Selling AI search reports to clients: options for agencies
Short answer: An AI search report shows a client how often ChatGPT, Perplexity, Google's AI Overviews and similar tools name their business when buyers ask questions. Agencies sell it as a one-off audit, a monthly report or both, plus fix work on the client's pages. The careful version promises a process and a measurement, never a result, because nobody can guarantee an AI tool will name a client. Ploy, a software subscription where an AI agent builds and edits websites, can track AI visibility, but it is one option among several.
Clients ask whether they show up in ChatGPT, and your agency wants a paid, repeatable report to answer it without promising something nobody can deliver.
I'm Ranjit, and I run Blur Studio: 15+ years building software, 120 apps taken live, and clients who have gone on to raise $10M+. I build AI systems and processes that help businesses do a lot more with a lot less time. I also work differently from most of what else is out there, more on that further down this page. Say hello.
You can sell an AI search report as a measured service with clear limits. Offer a baseline audit, then a monthly report that states the number of prompts, engines, runs and dates, and then optional fix work on the client's pages. Price it on your hours and tool costs, not on promised citations. Put the limits in the contract in plain words: AI answers vary from run to run, no one can buy a citation, and the report is a measurement, not a guarantee.
This page is for agency owners and freelancers who want to add the service without overselling it. It covers the product, a price method, a report template, the limits you must state, tool costs per client, and where Ploy fits as one option of several.
This page may contain affiliate links from which Blur Studio earns commissions.
What an AI search report actually contains
AEO means answer engine optimization: making your pages easy for AI answer tools to find, read and cite. GEO, generative engine optimization, is a near-twin label. The terms are defined differently by different vendors, so use one in your proposal and define it. Our SEO vs AEO vs GEO guide explains the difference, and our guide to getting mentioned by ChatGPT and Perplexity covers the page fixes that follow an audit.
What can be measured, based on what the evidence supports:
- Mentions: was the client named in an answer?
- Citations: was a client page linked as a source?
- Accuracy: was the description of the client right?
- Visibility percentage: in what share of many runs was the client named?
- Competitors named: who appeared instead, and how often?
- Referral traffic: visits from AI tools in the client's analytics. SparkToro's write-up of third-party samples found this small, so report it without hype.
Never report "AI rank". SparkToro's study (28 January 2026, about 600 volunteers, 2,961 runs on ChatGPT, Claude and Google's AI) found under a 1 in 100 chance of getting the same list of brands twice in 100 runs, and an identical order under 1 in 1,000. Visibility as a percentage over many runs was steadier. The study is not peer reviewed, the authors are not professional researchers, and one co-author works at Gumshoe.ai, a company that sells AI tracking. It dates from late 2025. Treat it as a reason to show counts, not ranks. The author suggests about 60 to 100 runs per prompt, and asks vendors to show their math. You should show yours too.
Three ways to package it
| Offer | What the client gets | Your effort | Main risk |
|---|---|---|---|
| Baseline audit (one time) | Crawler access check, page and schema review, a baseline citation check, a fix list | A few hours to a day, depending on site size | Client expects results from a list |
| Monthly visibility report | Visibility percentage by engine, competitors named, accuracy notes, changes since last month | Setup once, then a few hours a month | Numbers that bounce and confuse the client |
| Fix work | New or rewritten pages, answer-first structure, corrected facts, schema where it matches visible text | The largest part | Promised outcomes you do not control |
Start clients on the audit. It has a clear end, produces a fix list you can quote against, and lets the client see the measurement before agreeing to a monthly fee.
A baseline audit usually covers: whether the AI search bots are allowed in robots.txt, whether key pages are indexed in Google, whether each key question has a page that answers it directly, whether facts like prices and areas served are stated and consistent, which sources the AI tools cite for the client's category, and a first measurement. Google says there are no additional requirements to appear in AI Overviews and AI Mode, and no special schema, so do not sell schema as the key to AI visibility.
How to price it, with the conflict in the market
There is no dataset of what agencies charge for this. What exists is blog posts from agencies, and they disagree. Here is what two of them say, so you can see the spread. Do not average them.
| Source | Audit | Monthly |
|---|---|---|
| Alejandro Rioja, 22 August 2026 | Flat audit $500 to $750 for a single site under about 50 pages; $750 to $1,500 for larger or multi-location; $1,500 or more for a portfolio | Retainer 15 percent to 50 percent or more of the audit fee, each month |
| Cansu Ekren, Digital Agency Network, published 21 January 2026, updated 31 March 2026 | Audit project $2,000 to $8,000 one time | Retainers from $2,000 to $4,000 (entry) up to $8,000 to $15,000 (enterprise); an add-on to an SEO retainer of $1,500 to $3,000 for small businesses |
The scopes differ, which explains part of the gap. Both posts say outcomes cannot be guaranteed. Digital Agency Network's article says inclusion is probabilistic, so agencies "can promise a process, not fixed results", and its price ranges carry no source. Rioja says there is no mechanism that makes a model cite a specific domain for a fee, and that measurements go stale within weeks. These are practitioner opinions, not a rate card, and I do not treat either as "the going rate".
