AI sales coaching platforms price in one of five ways: per seat per year, per active user, per usage or scenario volume, a platform fee plus seats, or a tier set by how many reps, agents and managers you have. Which one is cheapest depends on how many people actually run sessions each month. Pilots and negotiated enterprise agreements sit on top of any of the five, they are not a sixth model.
Example. A head of sales enablement budgeting for 400 reps gets three quotes: one priced per licensed seat, one priced per active user per month, and one a flat platform fee plus a per-seat add-on. The three numbers look close in a 50-rep pilot. At 400 reps they diverge by a factor of two, because only one of the three charges for people who barely log in.
Every vendor conversation about an AI coaching platform eventually turns into a pricing conversation, and every pricing conversation starts to sound the same: a number, a caveat about "it depends on usage," and a promise to send a formal quote after a scoping call. That is not evasiveness. Coaching platforms genuinely do not price the same way twice, because the thing being priced (a rep's or agent's practice time) varies more than a normal software seat does. A rep who runs one scenario a month and a rep who runs twenty cost the vendor very different amounts to serve, and pricing models exist to decide who absorbs that difference: the buyer, or the vendor.
This piece describes the pricing models in general use across the category, not any single vendor's price list. No Retorio price and no competitor price appears anywhere below. The goal is narrower and more useful before a call: know which model a quote is built on, and know what to ask so the vendor has to say so.
Five shapes cover almost every quote a buyer will see in this category. They are not mutually exclusive. Many vendors mix two of them in the same contract (a platform fee plus a per-seat rate, for example). What matters for budgeting is not the label a vendor uses, but which of these five mechanics is actually doing the pricing work underneath it.
Every rep, agent or manager with an account is billed, whether they open the platform once a week or once a quarter. This is the model most buyers already know from other enterprise software, which is exactly why it is easy to compare across vendors and easy to budget a year ahead.
What it costs you as your population grows: cost scales linearly with headcount, not with how much coaching actually happens. A team that provisions accounts for everyone at onboarding and only has a third of them coaching regularly is paying full price for the other two thirds. The buyer absorbs the gap between provisioned and active.
Question to ask the vendor: "Do we pay for every provisioned seat, or only for the ones a manager actually assigns to a scenario?" A vendor who answers with a headcount number, not a usage number, is selling per seat.
You are billed for the reps, agents or managers who actually log in and run a session in a given period, not for the full roster. This looks cheap in a small rollout, because a pilot group is, by definition, a group that is actually using the platform.
What it costs you as your population grows: the price per person can hold steady while the total bill still climbs faster than expected, because adoption tends to rise once a rollout moves past the enthusiastic pilot cohort into the rest of the organization. A model that looked inexpensive against 50 active pilot users can look expensive against 2,000 active users a year later, even at the same unit rate.
Question to ask the vendor: "What counts as active, how is it measured, and can you show us the active-user trend line from a customer of comparable size at 12 months?" If the vendor cannot produce that trend line, you are pricing blind.
Price follows how much practice actually happens, billed by conversation, scenario or practice volume rather than by headcount. It suits teams whose practice needs differ sharply by role, where a field rep and a back-office agent would never consume the same amount.
What it costs you as your population grows: the invoice follows volume, so a team that cannot forecast how often reps will practise gets an unpredictable bill. A vendor may cap sessions per user per month and charge an overage rate above the cap, which turns a usage model into a seat model with a penalty attached.
Question to ask the vendor: "Is practice metered? What is the exact cap per user per month, and what does one session over the cap cost?"
A fixed annual fee covers implementation, administration, reporting and support, and a per-seat or per-active-user rate is added on top. The base fee is meant to cover the fixed cost of running the platform for your organization regardless of size, and the variable rate scales with who is actually coached.
What it costs you as your population grows: the base fee amortizes, so cost per person falls as headcount rises, which makes this model friendlier to a large rollout than a pure per-seat model and friendlier to a small one than a pure platform fee. The risk sits in what the base fee is actually buying. If it is billed as covering "unlimited administration" but a second region or a second language later triggers its own new base fee, the model was never as flat as it looked.
Question to ask the vendor: "What exactly does the platform fee cover, and does adding a second business unit, region or language trigger a new one?"
Price is set by a band your headcount falls into (for example, up to 100 reps, 101 to 500, 501 to 2,000) rather than by an exact per-seat multiplication. Each band carries a flat rate, so two organizations at opposite ends of the same tier pay the same amount.
What it costs you as your population grows: cost is a step function, not a smooth line. Adding one rep who pushes you from 500 to 501 people can move you into the next tier and raise the bill by far more than one rep's marginal cost would suggest. Buyers near the top of a tier should ask what happens at the boundary before they sign, not after headcount grows into it.
Question to ask the vendor: "Where are the tier boundaries, and what happens to price the month we cross one?"
Two quotes built on the same model can still differ, because four things move the number inside every one of the five.
The licence is also not the whole cost. Four categories sit outside it and belong in the same budget line.
