SEO client reporting: what agencies should show about AI answers

SEO client reporting now needs a page on AI answers: how often each client comes up in ChatGPT, Gemini and Perplexity, and in AI Overviews and AI Mode on Google, plus which competitors and sources appear beside it. An agency gets those numbers by running the same prompts on a fixed schedule, then puts them on one page of the monthly report, after rankings and traffic. For the tools themselves, our comparison of AI visibility tools sets four of them side by side, with other vendors' prices read on October 22 and 29, 2026.

Hypervisible, which we publish, was built for that page of the report: one account holds up to 80 client brands, and every Monday it reads three or five engines on the same questions and shows each change in points, the unit the template below uses.

Why clients now ask about AI answers

Clients ask because Google now answers many of their customers' searches with AI-written text, and because Search Console started showing them AI numbers in 2026. The owner of a roofing company in Dallas who types “commercial roofing contractor Dallas” into Google and reads a paragraph naming three rivals is going to call the agency.

The scale is hard to wave away. At Google I/O on May 19, 2026, Google said AI Mode had passed one billion monthly users, one year after its debut. How AI Mode decides which brands make it into an answer is the subject of our Google AI Mode guide.

Search Console made the topic visible to anyone with a login. Google began rolling out new generative AI insights on June 3, 2026, starting with a subset of UK sites, and extended them to every website worldwide on August 31, 2026. They show impressions and which pages appear in AI responses, by country. A marketing director who spots that new chart in the client's own account will ask what it means, and whether it is good news.

The chart can't answer the question on its own. It covers the client's verified site only. It doesn't show the text of the answer, the competitors named alongside, or a mention that rests on a review site or a Reddit thread instead of the client's pages. To know what the AI says, someone has to read the answers.

Two more pressures land on the same desk:

Fewer clicks. Pew Research Center's March 2025 panel of 900 American adults clicked a regular result on 8% of Google visits that showed an AI summary, and on 15% of those that didn't. A client whose rankings hold while traffic slips may be looking at exactly that. We dig into it in zero-click search in the AI era.
ChatGPT ads. OpenAI opened self-serve buying of ChatGPT ads on May 5, 2026. A client weighing that budget will first want to know whether ChatGPT already recommends them for free. Our ChatGPT ads guide covers how the ads work.

Our view: bring the first numbers before the client raises it. Explaining a competitor's lead is easier when you found it first.

What to track for each client

Track a fixed set of prompts per client, run on the same engines at the same rhythm, and five measures: mention rate, share of voice, position in the answer, cited sources, and what the answers say about the brand.

Prompts with and without the brand name

Keep two lists. Unbranded prompts are what a buyer types before knowing the client exists, such as “best project management software for a 15-person architecture firm” or “estate planning attorney in Raleigh for a blended family.” They show whether the AI recommends the client. Branded prompts (“[brand] reviews,” “[brand] vs [competitor]”) show what the AI says once the name is on the table. Report them apart. A prompt that contains the brand name almost always produces an answer that mentions it, so a blended rate flatters the client.

Write the list with the client's sales team, who hear buyers' questions every week. Then freeze it. If half the prompts change mid-quarter, next month's chart compares nothing. A quarterly review is enough.

Competitors

Start with the competitors the client names, then add whoever the AI recommends in their place during the first readings. Those are often not the rivals the client had in mind: a directory, a marketplace or a national chain can take the spot.

The five measures

Mention rate. The share of answers that name the client. Named in 21 of 70 answers, the rate is 30%.
Share of voice. The client's mentions as a share of all brand mentions in the same answers. If the answers name tracked brands 250 times and the client 40 times, share of voice is 16%. The formula is explained in AI share of voice.
Position. Named first among five firms is a stronger result than named fifth.
Cited sources. The domains and pages the AI relies on. This column usually turns into the action plan, because it lists the review sites, directories and trade press where the client is missing.
What the answers say. Tone, and plain errors: an old address, last year's prices, a product the client no longer sells.

