A $29 AEO seat looks cheap until your account manager spends Friday rebuilding its charts and publishing the drafts it called “content included.” Agencies made this mistake with percentage-of-spend PPC tools: we priced the software line and ignored the unbillable work behind it. On a 10-client book, that work can cost more than the tool.

Myth 1: More tracked prompts means more value

It sounds fair. You track 15 questions for the dentist, 40 for the SaaS client, and pay for what you use. I used to tell clients the same thing about keywords in 2016, back when I billed for building big keyword lists by hand. I was wrong then, and the logic is wrong now.

A client’s question set is not fixed. The week ChatGPT starts surfacing a competitor for “best CRM for contractors,” you need more prompts to find out how wide the gap is. If each block costs extra, you either eat the cost or stop testing. The pricing unit starts deciding how thoroughly you do the job.

Look at what the caps mean in practice. The Semrush AI Visibility Toolkit runs $99 a month per domain for 25 tracked prompts, with $60 for 50 more prompts, $99 for another domain and $99 for another seat. Otterly has a similar shape: $29 for 15 prompts on Lite, $189 for 100 on Standard, $489 for 400 on Premium, plus $99 a month for another 100 prompts. I broke down why per-prompt pricing repeats the old percentage-of-spend trap elsewhere. Price coverage, not the number of questions you are allowed to ask. Otherwise, the bill rises just when you need to investigate a change.

Myth 2: White-label means your logo on their dashboard

I understand why agencies accept the logo test. A branded PDF looks like something you can send to a client, especially in a demo. Then the account manager opens the login, finds the vendor’s name in the URL or footer, and starts exporting charts into a separate template. They still have to write the paragraph explaining what changed.

For 20 clients, one review found reporting still sits outside the product because there is no branded client-ready output. That last mile may look small for one account. Repeat it across a book and it becomes a standing Friday appointment. What the deck calls “white-label,” I call homework with your logo on it.

The term covers more than one product experience: a comparison of agency AEO tools puts white-label anywhere from a logo on a PDF to a full domain. Ask to see what the client actually sees, not just the sample report. Your branding should survive the login screen and the support process. Your strategist should be able to work behind your team, and the action log should be clear enough to forward without translating it into a second document. If the demo still shows the vendor’s domain in the browser bar, budget for that last mile.

Myth 3: Per-domain tiers scale fine

Per-domain pricing sounds built for agencies. At two clients, nearly every tier looks manageable. Run the same quote against 10 domains before you put it in a proposal.

Say you charge $1,500 a month for AEO and hold 10 client domains at 40 prompts each. One workload comparison puts that setup at $399 for Geoptie, $489 for Otterly Premium, $795 for Peec Agency Scale, $990 for Semrush with only 25 tracked prompts per client, and $4,089 for Profound on a $99 base plus $399 per client workspace. Those are not interchangeable delivery packages. The point is to expose the units: prompts, workspaces, domains and seats can all move the bill differently as the book grows.

I use $199 per domain as a delivery baseline because groas lists SEO and AI Search at that price per client domain, with content, technical fixes and citations included. That is $995 for 5 domains, $1,990 for 10 and $2,985 for 15. The number moves in a straight line you can quote. A lower monitoring bill does not automatically mean lower delivery cost if your team must still turn its findings into work.

Per-workspace and per-seat plans can jump when you add a client or a colleague; Ahrefs-style per-seat pricing ties part of the bill to headcount and client count. Model the 10-client bill and the work it leaves you, not the price of the first domain. That is the number your retainer has to carry.

Myth 4: You need a developer to go live

Most agencies hear edge proxy and picture a two-week ticket with the client’s IT person. Fair enough. The old version of this work could mean DNS changes, Cloudflare workers and a developer who disappears mid-project. I have lived that ticket. The client wants visibility; everyone ends up waiting for website access.

Packaging matters here. groas describes agency onboarding as one connection per client for ad accounts and website, in a couple of clicks on any platform or CMS. That is a different handoff from asking your account manager to edit a theme or open the client’s Cloudflare dashboard. One gives your team a service to run. The other gives them an implementation queue they may not be able to bill.

Test onboarding on the awkward client site, not the pristine demo account. If the trial needs a developer ticket, count that work before promising a launch date. A short setup claim is useful only if it survives the site you actually have.

