitive field
The Leaderboard is more than a single brand ranking — it gives you six distinct views of the field, each answering a different competitive question:| View | Answers |
|---|---|
| Models | Who wins in ChatGPT vs. Gemini vs. Perplexity? |
| Topics | Who owns each subject area you track? |
| Audiences | Who does AI recommend to an enterprise buyer vs. a solo founder? |
| Markets | Who leads in each geography + language? |
| Sentiment | Who does AI speak most favorably about? |
| Citation Share | Whose content is AI actually learning from? |
From reading to acting
- Overview — spot a gap or a trend change
- Conversations — read the actual answers to understand what’s happening
- Diagnostics — get the root cause
- Playbooks — turn it into a prioritized plan
- Back to Overview — measure whether it worked
Related
- Rank — how the Leaderboard ordering is computed
- Citation Share — the metric behind the citation panel