Keep Google's own numbers in a separate block: AI impressions from Search Console, and AI referral visits in Google Analytics 4. For definitions and what counts as a good score, see what AI visibility is.

How often to check

A weekly reading on the same prompts is enough for a monthly report, since four or five data points show a trend. AI answers change between sessions, which is why one screenshot settles nothing. Daily checks earn their cost during a launch, a recall or a press story, when the client needs to know within a day.

A one-page monthly report

One page, in the same order every month, so the client compares this month with the last without rereading the method. Full prompt lists and raw answers go in an appendix or stay in the tool. If you already send a monthly SEO report, add this page after rankings and traffic instead of creating a separate deliverable.

One-page AI visibility report template, to adapt for each client.
Section	What goes in it	Where the numbers come from
1. Scope	Month, engines, number of prompts, reading dates	Tracking tool
2. Headline number	Mention rate on unbranded prompts, change in points from last month	Tracking tool
3. Against competitors	Share of voice and rank among the 5 to 10 tracked competitors	Tracking tool
4. Prompts won and lost	Prompts where the client appeared or dropped out, with one answer quoted	Reading the answers
5. Sources and claims	Most cited sites in the client's market, those where it is absent, what the answers say, errors found	Tracking tool and reading the answers
6. Google and traffic	Impressions in generative AI features, visits referred by AI platforms	Search Console, Google Analytics 4
7. Next month	Two or three actions, each with an owner and a date	Agency and client

For a busy owner, four habits keep the page readable:

One headline number. The unbranded mention rate tells the client whether AI recommends them to strangers.
Changes in points. Going from 10% to 15% is 5 points. Calling it “up 50%” oversells a small gain.
One real answer. A quoted ChatGPT sentence recommending a competitor lands harder than any chart.
Plain words. Say “AI answers” and “questions”; “LLM outputs” and “prompt clusters” lose most owners.

Send the page the day before the monthly call. The call can then go to section 7.

White-label or branded: what clients expect

A client needs to read the report in your agency's format, with your logo and your commentary. A rebranded dashboard only matters if the client logs in to the tool. Agencies still look for it: Americans search “white label seo report” about 480 times a month (Google Ads estimate, October 2026).

What the tools in this guide offered agencies on October 29, 2026, according to their own pages:

Hypervisible, which we publish, has no white-label option. Agencies take the figures into their own report, which is what the one-page template above is for.
AthenaHQ has a page for agencies that lists several brands managed from one platform with team permissions, pitch workspaces for prospects, and discounted licenses to resell alongside the agency's own services. It doesn't mention white label.
Scrunch points agencies from its pricing page to a separate agency offer. The pricing page doesn't mention white label.
Similarweb includes one user in each AEO plan, and its pricing page says nothing about sharing with clients.

None of these pricing pages listed white label for AI visibility tracking. If your clients insist on a rebranded dashboard, ask the vendor in writing before you sign.

Pricing AI visibility tracking as a service

Agencies that sell this tracking tend to use one of three models: fold it into the SEO retainer, bill it as a separate line per brand, or sell a paid baseline audit first and the monthly tracking after. We found no independent public survey of what agencies charge. The ranges in circulation come from agencies and vendors that sell the service, so we don't quote one here.

Inside the retainer

The AI page becomes one more section of the existing report. It's the easiest sale to a client already under contract. The catch is that monthly work goes unbilled, which only makes sense if the retainer already pays for the content and PR that get a brand cited.

A separate line per brand

The client sees what the tracking costs and can stop it without touching the rest. Build the price from two inputs. The first is the tool, divided by brand: Hypervisible Business starts at €169 a month for 10 brands on three engines, or €16.90 per brand (€36.90 on five), and Pro costs €89 for 5 brands (€199 on five). The second is the time spent each month reading answers, writing the commentary and presenting the page. In our view, that time sets the price far more than the subscription does.