A safer method is to price from your own costs:
- Estimate hours for setup, each monthly run, analysis and the client call.
- Add tool costs per client (see below).
- Multiply by your normal rate and add your usual margin.
- Compare to the two ranges above, then decide whether you are priced as a small, mid-size or premium shop.
A worked example with made-up numbers: a monthly report takes 4 hours at $120 an hour ($480), plus $60 of tool cost per client, so $540 cost to you. If you sell it at $900, the gross margin is $360 or 40 percent. If the tool cost doubles because the client wants more prompts, the margin shrinks to $300, so cap prompts in the scope. These figures are invented to show the arithmetic.
A report layout you can copy
A client report should let a reader see how the number was made. Use this order:
- Summary: one paragraph in plain words, with one number and one next step.
- Method box: date range, engines, number of prompts, runs per prompt, whether web search was on, and who ran it.
- Visibility table: percentage of runs where the client was named, by engine and prompt group.
- Competitor table: the same percentage for each named competitor.
- Citations: which client pages were cited, and which outside sources the tools used instead.
- Accuracy: wrong facts found, where they came from, and who will fix them.
- Changes since last month: what you edited and what moved. Say plainly when movement is inside normal variation.
- Next month: two or three actions with owners.
- Limits: the paragraph below.
A sample limits paragraph you can adapt, written for a client:
This report measures how often AI tools name your business when asked the questions listed, using the engines, runs and dates shown. AI answers vary from one run to the next, so we report percentages across many runs and do not report a rank. No one can buy or guarantee a mention, and we do not promise any change in mentions, traffic or leads. Results reflect the tools as they behaved on the dates shown and may change. Pages that our AI tools help draft are reviewed by a person before publishing.
Also refuse these in your own pitch: guaranteed citations, "we will get you into ChatGPT", any share-of-voice number without its runs, engines and dates, and any claim that llms.txt or schema alone will do it. On llms.txt, a proposed file listing key pages for AI tools, SE Ranking studied nearly 300,000 domains in October 2026 and found no correlation with citation frequency, and Ahrefs log data reported in June 2026 found 97 percent of valid files got no requests. If your audit includes it, call it harmless but unproven.
Tool costs per client
You can run the first few clients by hand in a spreadsheet. Past that, a tool saves hours. These list prices come from each vendor's page in November 2026, and plans change, so check the vendor page before you quote a client.
| Tool | What it is | Listed price and limits |
|---|---|---|
| Ahrefs Brand Radar | An add-on to Ahrefs, a maker of SEO tools | Custom prompts listed from $50 a month; tiers are measured in checks, which equal prompts times platforms. Check the vendor page |
| Semrush | An SEO and marketing suite | AI features inside its "SEO + AI Search" plans; Starter listed $199 a month ($165.17 a month billed annually) with 50 prompts daily and 1 domain. Check the vendor page |
| Peec AI | A tracker for brand visibility in AI answers | Starter listed at EUR 85 a month, billed monthly, for 50 prompts and 3 models, as of November 2026. Check the vendor page, as the currency can differ by region |
| Otterly.ai | An AI search monitoring tool | Lite listed at $29 a month, excluding tax, for 20 prompts, as of November 2026. Check the vendor page, as it can show euros or dollars by region |
| Scrunch | An AI visibility platform | Core listed $250 a month on monthly billing, 1 brand. Check the vendor page |
| Profound | An enterprise visibility platform | Enterprise custom price only |
Three questions to ask any vendor before reselling its numbers. How often does it run each prompt? What counts as a mention? Can you export raw answers so a client can audit them? Vendors have not published their methods, and none of the tools above has an independent accuracy test behind its numbers, so keep your own sample hand-checked.
Where Ploy fits, and where it does not
Ploy's AI search visibility docs describe a saved workflow, a Ploybook, called "Track Your AI Visibility". It sets up a dashboard that asks tracked questions daily and records where the brand and competitors appear, using AI credits. The docs list six engines (ChatGPT search, Gemini, Claude, Perplexity, Google AI Overviews and Google AI Mode). Ploy's feature page lists five, so check which set you get. Ploy says no tool can guarantee a mention, that you should judge trends over weeks, and that results can differ from what people see. Another Ploybook, the agent readiness audit, gives a score from 0 to 5 and fixes gaps it can. Ploy can also draft pages and FAQs, which stay private until published and need human review.
What Ploy does not give you, going by its docs and pages:
- No confirmed white label. Ploy's homepage uses the phrase "white-label ready" for agencies, but its docs and agency page do not describe it. Do not promise a client your own brand on the tool. Ask Ploy in writing before you build a service around it.
- No confirmed multi-client dashboard. Per the agencies docs, you run one organization with one workspace per client, and you can either pay for one subscription across client workspaces or have the client pay in a separate organization. Credits are shared across an organization.
- No published caps or credit prices. Ploy does not publish how many questions or engines a plan tracks, or what a tracking run costs in credits.
