Your AI Visibility Report Is 14 Pages Long. What Changed?

Q3 AI VISIBILITY EXCELLENCE REVIEW — Prepared for [Your Business]

Reporting period: July 1–September 30 | Pages delivered: 14 | Pages edited: 0 | Account manager: Your Dedicated Insights Team (do not reply)

Dear Valued Partner,

Your brand sentiment is Vibrant. Your citations remain at zero. We have enclosed a donut chart so these findings can sit comfortably beside each other.

This quarter, our system captured 212 screenshots and raised your AI Presence Index by 3 points. Recommendations executed: see Appendix B. We appreciate your continued trust in the distinction between observing a problem and owning it.

Executive summary: momentum without movement

Your brand achieved continued momentum across monitored AI surfaces. We ran buyer prompts on schedule, recorded mentions, positions, tone, competitors and cited sources, then placed the results in a PDF. The PDF has grown since last quarter. Your site has not.

Share of Model Voice increased from 11% to 12% against a competitive set selected for comparability. The AI Presence Index rose 3 points to 74. Its methodology is proprietary, as is the threshold at which an increase would require anyone to change a page.

We monitored ChatGPT, Gemini and Google AI Overviews under your plan. Perplexity appears in comparative observations. If you are wondering which buyer prompt brought a customer to your site, that information does not appear on this page. Please proceed to the sentiment section, where the news is warmer.

Brand sentiment: Vibrant

This quarter: Vibrant. Last quarter: Stable-Positive. The quarter before that: Quietly Confident.

Our language model read 1,400 AI answers in your category and rated your brand Warm, Trustworthy and Directionally Premium. No customer was interviewed. No buyer prompt resulted in a click. We have nonetheless assigned an upward arrow to the finding.

In one analysis of 46,350 AI-visibility mentions, a score can look like a ranking and read like progress without changing who gets cited. We have addressed that risk by calling ours an index. Its upward arrow is available as a separate image for your next board deck.

Client action requested: Please do not ask whether Vibrant is a buyer intent category. It is a sentiment category, and the sentiment is positive.

Citations earned: see Appendix A

Citations earned across monitored answers: 0. Third-party mentions of your brand inside answers: 2, both Reddit threads from 2022. URLs on your domain cited: 0.

This distinction matters. Brands are 6.5x more likely to be cited via third-party sources than their own domain, and Reddit is the most-cited domain across ChatGPT, Gemini, Perplexity and AI Overviews. Your two old mentions therefore occupy an encouraging neighbourhood, if not an answer a buyer saw this quarter.

The report contains 19 screenshots proving you were absent. Appendix A is blank for layout reasons. We could have used the space to list pages worth improving, but that would have made Appendix A an action plan.

Client action requested: Circulate the screenshots to anyone who has asked whether AI search is important. They establish that we looked.

KPI tiles: three ways to make zero look busy

Please admire the tiles before asking what was fixed. Their trend lines point up and to the right, which is what tracker tools measure well. The tiles are the deliverable.

Tile 1: AI Presence Index 74, up 3 points. Inputs include weighted mention rate, sentiment warmth, screenshot resolution and momentum. Calculation: proprietary. What moved the score: recalibrated weighting. What shipped on your site to earn the increase: nothing.

A tracker runs buyer prompts, then records mentions, positions, tone, competitors and sources. Our contribution was placing those observations in a tile with rounded corners. Should the index fall next quarter, we will investigate the weighting with the same care we applied when it rose.

Tile 2: Share of Model Voice 12%, up from 11%. Competitive benchmark: Chadwick & Sons Industrial Supply, an Ohio business with 14 employees that is not your competitor. You beat them in Gemini. They beat you in Perplexity. Neither of you was cited by Google AI Overviews. We selected this comparison for comparability, not because either company sells to your buyers.

The 1-point increase is genuine within the selected set. Whether your buyers ask either engine about Chadwick & Sons remains outside the scope of this tile. Adjusting the competitive set would affect the number, so we recommend keeping the set stable until renewal.