Baseline audit, then tracking

Sell a one-off audit first: the market's prompts, the competitors named, the sources, what the answers say about the brand. It gives the monthly report a starting point and the client a plan. Our guide to the AI visibility audit lists what it checks. To size the job before quoting it, run the prospect through Hypervisible's free AI visibility checker: three AIs (ChatGPT, Gemini and Google AI Overviews) get 10 questions, half about the industry with no brand named and half naming the brand, and you have the 30 answers read 5 to 10 minutes later. There's no card to enter.

Whatever the model, don't sell a ranking. Answers vary from one session to the next and, as on Google, no serious vendor or agency promises a fixed spot. Sell the measurement, and the work that gets the client cited.

Choosing a multi-brand tool

Judge a tool on its price per client: brands or projects included, prompts per client, seats, and what the client can see.

Other tools' prices read on official websites on October 29, 2026; Hypervisible prices as of November 11, 2026; monthly billing, in the currency each site showed.
Plan	Monthly price	Clients covered	Prompts and seats	Client-facing reporting
Hypervisible Pro	€89 on 3 engines; €199 on 5	5 brands (or 1 brand in 5 countries)	40 questions per brand, 10 competitors, 5 users; weekly	No white label
Hypervisible Business	From €169 on 3 engines; from €369 on 5	10 to 80 brands	15 competitors, 15 users; weekly	No white label
AthenaHQ Starter	$295	Not stated on the pricing page	3,600 credits (1 credit is 1 AI response), unlimited seats	CSV export; API as a paid add-on
Scrunch Growth	$500 ($417 billed annually)	Not stated	700 custom prompts, 2,500 industry prompts, 5 seats	Reporting dashboard; Data API on Enterprise only
Similarweb AEO Intelligence	$129 ($99 billed annually)	Not stated	150 prompts, 1 user	Not stated

Divide by client. Hypervisible Business comes to €16.90 per brand on three engines (€36.90 on five), and Pro to €17.80 (€39.80). The other three pricing pages don't give a number of brands, so the cost per client depends on how you split the volume: Scrunch Growth's 700 custom prompts shared by ten clients leave 70 each, and AthenaHQ Starter's 3,600 credits buy 360 answers per client a month.

Then check the rest:

Engines and frequency. Hypervisible reads ChatGPT, Gemini and AI Overviews once a week, on Mondays, plus Perplexity and AI Mode on its five-engine plans. Similarweb checks ChatGPT, AI Mode, Gemini and Perplexity daily. Every Scrunch plan reads seven engines, Claude among them, and AthenaHQ Starter lists 11. If a client's buyers use Claude or Copilot, Hypervisible won't see those answers.
Client access. Hypervisible counts 5 users on Pro and 15 on Business. AthenaHQ Starter has unlimited seats with role-based access control, and Scrunch charges $25 a month for each seat beyond those in the plan.
On-demand readings. On Hypervisible Pro and Business, an on-demand reading, paid as you go with prepaid credit, refreshes a client's numbers before a quarterly review without waiting for Monday.

Running client reports on Hypervisible

We publish Hypervisible, and we put it first in the table above for agencies whose monthly report looks like the template in this guide. Here is what that rests on.

Its readings map onto the page section by section. Each Monday, the three or five engines on your plan get the same questions again. Per client, the account shows how often it is cited, where it sits in the answer on average, how it ranks beside tracked competitors in the same answers, which sites the engines lean on, and the tone and content of what they write about the brand. Every figure carries its change in points since the previous reading, which fills section 2 without a spreadsheet.

Google AI Overviews are on every plan, and AI Mode on the five-engine plans. That matters for a client who has just found AI impressions in Search Console and wants to know what the answers actually said.

Clients outside the US are covered too. The tool reads ten countries, the UK, Canada, France and Germany among them, in the language of each, so a client selling in Chicago and in London gets one reading per market, each counted as one brand. NEXUS GEO, the agency behind the tool, does only GEO work and builds the readings on data from Google and the AI engines.

Weigh its limits too: weekly rather than daily, Claude and Copilot left out, no white label, and a product that launched recently. Before pitching a prospect, run the free AI visibility analysis on their brand: 30 answers from three AIs, read in 5 to 10 minutes, with no card. Plan details are on our pricing page.

