Article
Seat pricing looks like a billing detail. It is closer to an org chart: it decides who is allowed to help a customer, and the answer it gives is usually wrong.
A customer reports something broken. The person who can answer it in thirty seconds is the developer who wrote that feature last month. They do not have a helpdesk login, because logins cost money and they are not support.
So the ticket goes to an agent, who reads it, does not know, messages the developer in Slack, waits, gets an answer, paraphrases it back into the desk — and something is lost in the paraphrase, so there is a second round. The customer waits a day for an answer that existed immediately.
Nobody decided this. It is what the pricing model decided, and it was decided once, quietly, when someone reasonably chose not to buy seats for people who are not support.
Shared logins. Teams pass a seat around to avoid paying for another. Now your audit trail says "support@" did everything and you have lost the ability to answer who actually replied to a customer — which you will want on the day it matters.
Nobody browses. When only the support team has access, only the support team reads support. The product manager who would learn most from an afternoon in the queue never goes there, and the company's clearest signal about what is broken stays in a tool most of it cannot open.
Escalation becomes a ceremony. If a specialist cannot simply open the ticket, escalation means re-typing it somewhere else. Context is lost at every hop, and the customer is asked to repeat themselves.
Seat pricing is at least predictable. Per-resolution AI pricing — increasingly common — is not, and it has a sharper version of the same problem: your bill scales with your support volume.
The events that generate a spike of tickets are bad releases, outages and confusing billing changes. Those are the months the invoice climbs, and the only lever you have is to make the AI do less of the work precisely when there is most of it. The pricing model and the crisis pull in opposite directions.
It genuinely can be, and it is worth saying so rather than pretending our model wins everywhere.
If you are one person doing support, per-seat on a free tier is cheaper than any per-workspace price including ours. If you run a large contact centre with high turnover, per-seat is honest: you really are consuming capacity in proportion to headcount. And if you need per-seat licensing to control access for compliance reasons, that is a real requirement.
Per-workspace pricing suits the middle — small and medium teams where the people who should answer a question are not all in the support department. That is most businesses, and it is why we chose it.
A single price for the whole desk rather than a price per agent, so adding a colleague does not change the bill. It suits teams where the people best placed to answer a question are not all in the support department.
No. For a single person it is usually cheaper, especially on a free tier. For a large contact centre it is an honest reflection of consumption. It is worst for small and medium teams, where it quietly decides that specialists outside support do not get access.