Myth 5: “Content included” means content published

This one costs agencies more than prompt caps. A plan advertises 50 articles a month; what arrives is 50 Google Docs. Somebody still has to fact-check them, format them, load them into WordPress, set the slugs, add internal links and hit publish. On a white-label article engine priced at $199.99 to $499.99 for 50 to 200 articles, the drafts can be the cheap part. Revision and account handling are where the margin goes.

Monitoring tools leave a similar gap. Brand Radar shows where you were mentioned but offers no briefs, content scoring or handoff. Otterly reviews describe a platform that shows visibility rather than changing the site itself; it is not always clear what specific action to take. Those tools may tell you where to look. Your team still has to close the gap.

Say your manager spends three hours per client per month turning drafts into published pages. At 10 clients, that is 30 hours. At a $75 loaded hourly rate, it is $2,250 in labor on top of the software bill. You can reasonably choose to staff that work. You cannot reasonably leave it out of the price.

Ask where the workflow ends. A draft, an approved page and a published page are three different deliverables. If “publishing” is not in the feature list, price those hours as if you hired someone to do them, because you did.

Myth 6: The cheapest seat is the cheapest tool

I hear this one from smart owners who should know better, partly because I said it myself for years about PPC optimizers. A $29 seat looks like a rounding error next to a $1,500 retainer. Then you staff it. Someone has to learn the UI, QA the prompts, chase CMS access, paste drafts into WordPress, rebuild the PDF under your brand and answer the “so what changed” email on Monday. None of that appears on the vendor invoice.

Run it per client. At $29 for the seat and four manager hours a month at a $75 loaded rate, you have $300 in labor and $329 in delivery cost before content or citations. Across 10 clients, that is $3,000 in labor attached to a $290 software bill. The hours are the larger line item, even though the seat is the number everyone remembers from the demo.

That is why per-seat and per-project caps can hurt agencies twice: the bill can climb with headcount and client count while the work stays on your desk. Cheap software with expensive labor is expensive software.

Before you sign, make the rep show you four things:

  • A 10-domain quote with the prompt limits visible. Can you track 25 prompts or 125 on a client domain without changing the price? If not, model the extra blocks rather than hoping you will not need them.
  • The client-facing experience. Open the login and the report. Check whose brand appears and who answers support requests. A logo on page one does not settle the question.
  • The last step of execution. Find out who fact-checks, approves and publishes content, technical fixes and citations. A draft in a queue is not work live on a site.
  • A real onboarding path. Connect a client, run the audit and inspect how changes are logged. If you need a developer or a freelancer to get started, put those hours in the proposal.

A pricing page cannot answer those questions for your team. A trial can. Put one real client through it, count every hour your team touches the account over seven days, then use that experience to challenge the 10-client model. Do not mistake an unusually quiet trial week for a monthly average, but do not leave the hours you observed out of the quote either.

Myth 7: Clients pay for the visibility report

This is the hardest one for me to kill. I sold PPC reports for years where the deliverable was a clean PDF and a hopeful paragraph about impression share. It felt like work because it took all Friday to build. I understand why a visibility dashboard feels like the product now.

But a chart cannot change a page. If ChatGPT stops naming a client for “emergency plumber Austin,” the client needs someone to investigate, make a fix and explain what changed. A visibility score without that link is a weather report. It tells them it rained and charges them for an umbrella they never got.

Here is the margin problem. One breakdown of a $2,500 monthly retainer counts $1,250 in partner cost, $350 in account labor, $150 in tools, $250 in sales and $100 in infrastructure, leaving $400 net. That $350 is not spare capacity. If your manager must export tracker output, interpret every score change and organise the next action, monitoring consumes labor before the client sees a fix. I worked through the 10-client AEO retainer math at $14,050 to $23,700 a month; the report itself is not the expensive part. Paying for monitoring and then paying your team to turn it into action is.

I am not arguing that clients should never see a report. They should be able to see what happened and why. I am arguing that the report is evidence of delivery, not a substitute for it. That distinction matters when an account manager spends Friday polishing a chart while the fix waits for Monday.

The best AEO dashboard still will not save the margin if the tool stops at monitoring. Before you sign, ask one question and make the rep answer without a slide: Who publishes the fix? If the answer is your team, price those hours first. That answer is your true cost.