Measuring is only half the job: what the report should lead to

A report tells the client where it stands. Moving the numbers takes work on what the AI engines read, and Hypervisible gives the agency the material for section 7 of the page:

The sites the engines cite on the client's prompts: comparison pages, review sites, videos, forums. Those are the places where the client needs to be named.
What the answers say about the brand, and in what tone. A wrong address or last year's prices gets fixed on the client's site and listings.
Who is cited in the client's place, prompt by prompt.
The change in points each Monday, so you know whether last month's action worked.
A team that can review the tracking and fit it to the client's market, from the prompts to the competitors.

Hypervisible stops there. It measures and shows where to act, but it doesn't write the client's content, and it doesn't get other sites to mention the brand. Deeper work (content, site structure, structured data, a presence on the sources the engines read, reviews) calls for a GEO agency working with the brand or alongside its SEO agency. NEXUS GEO does that work; it's the agency that publishes Hypervisible, which means us.

SEO works on where a client's pages rank among Google's links. GEO works on whether the client is named and cited inside the answer an AI engine writes. The two help each other, since AI Overviews and AI Mode draw on indexed pages. The assistants also read forums, videos, reviews and comparison pages, and they answer with brand names, so the methods and the metrics differ.

Your brand

The free analysis asks 10 questions to three AI engines (ChatGPT, Gemini and Google AI Overviews), reads the 30 answers in 5 to 10 minutes and keeps your dashboard in your account.

Frequently asked questions

How do you report AI visibility to a client?

Use one page a month, always in the same order: scope (engines, prompts, reading dates), the mention rate on prompts that don't name the client with its change in points, share of voice against tracked competitors, prompts won and lost, cited sources and what the answers say, then two or three dated actions. Quote one real AI answer; it convinces faster than a chart. Put raw data in an appendix.

What should an SEO client report include in 2026?

A 2026 SEO client report still covers rankings, organic traffic, conversions and the work done that month. It should now add Search Console's generative AI impressions, AI referral visits from Google Analytics 4, and one page on AI answers: how often ChatGPT, Gemini, Perplexity and Google's AI features name the client, against which competitors, and which sources they cite. End with next month's actions.

Can AI visibility reports be white-labeled?

Among the tools in this guide, none listed white label for AI visibility tracking on October 29, 2026. Hypervisible, which we publish, has no white-label option: agencies copy its figures into their own branded report. The AthenaHQ, Scrunch and Similarweb pricing pages don't mention white label. If a client needs a rebranded dashboard, get the vendor's answer in writing before you sign.

How much should an agency charge for AI visibility tracking?

There is no independent public benchmark. Agencies fold tracking into the SEO retainer, bill it as a separate line per brand, or start with a paid baseline audit. To set your own price, add the tool cost per client (€16.90 per brand per month on Hypervisible Business with three engines, or €36.90 with five, for example) to the hours spent each month reading answers, writing commentary and presenting the report, then apply your usual margin.

Which tool tracks AI visibility for multiple clients?

Hypervisible, which we publish, is the one we suggest: on three engines, Pro covers 5 brands for €89 a month and Business 10 to 80 brands from €169 (€199 and from €369 on five engines), read every Monday with changes shown in points. On October 29, 2026, Scrunch Growth offered 700 prompts and 5 seats for $500. Compare price per client, prompts per client, engines and frequency.

How often should you check a client's AI visibility?

Weekly is enough for a monthly report: four or five readings on the same prompts show a trend, while one answer on its own can change between sessions. Check daily during a product launch, a recall or a press story, when the client needs to see a shift within a day. Hypervisible reads every Monday; Similarweb checks daily.

Does an AI visibility tool improve a client's results on its own?

No. A tracker such as Hypervisible, which we publish, shows how often the AI engines name the client and where to act: the sites they cite, what they say about the brand, who is cited instead, and each Monday's change in points. It doesn't write content or earn mentions on other sites. That work falls to the agency, or to a GEO agency such as NEXUS GEO, which publishes Hypervisible. GEO and SEO support each other, but they use different methods.