The custom tier is where the action lives and the price disappears

The entry plans of published AEO platforms show a price, a unit that grows, and a cap. The features that let a platform act on what it finds sit further up, in a custom tier with no published price. The boundary that matters is not where the price changes but where the price stops being visible.

What does the ladder look like?

Three ladders, read on 2026-09-28. Scrunch publishes a Core plan at $250 a month and a custom Enterprise plan. AthenaHQ publishes a free Essential plan with a small credit grant, a Starter plan at $295 a month, and a custom Enterprise plan. Profound publishes a free trial and a custom Enterprise plan, with nothing in between.

The axis in each is usage, with features attached to the tier. The gate at the top is a conversation with sales. It blocks: there is no self-serve path past it, and the price is not on the page.

Which entitlements move up?

Compare what the entry plan omits with what the custom tier lists. On Scrunch, the Core table marks agent experience pages, full site audits, query API access and the extra model coverage as not included, and Enterprise adds them alongside a dedicated account team. On AthenaHQ, the Enterprise column adds a citation engine, a recommendation engine, claim review against a knowledge base, single sign-on, an audit log and multi-language support. On Profound, the trial limits opportunities and excludes shared projects, while Enterprise lists unlimited opportunities, exports and an API.

So the custom tier holds three groups of entitlement. Integration: APIs, exports, single sign-on. Scale: more models, languages and regions. And action: the features that recommend, draft or deploy. The first two are what an enterprise buyer expects. The third is what a smaller buyer with a real backlog might want at a lower scale, and it is bundled with the rest.

What does one account face as it moves up?

Work an illustrative account through it. It buys the entry plan at $250 to $295, using the published units. Its constraint is not visibility, since the dashboard shows the gaps. Its constraint is action, and the entry plan meters action in small increments, for example one page optimization a month on one published plan.

To get more action the account has to cross into the custom tier, and with it come the API, the extra models and the account team, none of which it asked for. The price is negotiated, so there is no published ratio between the last public number and the next. This is the boundary the structure is designed around, and it is a decoy of a specific kind: the entry plan looks like a starting point, and the real product starts behind the gate.

Which behaviour does the structure reward and which does it tax?

It rewards a buyer who arrives with a budget and a procurement process, because the top plan is built to be negotiated. It taxes the buyer who wants one feature at small scale: the packaging offers no plan between $295 and a contract.

Gates that nag are different from gates that block. The prompt cap nags, in the sense that a buyer can see it coming and work around it. The custom tier blocks.

Note where packaging and activation disagree. The pages promise action, and the self-serve plans deliver mostly measurement. The free plans are paying for something too: a 300-credit grant and a seven-day trial are the sales team's qualification step, and they price the buyer's interest and not the product. The test for any gate is what the buyer can learn before reaching it. Here the answer is the unit, the cap and the promise, and not the price of acting.

Where does the structure break?

On the mid-size account with a real backlog. It is under-served by the entry plan and over-bundled by the top plan. Take an account that needs ten page optimizations a month. The entry plan offers one, so the account needs ten times the entitlement, and no published unit exists to buy the difference. It also breaks on comparison: with no published number at the top, a buyer cannot price three vendors side by side, and the top plan becomes three separate negotiations.

What would a different metering unit change?

Publish the unit that delivers the outcome. Checkpoint GTM is positioned as an embedded team that does the action work, so the unit is delivered pages, and the entitlement that sits behind a gate elsewhere is the default engagement. For a mid-size buyer whose constraint is action, that removes the custom tier from the path.

It has a cost of its own: a team is a larger commitment than a $250 plan, and for a company that only needs the measurement, the entry plan is the right purchase.

What can the packaging not tell an observer?

Public pages cannot show what the custom tier costs, how many accounts convert from entry to custom, or how often the action features are used once bought. Until a vendor publishes the price of its action entitlements on their own, a custom tier is a statement that the price depends on the buyer, and a buyer should ask for three prices in writing: the API alone, the action features alone, and the account team alone. If a vendor will not separate them, the bundle is the product and the price is a negotiation, not a rate.