Every CPQ on the market prices the same way. Charge per seat. Add seats when the sales team grows. Bill more every year whether the tool gets used or not.
This model is broken. Here is why, and what should replace it.
What you are actually paying for
Most CPQs list at $75 to $200 per user per month. Multiply by your reps. Add approvers. Add deal desk. Add the admin who keeps the config running. You are spending six figures a year before anyone writes a quote.
Then watch how reps actually use it. They do not. They avoid CPQ for as long as possible. They build shadow quotes in spreadsheets, get the deal shaped up, and only drop into the tool at the last possible moment. When they finally are in it, they are stuck for hours clicking through screens to produce a document the customer will spend thirty seconds reading. Most CPQs work this way. Reps hate them, avoid them, and pay the price in hours when they cannot avoid them anymore.
That is what your per-seat bill is paying for.
The math assumed something that is no longer true. It assumed CPQ was work done by people, at a rate that grew with headcount. More reps meant more quotes. More quotes meant more CPQ usage. Per-seat tracked value because per-seat tracked work.
The premise is gone, and the bill keeps coming.
The tax has three parts
A tax on hiring. Every new rep costs you a seat before they write a single deal. You want to ramp a rep who signs ten deals in their first quarter. You are paying for their CPQ seat for all of it, including the two months they spend reading the playbook.
A tax on keeping the seat full. Reps leave. Seats sit assigned. Admins forget to reclaim them. Procurement runs a true-up and you find out you have been paying for forty seats on a team of twenty-eight. The vendor has no incentive to tell you.
A tax on being efficient. If your sales org gets better at its job, you are punished. Higher deal volume per rep, same per-seat cost. Your CPQ bill is stuck to your org chart, not to your revenue.
None of this gets better once an AI agent joins the team.
AI agents killed the seat
The per-seat model assumed a human sat in the seat.
That assumption is dead.
Klarna cut 700 FTE and the licenses that went with them. Customer support platforms are seeing 80 to 90 percent seat compression in teams that adopted AI agents early. Seat-based pricing dropped from 21 percent of SaaS companies to 15 percent in twelve months.
Gartner expects 40 percent of enterprise SaaS contracts to include outcome-based components by 2026. IDC thinks 70 percent of vendors will move off pure per-seat by 2028.
The market decided. CPQ is the last room to get the memo.
The unit is the contract
Outcome-based pricing charges for the result, not the chair.
Intercom charges $0.99 for every ticket their AI agent resolves. Zendesk charges $1.50 to $2.00 per automated resolution. Salesforce Agentforce charges per agent conversation, not per user. No one is counting humans.
For CPQ, the outcome is not ambiguous. A signed contract is the point of the sales motion. Everything a CPQ does is in service of getting there. Configure the product. Price the deal. Build the quote. Route the approvals. Generate the document. Collect the signature.
One unit connects all of that. The contract.
That is what a CPQ should charge on. Not seats. Not modules. Not “platform access.” A signed contract is the moment value is delivered to both sides of the deal, and it is the only number your CFO, your CRO, and your board actually care about.
This flips three things at once. Your CPQ bill tracks your revenue, not your headcount. The vendor only gets paid when you get paid, so a CPQ no one uses stops billing. And a new rep costs you nothing until they close.
Your finance team can model it. Your CFO will understand it on the first read. Your reps stop being a line item on a software invoice.
Why has no legacy CPQ done this
Because seats are easy to bill and hard to lose. Legacy CPQ vendors are locked into the seat model the same way their customers are. Their sales comp runs on seat expansion. Their forecasting runs on seat count. Their enterprise agreements are priced, discounted, and renewed by the seat.
Moving a legacy CPQ to outcome-based pricing is not a pricing change. It is a rebuild.
Now watch what happens when they try to bolt an agent onto the seat model. If the legacy CPQ stays on per-user pricing and adds an AI agent, they have to charge for the agent too. A seat for the human and a seat for the agent. Your bill goes up while the rep does less work. The vendor cannot resist the seat math, so they double-bill, and the customer ends up paying more for a tool they already hated.
The seat model and the agent cannot coexist. The moment a vendor adds an agent, per-seat pricing stops making sense. They either keep billing by the seat and your costs balloon, or they cap the agent at one “AI seat” per user and pretend that reflects the work being done. Neither one holds up at renewal.
There is a harder reason underneath all of this. Most legacy CPQs are not confident enough in the outcome to price against it. If the tool only gets the quote 15 percent of the way there and the rep finishes the job in a spreadsheet, the vendor cannot charge for the contract. They can only charge for access to the tool and hope no one notices the gap.
An outcome price requires a product that actually produces the outcome.
What we are doing at Veles
We built Veles so the agent builds the quote and the rep reviews. The unit of value is a signed contract, not a seat. So that is how we want to charge.
Every team starts on a free tier. Run up to 25 contracts before you pay us a dollar. After that, pricing tracks the work. Contracts executed, not seats assigned. Ramp a new rep at no license cost. Hit a slow quarter and the bill drops with it. Hire ten more reps and your CPQ bill only moves when deals move.
That is what pricing should do. Track the contract, not the chair.
If you are running CPQ today and watching your per-seat bill grow faster than your revenue, the problem is not your config. The problem is the pricing model. That is the part no rep will ever fix, no matter how many seats you buy.
Simon, Veles’ CEO