Tile 3: 212 screenshots captured, 0 pages edited. Coverage: complete. Action log: see Appendix B. Storage of screenshots: permanent. The screenshots faithfully preserve the state of your visibility before and after no changes were made.

Only 1% of visits click a source link inside an AI summary, which makes visibility within an answer worth examining. Our 212 images document where that visibility did not occur. They do not show what happened after a buyer encountered an answer, but they are exceptionally legible at full resolution.

Client action requested: Download the tiles before asking which URL gained a citation. The download button is included in your plan.

Engine-by-engine breakdown: monitored continuously, touched never

ChatGPT: one warm mention. For the prompt “best [your category] for small business,” the answer named a competitor and omitted you. We described the tone of that omission as neutral-positive. Cited sources included Reddit, YouTube and Wikipedia. Your site was not cited.

We have preserved the answer in a screenshot. The screenshot will help your team recognise the gap if it chooses to address it. No team member from ours has been assigned to do so.

Gemini: directionally aware of your category. For “who should I hire for [your service]?”, the answer listed three national brands and one Reddit thread. Your position was below the fold that has no fold. We recommend reviewing this answer during your next content-planning meeting. We can provide the screenshot again if the first copy has been misplaced.

Google AI Overviews: monitoring continues. Overviews appear on 20%+ of Google searches. 68% of US searches end without a click, rising to 83% when an Overview is present. Across 40 buyer prompts, your Overview citations totalled zero.

Position-one organic CTR is suppressed 58% when an Overview appears. We include that figure here so your zero sits within a larger industry concern. It does not make your zero smaller. We have coloured this section green because the monitoring ran successfully.

Perplexity: comparative observation only. Chadwick & Sons appeared where you did not. This result is not included in your three-engine monitoring allowance, but we felt it added useful context to the share-of-voice tile. We have not made a corresponding change to your site.

Client action requested: Decide which omission concerns you most. We will continue to capture it while you decide.

Recommendations: your work, presented as our insight

Recommendation 1: Consider creating authoritative content. Our models detected that you are missing from 60% of answers. Coverage of visibility tracking quotes a critic describing tools that report a gap without optimisation, execution or attribution. We agree with that diagnosis and have reproduced it here as a recommendation.

Suggested action: write something authoritative. Topic: to be determined by your team. Publishing cadence: to be maintained by your team. Technical infrastructure, entity clarity, schema and freshness: also yours. Once you publish, our role is to monitor whether the result is mentioned warmly.

We have not identified a URL to edit in this recommendation. Selecting one might imply we had accepted responsibility for the work. Please consult your internal owner, who has not yet been assigned.

Recommendation 2: Explore entity optimisation opportunities. Evidence-rich fixes such as quotations and statistics lifted citation rates by +37% and +22% in the Princeton GEO study cited here. We have not added a quotation to your pages. We have not added a statistic. We have added this sentence about quotations and statistics to your report.

Third-party validation is a related opportunity. The gap between diagnosis and execution is where programmes stall when analytics tools do not produce structured content or deploy fixes. We recommend closing that gap. Our contract recommends that your team close it.

Next quarter, we will rerun the prompts to determine whether your efforts succeeded. If you have not yet started, we will still rerun them. The schedule is the part we manage.

Renewal: protect the momentum you can measure

Your quarterly plan includes monitoring for up to 50 buyer prompts across 3 engines. Additional engines, prompts, projects and edits to your website are out of scope. Entry monitoring in this category runs from $29 a month for 15 prompts to $95 a month for 50 prompts and 1 project; enterprise plans are custom after a 7-day trial of 50 prompts and 3 engines, according to an October 2026 pricing snapshot.

You are on the annual Excellence tier. Renewal before October 31 locks in a 12% discount and another quarter in which screenshots can be collected. Non-renewal risks loss of access to historical screenshots, sentiment warmth and Appendix B. It does not alter the number of pages we edited last quarter.

Before Q4, please:

1. Approve renewal.
2. Assign an internal owner to Recommendation 1.
3. Assign an internal owner to Recommendation 2.
4. Confirm that execution lives with your team.

