How AI rankings work
This section explains the ideas the product is built on. Someone asks ChatGPT or Perplexity which supplier to use and gets one synthesised answer back rather than ten links — your brand is either named in it or it isn't. AI search explained covers why that shift matters and how it differs from a ranked results page.
From there, three concepts do most of the work. Prompts are the questions we send to the models on your behalf, and how realistic they are determines how useful everything downstream is. Visibility, position and sentiment are the three numbers every scan produces — how often you're mentioned, where you sit in the list when you are, and how favourably you're described. Citations are the external sources the models lean on, which is usually where the explanation for a competitor's strength turns up.
Read this section if you're new to AI visibility, if you're being asked to explain the numbers to someone else, or if a figure on your dashboard doesn't match your intuition. It's concept-first rather than click-by-click — for the dashboard pages themselves, see Using the dashboard. Everything here applies on every plan, though which models a scan covers varies by tier.
Start here
- What SimplyRank measures — visibility, position and sentiment, with ranges and how to read them.
- How prompts work — what makes a prompt useful, patterns that work, and how many you need.
- The AI models we track — ChatGPT, Claude, Perplexity and Gemini, and how their behaviour differs.
All articles
AI search explained
Why AI models are becoming a discovery channel like Google, and what that means for your brand.
Citations and sources
How AI models cite external sources, what those citations mean for your brand, and how SimplyRank tracks them.
How prompts work
What a prompt is in SimplyRank, what makes a good one, and how to write prompts that produce useful data.
The AI models we track
ChatGPT, Claude, Perplexity, and Gemini — what each one is, how they differ, and which plans include them.
What SimplyRank measures
The three core metrics — visibility, position, and sentiment — and how to read them.
Frequently asked questions
What is a visibility score?
It's the percentage of your prompts where the AI mentioned your brand in its answer, on a 0–100 scale. Thirty-six mentions across fifty prompts is 72%. There's no universal benchmark — a good number depends on how crowded your category is, so compare against your strongest competitor rather than against 100%. See What SimplyRank measures.
What counts as a good prompt?
One a real customer would plausibly type: specific, framed around a decision, and usually asking for a recommendation, a comparison or an alternative. Avoid one-word queries and avoid naming your own brand, since that tests recall rather than discovery. How prompts work has patterns and examples for each type.
Which AI models does SimplyRank scan?
ChatGPT, Claude, Perplexity and Gemini. They behave differently — Perplexity cites every claim and weights freshness, Claude leans on editorial sources, Gemini tracks closest to Google's index — so a brand can be strong on one and near-invisible on another. Every plan scans all four; see The AI models we track.
Why does my visibility differ so much between models?
Because the models were trained on different content and retrieve differently at query time. That variance is useful rather than a fault: it tells you which engine to focus on and which signal to work. Compare them side by side in the dashboard and read The AI models we track.
What is a citation, and why does it matter?
A citation is an external source — a review site, a publication, a forum thread — that a model references when it answers. Citations are the clearest available evidence of why a model recommends who it recommends, which makes them the most actionable data in the product. See Citations and sources.