Plan facts from Ploy's pricing page in November 2026: Free is $0 (500 credits a day, up to 2,000 in total), Starter $50 a month (4,000 credits), Pro $300 a month (24,000 credits), Enterprise custom. Credits do not roll over. See Ploy for agencies and white label website tools for agencies for the wider picture. Our guide to what to automate in SEO and AEO retainers covers the work around the report.
Ploy suits an agency that also builds and hosts the client's site on Ploy, so tracking and fix work sit in one place. It suits less well an agency that wants a white-label, branded reporting portal or tracking across dozens of client sites from one screen. For that, a dedicated tracker or your own reporting is the better fit. For client reports in general, see client reporting software for agencies.
A pre-sale checklist
- Write your definition of AEO or GEO in two sentences.
- Decide the sample: prompts, engines, runs, cadence.
- Run the method on your own agency first.
- Write the limits paragraph and put it in the contract.
- Pick the tool, and check its current price and export options.
- Set a price from your hours and tool costs.
- State who reviews AI-drafted pages.
- Define what you will not promise.
- Agree what the client must do: supply facts, approve edits, fix their own listings.
Who should skip this service
- Agencies with no one who can read and fix web pages. A report with no fix work is a number and nothing else.
- Clients whose buyers do not use AI tools to choose. Check their analytics and sales calls first.
- Any pitch that depends on a guaranteed citation. If a client insists, decline.
How I'd approach this
I would sell the audit first, run it by hand for the first two or three clients, and only then buy a tool. By hand you learn what a good prompt list looks like and where tools disagree with reality.
First-week plan:
- Day 1: Write the offer: audit, monthly report, fix work, with one-line limits for each.
- Day 2: Build the spreadsheet and report template from this page. Run it on your own site.
- Day 3: Price from your hours. Compare with the two public ranges and decide where you sit.
- Day 4: Pick one friendly client and offer the audit at a fair price in exchange for feedback on the report.
- Day 5: Test one tool on a free plan or trial and compare its numbers to your hand log. Do not resell numbers you cannot explain.
Decision rule: if you also build client sites and want drafting and tracking in one place, try the Ploy Free plan on one client workspace and read the limits above. If you need a branded portal, pick another tool. Blur Studio builds custom content systems and looks after the pages AI tools read, if you want that handled for you: see what Blur Studio builds.
A closely related page covers the same ground from another angle: Are Competitors Named Instead of You in AI Answers? How to Check.
Related guide: Make Your Website Readable for AI Agents: Where to Start.
Questions people ask
Can I promise a client that AI tools will mention them?
No. Both agency pricing posts I cite say you can promise a process, not fixed results. SparkToro's study also found lists of brands rarely repeat. Promise measurement and fix work, and state limits in the contract.
What should an AI visibility report include?
A method box with dates, engines, prompts and runs, a visibility percentage by engine, competitors named, which pages were cited, accuracy notes, changes since last time and a limits paragraph. Never report an AI rank.
How much should I charge for an AI search audit?
Public posts disagree: Alejandro Rioja lists $500 to $750 for a single small site, while Digital Agency Network lists $2,000 to $8,000 for an audit project. Neither is a dataset. Price from your own hours and tool costs.
Does Ploy offer white label AI citation reports?
Ploy's homepage says "white-label ready" for agencies, but its docs and agency page do not describe a branded report. Do not promise it to clients until Ploy confirms it in writing. See Ploy for agencies.
Do I need a tool, or is a spreadsheet enough?
A spreadsheet is enough for the first few clients. Tools save time at scale. Ask each vendor how often it runs prompts, what counts as a mention and whether you can export raw answers.
Should the audit include schema and llms.txt?
Google says there is no special schema requirement for AI Overviews and AI Mode. Studies found no measured effect from llms.txt. Include them as optional, low-cost items, and make sure schema matches visible page text. See adding FAQ and schema markup.
How often should I run the report?
Monthly suits most clients. Results can go stale within weeks, so state the date range in every report and keep the prompt list the same from month to month.
Who reviews AI-written pages?
A person on your team should review every page an AI helped draft, because AI output can contain mistakes. State this in the scope so the client knows who is accountable.
More in this topic: Websites: AI answers, AI bots and AI search reports (4 guides) →
Sources
- AIs are highly inconsistent when recommending brands or products (SparkToro)
- GEO pricing guide (Digital Agency Network)
- GEO audit pricing: what to charge clients (Alejandro Rioja)
- AI features and your website (Google Search Central)
- Optimizing for generative AI features (Google Search Central)
- llms.txt study (SE Ranking)
- 97 percent of llms.txt files got no requests (Search Engine Journal)
- Ahrefs Brand Radar (Ahrefs)
- Semrush pricing (Semrush)
- Peec AI pricing (Peec AI)
- Otterly pricing (OtterlyAI)
- Ploy AI search visibility docs (Ploy)
- Ploy agencies docs (Ploy)
- Ploy pricing page (Ploy)