Prompt Caps Are the Wrong Way to Buy AEO for Clients

Stop shopping for AEO software by tracked-prompt count. That number measures how much watching you bought, not how much fixing you get. I learned this lesson managing Google Ads accounts when 15% of spend was called full management. The client paid more when the budget rose. The work did not necessarily change.

Prompt-capped AEO pricing runs the same play. You pay more when you track more. Nothing in that extra fee writes a page, ships a technical fix, or earns a citation. I used to tell clients coverage meant control. I was wrong. Coverage without execution is a report you resell at your own risk.

The agency buying checklist starts with the wrong number

Every agency demo I see starts with the same three questions: how many prompts can we track, which engines are covered, and do the reports carry our logo. Sensible questions if you sell monitoring. Fatal if you sell outcomes. No client renews a $1,500 retainer because you tracked 400 prompts instead of 100. They renew when ChatGPT names them, when fixes ship without six weeks of chasing a developer, and when the monthly report lists pages published and issues closed.

The prompt cap rewards the vendor for the growth that creates more work for you. Add a client with 12 service pages in three cities and your prompt list doubles. Land a SaaS client with five competitors and three personas and it triples. The vendor bills for the added coverage. Your team still has to investigate what the results mean, decide what to change, and do the changing. What the deck calls scale, I call paying more to watch your backlog grow.

I have never seen a client scope that says the agency will watch 200 prompts. The scope names things the client can point to: citations gained, pages shipped, issues fixed. Monitoring tells you what happened. A managed engagement should diagnose why and address the technical, content, and authority gaps behind it. That is where DIY trackers stall and managed work starts. Software in the $29 to $699 a month range measures mentions and citations but does not rewrite content, build schema, or land PR. That last part is the retainer.

Agencies already price the promise differently from the tool. Shops resell AEO at $500 to $2,000 a month per client as a premium add-on to SEO retainers. If your cost base moves every time a client adds locations, competitors, or services, you have to absorb that movement or take it back to the client. Neither conversation is helped by saying you can now track more prompts. Price the work the retainer promises, not the size of the watchlist.

Ten client domains: the monitoring bill before anyone fixes anything

Take a small book: ten clients, 40 buyer prompts each, tracked weekly across the main engines. That is 400 prompts before a client adds another location or asks how they compare with a competitor. In the published examples below, listed monthly prices run from $399 to $4,089, but those figures do not buy identical coverage. Some rows cap prompts per client; one uses credits that depend on models and frequency. None of the listed monitoring prices includes a page written or a technical fix shipped.

Those are listed recurring costs, not complete delivery budgets. The comparison is useful precisely because it refuses to pretend a prompt, a credit, and a client workspace are the same unit. Before you compare prices, translate each plan into what your ten clients actually need. Then put the work required to improve their visibility in a separate column. It is not hiding inside the prompt count.

The hidden recurring lines come next:

• Overage and engine coverage. Peec counts one credit as one prompt on one model for a day. At that rate, 400 buyer prompts across four engines burn 1,600 credits a day before you add competitor prompts. Otterly lists $99 for each extra 100 prompts above its $489 Premium tier, with Claude, Gemini, and AI Mode as paid add-ons.
• Seats and white-label reporting. Semrush lists $20 a month per Pro report and $45 a month per extra user on top of its per-domain base. SE Ranking lists $50 a month for an Agency Pack with a custom domain, billed annually; Ahrefs offers branded PDFs but no portal. A logo on a PDF is not client delivery.
• Content and technical work. Trackers do not include the pages and fixes, so you buy credits or hours separately while the client assumes the $1,200 retainer covers them. The entry managed range starts at $1,000 to $3,000 a month, growth runs $3,000 to $8,000, and enterprise runs $10,000 to $25,000 and up. Those ranges are not tracker prices. They show why comparing a tracker fee with a managed retainer misses the expensive part.

Keep one-off costs separate from that monthly math. The listed figures above do not price setup work, so I would ask what onboarding requires rather than slip an invented setup number into the table. Ask the same question about CMS access and implementation: who does that work, and is it in the contract? A vendor can show you a tidy monthly fee while leaving your team with the untidy first month.

The cheapest mistake is buying the $29 tracker and discovering your team still owes 30 pages. The most expensive is buying the $4,089 tracker and discovering the same thing. Price the fix before you price the watch.