Our system will continue running the same prompts on schedule and recording the same fields. Your Q4 report will compare the new tiles with these tiles and describe the difference as progress toward presence. We noticed everything. We changed nothing. That was the service.

Package complete. Total pages delivered: 14. Total pages edited on your domain: 0.

Appendix B: the action log we do not provide

Some clients ask what an AI visibility dashboard should show beyond screenshots. We support informed purchasing decisions. A useful dashboard can show mentions, citations by URL, position in an answer, share of voice, prompt coverage, source mix and what reached your site. It can also connect those observations to GA4 AI-assistant referrals or assisted revenue.

Those fields tell you what happened. Management requires a second record: what changed, where, when and why. We have therefore prepared a five-line specimen to help your team understand the work it could request from an execution provider. The lines below are examples of a different service. They do not describe work we performed, and none should be entered into your Q3 action log.

A dashboard without that record is a thermometer. It takes your temperature every week and never brings soup. For archival purposes, we have illustrated the arrangement.

1. What changed, on which URL, and when. Specimen entry: edited /pricing/ on Sept 12; added a direct answer block, three buyer questions, one sourced statistic and schema. Our corresponding entry: no URL, no date, no edit.
2. Which engines moved on which buyer prompts. Specimen entry: ChatGPT changed from omitting the business to citing its pricing page for “best [category] for small business” on Sept 19; Gemini and AI Overviews did not move. Our corresponding entry: a share-of-voice tile rose 1 point. The tile does not identify a changed page.
3. Which buyer prompts still exclude you, and why. Specimen entry: “Who should I hire for [service]?” still returns three national brands and Reddit; the services page has thin entity clarity and lacks third-party validation. Our corresponding entry: we recommend exploring opportunities.
4. What the vendor did versus what it asked you to do. Specimen entry: vendor shipped two page rewrites, one FAQ and citation outreach to two directories; client approved the work. Our corresponding entry: your team owns the writing, publishing, technical changes and follow-up. We own the screenshots.
5. What ships next month. Specimen entry: rewrite the services page, publish a comparison page on a topic where competitors get cited, and fix crawl blocks on two URLs. Our corresponding entry: rerun prompts and issue the next report.

Please note the difference between these specimens and your existing tiles. Each specimen can be checked against a URL, a date or a buyer prompt. A tile can be checked against last quarter’s tile. Both activities involve comparison; only one leaves your site different from how it was found.

Procurement note: request the log before the pitch

If you are evaluating software to automate AI search visibility reporting across ChatGPT, Gemini and Google AI Overviews without hiring an agency, ask to see last month’s action log before the renewal call. A vendor that sends tiles is offering monitoring. A vendor that sends URLs, dates, decisions and moved prompts is offering something else. The distinction is not visible in a sentiment gauge.

Monitoring is useful when you need to know which buyer prompts mention you, which sources an engine cites and where competitors appear. It becomes an expensive way to avoid a decision when the same missing answers arrive every month and no one changes the pages, fixes the crawl blocks or works on the third-party sources those answers draw from. The screenshot can confirm the gap. It cannot close it.

For that reason, we suggest comparing trackers separately from tools that help you get featured in Google AI Overviews rather than only monitored. groas operates in the execution category: specialised models work continuously while a named strategist sets direction, with actions logged and explained. Our Excellence tier does not include those edits. This is not a defect in the screenshot files.

Say you spend $20k a month on search and your buyer prompts still return Reddit threads from 2022. You could commission another 212 images of the gap. You could also ask for two pages to be fixed and a prompt to be checked again afterward. We can supply the images. Please direct the second request to someone authorised to change something.

Final renewal filter: ask for the previous month’s action log. If the reply contains only tiles, keep your money. If it contains URLs, dates, decisions and prompts that moved or still miss you, keep the vendor. Our Q3 package contains the tiles, Appendix A’s blank page and this specimen of the log. It does not contain the log itself.

I used to send the Google Ads version of this report, and the only question worth asking now is: what did anyone actually do last month?