Agency fit comes down to three things: client workspace separation, white-label reporting, and whether prompt allowance pools across accounts or is capped per account. Vendors do publish all three, but in units that do not compare, so the comparison has to be built rather than read.
One vendor counts sites, another brands, another projects, another brand dashboards. Every figure below was read from the vendor’s own pricing page on 17 August 2026.
Last updated: 17 August 2026. Agency tiers change more often than self-serve ones, so this page is rechecked quarterly.
The mistake agencies make is comparing the headline agency price. The number that decides your margin is the cost per client brand at the count you actually manage, and that number moves in completely different ways across these vendors.
What are the three axes that actually decide fit?
- The question
- How many distinct client accounts, and can a client see only their own?
- Published as
- Sites, brands, projects, or brand dashboards. Four words for roughly one thing.
- The question
- Can the output carry your branding rather than the vendor’s?
- Published as
- Usually a tier feature, sometimes a custom domain, sometimes absent entirely.
- The question
- Does one busy client eat another client’s prompts?
- Published as
- A total prompt count, or a per-brand count. This is the axis that decides how you sell.
What do the agency tiers actually cost?
| Tool | Multi-client tier | Client separation | Prompt model | White-label |
|---|---|---|---|---|
| Mentions.so | $99 / $199 / $399 | 5, 10, then unlimited sites | Pooled: 50, 100, 300 prompts | Yes, at $399, with custom domain |
| Rankscale | $99 / $385 / $780 | 10, 50, 100 brand dashboards | Pooled credits | Yes, from the $385 tier |
| Trakkr | £395 Scale | 10 brands | Per brand: 50 prompts each | Not published on the pricing page |
| Rivarise | $1,299 Agency | 5 client brands | Per brand: 60 prompts a day each | Yes, with client portals |
| Peec AI | Not published | 25 to 50+ projects by tier | Pooled credits | Not published on the pricing page |
| Otterly.AI | Enterprise, from $1,000 | Unlimited team members, one workspace | Pooled prompts | Not published on the pricing page |
Read the fourth column before the second. Mentions.so at $399 gives you unlimited sites sharing 300 prompts, so at twenty clients each account gets fifteen prompts. Trakkr at £395 gives ten brands with fifty prompts each, which is five hundred prompts that cannot be borrowed between clients.
Those are not the same product at a similar price. One is a pool you allocate, the other is a guarantee you can sell.
What does each client actually cost you?
Divide the tier by the client count you genuinely manage, not the maximum the tier allows.
| Tool and tier | At 5 clients | At 10 clients | At 20 clients |
|---|---|---|---|
| Mentions.so Business, $199 | $39.80 | $19.90 | Capped at 10 sites |
| Mentions.so Agency, $399 | $79.80 | $39.90 | $19.95 |
| Rankscale Growth, $385 | $77.00 | $38.50 | $19.25 |
| Trakkr Scale, £395 | £79.00 | £39.50 | Capped at 10 brands |
| Rivarise Agency, $1,299 | $259.80 | Capped at 5 brands | Capped at 5 brands |
We are the most expensive row here by a wide margin, and the cap is real: five client brands and no more. What that buys is nine engines on every brand and sixty prompts a day for each of them, which is a different shape of product from a large pool spread thin.
If your model is many small accounts, the pooled vendors win on arithmetic and it is not close.
If your model is a handful of accounts that each expect depth, per-brand allowances stop one client’s activity degrading another’s data. Decide which agency you are before comparing prices, because the cheapest option for one model is the wrong product for the other.
Which tools suit which agency?
High client counts on one bill, where per-client depth matters less than coverage of the roster.
- Strength
- Unlimited sites at the top tier with unlimited seats at every tier, plus white-labelling and a custom domain.
- Constraint
- 300 prompts pooled across every client. Depth per account falls as the roster grows.
Agencies that would rather manage a credit budget than a prompt cap.
- Strength
- 50 brand dashboards, 5,500 credits, white-label and REST API access, with ten engines named on Rankscale’s own pricing page.
- Constraint
- Credits have to be converted before you can compare anything, at 0.25 credits per engine query.
Agencies that want a fixed promise per client rather than a pool to ration.
- Strength
- Ten brands at fifty prompts each, with all eight models included at every tier and no per-model add-ons.
- Constraint
- White-label is not stated on the pricing page, so confirm it before pitching client-facing reports.
Small rosters where each client expects full engine coverage and the answer text behind every number.
- Strength
- All nine engines on all five brands, sixty prompts a day per brand rather than pooled, unlimited seats, client portals.
- Constraint
- Five brands is a hard ceiling, and $259.80 per brand at five clients is the floor rather than a waypoint. Above five clients this is the wrong product.
Agencies willing to run a sales conversation to get a number.
- Strength
- Explicit agency tiers scaling from 25 to 50 projects and beyond, and larger tiers can switch to weekly tracking, which its documentation says costs a third of daily.
- Constraint
- No price appears on any public page. You cannot size or compare it without contacting them.
What does the per-client allowance mean for the report you send?
It sets the floor on what you can defend in a client meeting. A prompt allowance divided thinly across a roster produces small samples per account, and small samples produce intervals wide enough to swallow the movement you are being asked to explain.
Work it through. Mentions.so Agency at twenty clients gives each account fifteen prompts. Checked daily on three engines that is 45 observations a day, which is enough to say something monthly and not enough to say anything weekly. Trakkr Scale gives each of ten brands fifty prompts, so the same daily checking on three engines is 150 observations a day per client.
Neither is wrong. But the first agency should be selling a monthly report and the second can credibly sell a weekly one, and pitching the wrong cadence for your allowance is how a retainer gets questioned. The arithmetic behind those thresholds is in is my AI visibility change significant.
What happens when you switch tools mid-retainer?
You lose comparability, and the client notices. Two tools querying different engines at different frequencies will report different numbers for the same brand in the same week, and the difference is a measurement artefact rather than a change in performance.
The practical protection is to run both for one reporting period and show the client the overlap before you cut over. That converts an unexplained step change into a documented methodology change, which is a very different conversation.
It also argues for storing raw answers rather than only scores wherever the tool allows an export. Scores from a retired vendor are not recomputable, so a switch without an export erases your history permanently. That is the same reason a baseline is worth keeping in your own spreadsheet, covered in how to baseline AI visibility.
What should you pin down before signing?
Four questions, and three of them are rarely on the pricing page.
Does a client see only their own workspace, or is separation just a filter on one account? This decides whether you can give a client a login at all.
Is the prompt allowance pooled or per brand? Ask in writing. It is the difference between a service level you can promise and one you have to ration.
Does weekly tracking reduce the cost, and does it reduce the value? Peec’s documentation states weekly tracking costs a third of daily. That is a real saving and it also cuts your sample by two thirds, which widens every interval you report, as set out in how many times to run a prompt.
And the one nobody publishes: how many times is each prompt queried per engine per day. Two agencies reporting different numbers for the same client usually differ on this rather than on anything real, which is the arithmetic in how much do AI visibility tools cost.
The wider field including single-brand tiers is in best AI visibility tracking tools, and the sub-$200 options in AthenaHQ alternatives.