Forrester makes the same point about revenue enablement platforms in general, that the cost does not stop at signature (Forrester on revenue enablement platforms). HBR's work on B2B value perception describes the internal budget argument a pricing model has to survive (HBR on B2B value perception).
Two commercial patterns are often mistaken for pricing models. Neither is. Both wrap around whichever of the five a vendor actually uses.
Negotiated enterprise agreement. A committed spend minimum over a multi-year term, in exchange for a blended rate that sits below the list price of the underlying model. The blended rate is the headline, the commitment is the obligation, and the two are worth reading in that order.
A discounted or fixed-fee pilot, usually three to six months and a limited group of reps, agents or managers, is priced separately from the multi-year enterprise contract it is meant to lead into. The pilot price is not the platform's real price. It is priced to get to a decision, and the vendor expects to make the economics back over the scaled contract.
What it costs you as your population grows: the pilot number tells you almost nothing about the price per person at full rollout, because the two are not the same commercial offer. A pilot quoted at a low flat fee for 50 users says nothing about what the same vendor will charge for 2,000, and a buyer who anchors a full-scale budget on the pilot number is anchoring on the wrong figure.
Question to ask the vendor: "What is the per-person price at the rollout size we actually plan to reach, in writing, before the pilot starts?" A vendor who will only commit to that number after the pilot is asking you to negotiate from a position where your switching costs are already sunk.
Across all five, the single most useful question a buyer can ask in the first vendor call is: "Walk me through what this costs at our current headcount, and again at twice that headcount, assuming half of it actually runs sessions each month." Ask it of Retorio, and ask it of every other vendor you are evaluating in the same call. That one question forces the vendor to reveal whether the model is seat-based, usage-based, tiered, or blended, because each model answers it with a different shape of number. A per-seat quote doubles cleanly. A per-active-user quote depends entirely on the adoption assumption you supplied. A tiered quote may not move at all, or may jump sharply, depending on where the boundary sits. Getting that answer in writing, before a contract is drafted, is what turns a pricing conversation from a sales pitch into a comparable budget line.
Two further questions are worth asking in the same call. What happens to unused seats, do they roll over, get reassigned mid-term, or expire. And what happens at renewal, is the introductory rate a promotion or the ongoing price, and is any committed-spend minimum or auto-uplift clause in writing.
Retorio prices per user by licence type, and the licence type determines which features are included.
None of this replaces asking what the coaching itself covers, how reps, agents and managers are scored, and how quickly a new hire is expected to reach a working level of skill. Pricing is the mechanism. The coaching is the thing you are actually buying.
The five models in general use are per seat per year, per active user, per usage or scenario volume, a platform fee plus seats, and a tier set by population. Most vendor quotes, including a quote from Retorio, are built on one of these or a blend of two, and the model matters more to your total cost than any single vendor's headline rate.
Both exist in the market, and many vendors offer either depending on deal size. Per seat charges for every provisioned account regardless of use. Per active user charges only for reps, agents or managers who actually run a session in a given period. Ask which one applies before comparing two quotes, because the same headline number means something different under each.
It depends on the model. Per-seat pricing scales linearly with headcount. Per-active-user pricing scales with adoption, which usually rises after a pilot, so the bill can grow faster than expected even at a flat unit rate. Tiered pricing moves in steps at population boundaries rather than smoothly. A platform fee plus seats tends to get cheaper per person as headcount rises, because the fixed fee is spread across more people.
A platform fee is a fixed annual charge meant to cover implementation, administration, reporting and support, billed regardless of headcount, usually paired with a smaller per-seat or per-active-user rate on top. Ask what the fee covers specifically and whether adding a new region, business unit or language triggers a second one.
A pilot is useful for testing fit and adoption, but treat the pilot price as a separate commercial offer from the full rollout price, not a preview of it. Ask for the per-person price at your actual planned rollout size in writing before the pilot starts, so the pilot's low cost does not become the anchor for a budget that later has to cover ten times as many people.
Ask what counts as a billable seat or active user and how it is measured, what a platform fee specifically covers, where tier boundaries sit and what happens when you cross one, and what the per-person price is at your actual planned rollout size, not just at pilot size. A vendor who can answer all four in writing before a contract is drafted is easier to budget against than one who answers only after you sign.
Neither is better in the abstract, they fail in opposite directions. Per seat is predictable and wastes money on people who never log in. Usage follows real practice and becomes unpredictable when practice volume varies month to month. Pick per seat when you need a fixed budget line and expect broad adoption, pick usage when adoption is uneven and you would rather pay for what is used.
Retorio is GDPR-compliant, EU AI Act-aligned, and ISO 27001-certified. Hosted on Google Cloud Platform with EU data residency. Your data stays in Europe.
Vodafone VOIS · AXA · Daimler Truck · Kraft Heinz · Merck · Madrigal
Built in Munich, Germany. 50+ enterprise clients across insurance, pharma, telecommunications, automotive, and consumer goods.