> ## Content Index > Fetch the complete content index at: https://www.appliedcomms.ai/llms.txt > Use this file to discover other available public pages before exploring further. # The dashboard said we were winning: I paid a GEO/AEO tool to track my firm's AI visibility for a month - URL: https://www.appliedcomms.ai/ai-visibility-dashboard/ - Published: 2026-10-14T07:00:41.000Z - Updated: 2026-10-14T07:41:03.000Z - Author: Michael MacLennan ## **The Brief** - **The result:** for a month, a paid AI visibility tracker ranked my consultancy first in its category. On the questions a buyer would ask without our name in them, AI engines named us in about one answer in nine, mostly on ChatGPT (27%, against Perplexity's 2%). - **What we did:** 15 buyer questions about AI adoption help in Scotland, run daily in four AI engines by the GEO/AEO tracker Otterly from 10 September to 8 October (more than 1,500 answers), and six of them put to Claude and Gemini by hand, twice. - **The takeaway:** the AI visibility number that matters is how often you are named when the buyer does not say your name, engine by engine. The dashboard I paid for did not lead with it. ## **The Story** On the last morning of the month, the AI visibility tracker I was paying for put Faur alone in its "Leaders" box, with a 100% likelihood to buy. Every other firm it tracked, Edelman included, was "Niche". I was pleased for about a minute. Then I opened the rows. [Faur](https://faur.site/?ref=appliedcomms.ai) is my consultancy, and the subject was our work helping organisations in Scotland adopt AI. The tracker was [Otterly](https://otterly.ai/?ref=appliedcomms.ai), one of the tools sold for generative and answer engine optimisation (GEO and AEO). For a month it tracked what ChatGPT, Copilot, Perplexity and Google's AI Overviews said on that subject, first on a free trial and then on its paid Lite plan. Every tracker figure below comes from Otterly's dashboard or its exports. ### **What the dashboard said, and what the rows said** ![](https://storage.ghost.io/c/e9/fc/e9fc5e96-e3eb-42e1-81da-72a4c05345f0/content/images/2026/10/dashboard-said-we-were-winning-dashboard-and-rows-16x9-2.png) The headline for the last fortnight was 19% coverage, first of nine firms, and a 49% share of voice. Two of my 15 questions contain the word Faur: "Is Faur a good choice for AI adoption advice?" and "Compare Faur with other AI adoption consultancies in Scotland." They scored 70% and 65%, and the 19% includes them. Take them out and Faur was named in 12% of answers that fortnight. Over the month, Otterly's export shows 149 answers out of 1,365, about one in nine: in part one's terms, a mention rate of 11%. Otterly's own top recommendation on 9 October was to remove the branded prompts. Its headline figure still counts them. The share of voice is arithmetic on a list I typed. Asked during the trial to suggest Faur's competitors, the tool offered Tesco, Sainsbury's, Asda, Morrisons and Aldi, so I entered eight names myself: Brunswick, Edelman, Burson, Weber Shandwick, Waracle, Evolve AI, Grayling and Hard Numbers. (Disclosure: I worked at Brunswick until 2023, have freelanced for Grayling since 2025, and Darryl Sparey of Hard Numbers features in part one.) The 49% is Faur's share of the mentions among those nine. Otterly did that sum correctly, but the firms the engines kept citing for these questions, among them Insightful AI, Summone and Eastside People, were never in it. Big brands meet the same flaw from the other side: [The Remarkables](https://www.the-remarkables.com/ournews?ref=appliedcomms.ai) quote a chief marketing officer "perplexed" to be "recommended alongside brands we would not consider as our competitors". So is one in nine good for a firm my size? I can't tell you, because the only firms the dashboard compared Faur with were the ones I typed in. ### **The useful numbers were further down** Split by engine, the one in nine came apart. Over the month, on the 13 questions without my name, ChatGPT named Faur in 27% of answers, Copilot in 7%, Google's AI Overviews in 6% and Perplexity in 2%. Five of the 13 never produced a mention at all. If a figure like this is going into a board report, that is the version to send, and the [Monthly Stakeholder Update template](https://www.commswith.ai/library/monitoring/?ref=appliedcomms.ai) on Comms With AI is one way to write it up. ![](https://storage.ghost.io/c/e9/fc/e9fc5e96-e3eb-42e1-81da-72a4c05345f0/content/images/2026/10/dashboard-said-we-were-winning-engine-split-16x9-1.png) ### **The number moves on its own** Nothing on faur.site changed from 2 to 7 October, and Faur's daily rate on those 13 questions still ran between 9% and 17%. The bigger movements did not line up with anything I did either. ChatGPT started naming Faur when asked for the recognised experts on AI for communications in Scotland on 15 September, Google's AI Overviews when asked about Edinburgh consultancies on 17 September, and Copilot on the experts question on 24 September, every day since. Three engines changed their minds on three different days. The rows show when, and what each one read, but never why. Otterly's own research is candid about this. Its [stability study](https://otterly.ai/blog/ai-search-visibility-stability/?ref=appliedcomms.ai) found that with 10 prompts, 90% of the brand coverage results it measured fell anywhere between 39% and 71%. Lite allows 15\. The newest days also arrive incomplete: when I exported on 9 October, Perplexity's answers for the last two days had not come in. Any change you make will line up with some rise or