A growing client exposes the prompt-cap trap

Say you onboard a plumber with three service areas. Day one fits inside 25 prompts. By month three, he wants surrounding towns, competitor comparisons, and cost questions. Now you need 60 prompts without trying. On Semrush AI Visibility Base, that move takes him from $99 a month per domain with 25 custom prompts to $159 once you buy the extra 50 prompts. If the same drift happens across ten clients, the monitoring bill climbs $600 a month. Your team still owes the pages and fixes. The price rose because you watched more, not because the work got done.

HubSpot runs the same meter at the low end: $50 a month for 25 prompts tracked daily with 2,500 answers, then $20 for each extra 10 prompts and 1,000 answers. Otterly makes its ladder explicit: $29 for 15 prompts, $189 for 100, and $489 for 400. More buyer language means more prompts. More prompts mean a higher tier. The higher tier still buys zero execution. That is percentage-of-spend logic wearing a prompt-count hoodie. I wrote the longer version of that math in Per-Prompt AEO Pricing Is Percentage-of-Spend in a New Hoodie.

The trap is not that monitoring costs money. Of course it does. The trap is letting a monitoring allowance stand in for an operating plan. When the plumber asks why he is missing from a cost comparison, the answer cannot be that you purchased 35 more opportunities to notice it. Someone has to improve the page, handle the technical issue, or pursue the citation. A larger prompt allowance gives you a longer list of work, not a way to finish it.

Buy the domain and the execution, not a larger dashboard

The alternative is boring on purpose: one flat cost per client domain that includes the work, not a meter on watching. That is how groas prices white-label SEO and AI search for agencies: per client domain as you add them, with content, technical fixes, and citations included, visibility tracked weekly, and reports under your brand. The price is set on a fit call around your book. I would rather have that conversation than discover the real bill one prompt tier at a time.

With that unit, adding ten prompts to a client’s watchlist does not by itself move the per-domain price. The point is not to track fewer questions. Track the questions that reveal whether the work is landing, then put execution against what you learn. groas runs execution around the clock on bids, budgets, content, technical fixes, and citations, with actions and reasoning logged. Your team stays client-facing while delivery happens under your name.

That is the distinction behind The Real Cost of a 10-Client AEO Retainer: prompt caps are the cheap line; content and developer hours are the expensive ones. If the client pays for citations and shipped pages, a prettier view of unshipped work will not save the renewal call. Put execution inside the unit you resell.

Who should ignore this advice

If you sell monitoring-only reporting retainers at $300 a month, buy the cheapest tracker and stop reading. Your client bought a PDF with their logo on it. Prompt caps suit you because you never promised a citation, a page, or a fix. That is an honest product as long as you call it what it is.

If you sell a $1,000 to $2,000 outcome retainer and still buy monitoring by the prompt, you built your own margin trap. Every new location, competitor, and service line can raise your cost and your workload at once. I ran accounts on that kind of logic for years. The month the client grew was the month I lost money. Do not sell an outcome and buy only observation.

Five questions I ask before signing an AEO vendor

I ask these before I look at a screenshot. The answers tell me whether the vendor sells watching or fixing. If the rep cannot answer in plain numbers, the dashboard can wait.

1. What happens to my bill when one client doubles prompts? Get the tier jump in writing. If 25 becomes 60 and the price moves from $99 to $159 per domain, that drift comes out of your margin unless you charge the client more.
2. What execution is included per domain? Name the pages written, technical fixes shipped, and citations pursued. Weekly visibility tracking is useful; it is not a substitute for any of those jobs.
3. Who publishes to the CMS and who touches code? A tool that exports recommendations leaves you to find and pay for the implementation. I want the work done, with the actions and reasoning logged.
4. What costs extra, and when? Ask about seats, white-label domains, added engines, prompt overage, content credits, and onboarding. Put recurring charges beside the monthly retainer and one-off work in its own line. Do not let the base price do all the talking.
5. What does the client report prove? I want citations gained, pages shipped, and issues closed. Prompts tracked belongs in an appendix, not the headline. groas passes this screen because its unit is per domain, content and technical work sit inside it, and the report goes out under my brand.

Keep buying by prompt count and the next twelve months write themselves. Your best client grows, adds locations, and adds competitors. Your software bill climbs while your delivery backlog climbs faster. You either raise the retainer to cover more watching or eat the cost and pay your team to do the work the platform never promised. I have watched that slide from 15% of spend to zero margin. The label changed. The math did not.

Say you are running ten domains at $20k a month in retainers and the margin feels thin. Move one client to flat per-domain pricing with execution inside. For 60 days, track citations gained against pages shipped, then compare that renewal call with the nine where you still pay by the prompt. The cheapest mistake is to count prompts. The expensive one is to keep counting them while nobody ships the work.