Methodology
Every number on our case studies comes from the method on this page. We publish it so you can judge our results on how they were measured, not just on what they say.
FirstMotion measures AI search visibility by tracking a fixed set of buyer prompts, repeatedly, across the major AI engines, and recording whether and how a brand is mentioned, which sources are cited and which competitors appear. We report visibility, share of voice, citation rate and mention position by engine, then connect them to AI referral traffic and pipeline.
Two things sit behind every engagement. ContextualJourney™ is our proprietary AI search visibility platform. It runs behind every engagement and is operated by your specialist team, so you see the outputs, tracking and reporting without needing to learn another dashboard. Tracking runs daily across Google, ChatGPT, Claude, Copilot and more, in every country a client operates in and for every buyer persona it targets.
PromptPath™ is our framework for mapping the prompts B2B buyers use inside AI tools when they evaluate a category, then building a GEO strategy that makes sure a brand is cited at each stage of the buyer journey. Most visibility tools stop at a score. Ours maps the buying unit and journey stage behind each prompt, so the team knows what to fix next and not just what the number is today.
The prompt set is the foundation of every measurement, so we build it from the buyer and not from the keyword tool. We start with customer and ideal customer profile work, then map the questions each buyer role asks at four stages: the problem they are trying to solve, the category they are researching, the vendors they are comparing and the due-diligence questions that close the deal.
In a full strategy phase this produces 150+ prompts across the funnel. We prefer a narrow set of high-intent prompts to a broad set of curious ones. A brand that appears for a definition query attracts readers. A brand that appears for a comparison or shortlist query attracts buyers. Once the set is agreed we keep it stable, so changes over time reflect real movement and not a shifting yardstick.
Each metric answers a different question, so we report them separately, by engine and by buying stage.
| Metric | Definition |
|---|---|
| Visibility | The share of tracked AI answers that mention the brand, across the whole prompt set. |
| Share of voice | The brand's share of all brand mentions across the prompt set, compared with named competitors. |
| Citation rate | The share of chats or answers that cite the brand's own site as a source. |
| Retrieval rate | The share of tracked AI chats in which the brand's domain is retrieved as a source. |
| Mention position | Where the brand appears among the brands named in an answer. Lower is better. |
| AI-referred sessions and pipeline | Visits from AI platforms to your site, connected to leads and pipeline in your analytics. |
For a deeper explanation, see our guides on measuring share of voice and the KPIs that matter for a GEO campaign.
These are the success criteria we set with clients for priority prompts. They are targets we aim for, not guarantees, and the right level depends on the category and the competition.
| Sentiment | Whether the tone of brand mentions is positive, neutral or negative. |
|---|---|
| Context | How the brand is described, for example as a trusted adviser, a budget choice or an innovative leader. |
| Accuracy | Whether AI represents the brand's products, pricing and features correctly. |
| Competitive position | How the brand's narrative compares with competitors on similar prompts. |
We combine specialist AI visibility tools with search data, and we name them so the method can be checked.
| Tool | What we use it for |
|---|---|
| Peec AI | AI visibility tracking: mentions, citations, retrieval and share of voice by prompt and engine. |
| Profound | AI visibility tracking: how brands are mentioned and cited across AI answers. |
| Ahrefs | Search, backlink and brand-mention data, including how third-party sites reference a brand. |
| Google Search Console | First-party search performance data for the client's own site. |
| AirOps | Workflow automation that lets the team ship more content and fixes for less effort. |
Tracking runs daily across the major engines, in every country a client operates in and for every buyer persona it targets.
Disclosure: Alex Price, a FirstMotion director, is an investor in Peec AI.
AI answers vary between runs, between users and between engines. One screenshot of a favourable answer proves very little, so we run the same prompt set repeatedly and report the pattern. ChatGPT, Copilot, Gemini, Google AI Overviews and the other engines also rarely move together, which is why we report each engine separately and never blend them into a single number.
There are limits to any measurement of this kind. A tracked prompt set is a sample of what buyers ask, not all of it. Tools differ in how they sample answers. And no agency controls what an engine says, so we do not guarantee specific rankings or citations. Where we publish a result, we state the source, the prompt count and the date range, so you can see how much weight it can bear.
Every case study we publish uses this method and links back to it. See the results for a legal technology platform, a UK wealth manager and FirstMotion itself, or read about the services built on it: our GEO agency programme.
In a full strategy phase we map 150+ prompts across the funnel. The set we track on an ongoing basis is agreed with each client and kept stable, so change over time is meaningful.
AI answers vary between runs, users and engines. A single screenshot tells you very little, so we repeat a fixed prompt set and report the pattern across runs.
ChatGPT, Copilot, Claude, Gemini, Google AI Overviews, Google AI Mode and Perplexity, reported separately because they rarely move together.
No. Engines change how they work often and nobody controls their answers. We measure honestly, remove the reasons a brand is being skipped, and report what changes.
Start with a free 20-minute call. We will show you where your brand and your closest competitors appear in AI answers, and tell you honestly whether we can help.