fall, whether or not it caused it. ### **What I changed, and what happened** On 15 September I planned five fixes and made one and a half: a redirect for a dead page Copilot had cited on 13 September, and tighter descriptions of Faur on my own site. A rewrite of the AI Adoption page, matching descriptions on LinkedIn and Companies House, and one corroborating LinkedIn post never happened, because September got busy with extra client work, which I suspect is true of most teams that plan fixes like these. Two things I had not planned went up instead: an article on whether a small organisation needs an AI use policy (22 September) and an "In practice" section on the AI Adoption page (30 September). The article was not cited once. After a technical tidy of the site's redirects on 29 September, citations of the AI Adoption page went from two to five a week to 31 in the last 10 days, and Otterly's report put them up 483% on the fortnight before. The rate at which the engines named Faur was 12% in the week before the tidy and 12% in the week after. Part one argued that the engines look for corroboration, being written about by somebody other than yourself, which "used to be called PR". Every one of the 149 mentions cited a site I run, and the one planned fix that was PR is the one I never made. ### **Then I did it by hand** Claude and Gemini are not in Otterly's Lite plan, so I asked them six of the same questions myself, on 15 September and again on 9 October: signed in to Claude with my profile switched off, and signed out of Gemini. That is 24 answers, and none named Faur or me. On 9 October they named the firms Otterly kept seeing in its citations, including Summone, 360strategy, Insightful AI and Eastside People. Asked who advises boards in Scotland on AI governance, Claude described the gap a small firm falls into without being asked: "my searches didn't surface a clear list of independent Scotland-based advisers focused on boards specifically, so there may be smaller practices I've missed." ### **What US$29 bought** The Lite plan cost US$29 (£22.42) for the month. By 9 October Otterly had also generated 60 recommendations for Faur, and Lite let me see four of them. The fairest was a page on what an AI adoption assessment is and who offers one in the UK, a question on which Faur was never named. Another was to "Publish a new page Unlock Senior Success: Top Partners for Executive Communications Support on your site", a close cousin of the self-ranking listicles part one warned against. On 8 October Otterly announced a [partnership](https://otterly.ai/blog/otterlyai-getfluence-partnership/?ref=appliedcomms.ai) with Getfluence, which sells "sponsored articles with do-follow backlinks" and "AI-Mention™ placements inserted directly into existing articles AI engines already source", with 25% off a first campaign for Otterly customers. Sponsored content is legitimate when labelled, and [Google's spam policies](https://developers.google.com/search/docs/essentials/spam-policies?ref=appliedcomms.ai) treat paid links as link spam unless they are marked rel="sponsored" or rel="nofollow". Across my 15 questions, news, media and blogs supplied about 6% of citations, against the "over 20%" Otterly quotes from its US benchmark. The sources that mattered, gov.scot and the Scottish AI Playbook among them, are ones nobody can buy their way into. ## **The Practice: what you can put into practice today** 1. **Run the free version first.** [Part one's method](https://www.appliedcomms.ai/gullible-engines-ai-visibility/) takes six questions, two engines, 12 answers and about 20 minutes. It tells you whether you are named at all before you pay to learn how often. 2. **If you pay for an AI visibility tracker, split the number.** Tag the questions that contain your name and report them on their own, then report the rest engine by engine. 3. **Read the rows before you believe a rise.** Check which page the engine cited and who wrote it, and compare any change with a week in which you changed nothing. If you are offered a placement, ask whether it will be labelled as paid and its links marked as sponsored. What is your number once your own name is taken out of the question? *Every Applied piece follows the same shape (The Brief, The Story, The Practice) and is made the same way: my argument and examples, drafted with Claude, every claim checked against its source, and a final edit by me.* --- **Coming up:** [Advocate and Sceptic: Billy Partridge on AI and Human Judgement](https://luma.com/azmtoigy?ref=appliedcomms.ai), a Comms With AI Leader Interview, Wednesday 28 October, 12:15 to 13:00 GMT, online and free. We'll discuss what healthy scepticism looks like in a team that uses AI every day, how to tell when an automated process should hand back to a person, and how to keep human thinking sharp when AI makes the obvious answer cheap to produce. **More like this** - [Gullible engines: what AI says about you, and the free way to check your GEO/AEO visibility](https://www.appliedcomms.ai/gullible-engines-ai-visibility/): the companion piece, with Darryl Sparey of Hard Numbers on why the engines are so easy to feed, and the free method this experiment followed. - [The Lag Problem: Comms With AI in the Monitor Phase](https://www.appliedcomms.ai/the-lag-problem-comms-with-ai-in-the-monitor-phase/): the Agent Series instalment on monitoring, the phase this work belongs to. - [Five layers of AI tooling, and the comms job each one is for](https://www.appliedcomms.ai/ai-tool-layers-for-comms/): how to match a comms job to the right layer of AI tooling.