per-user pricing

in English: per-seat pricing

basicwidespread

A pricing model where you pay a monthly or annual fee for the software multiplied by the number of users you grant access to, regardless of how much they work with it.

In plain terms

Per-user pricing works like a gym membership: everyone needs their own pass, and it costs the same whether you go daily or once a month. The software company does not measure how much work you did with the program, but how many people have access.

That is why it has become one of the most widespread pricing models for cloud-rented business software: it is easy to understand and easy to plan with. If ten colleagues need email, a spreadsheet application or a customer database, you pay the per-user fee ten times, and you know in advance how large the bill will be.

With the emergence of AI agents, this logic has shifted. If an agent performs the work of several people, a company may need fewer "seats," while the provider faces significant computing costs for the vast amount of machine work. Therefore, more providers are shifting to usage- or outcome-based pricing, or mixing these with per-head fees.

How it works

The provider sets a per-person fee, for example a few thousand forints or a few dozen dollars per user per month, often with multiple package levels.
The company purchases as many seats (licenses) as needed for its employees and assigns them individually by name.
The bill is proportional to the number of seats purchased: if a new colleague arrives, you buy one more seat; if someone leaves, you can reassign their seat, but with an annual contract, cancellation is often only possible at renewal.
Usage volume is usually not included in the price, limited at most by a "fair use" cap.
For AI features, many providers build on this: on top of the per-user fee they add a credit allowance or a usage-based surcharge, because agents' work incurs real computing costs.

An everyday example

Your twelve-person office subscribes to an office software suite expanded with an AI assistant and pays for twelve seats because everyone needs separate access. When two interns join in summer and you grant them seats, the monthly fee rises to cover fourteen people, even though they barely use the software.

Who is affected

entrepreneur, executiveYou must decide how many seats you pay for and whether the per-person fee is still worthwhile even if agents already handle part of the work.

employeeYour access to a tool depends on whether your employer buys a seat for you, which is why only a few team members may receive an AI subscription. → Work & HR

developerCoding agent subscriptions typically consist of a per-user fee and a usage quota, so exhausting your quota can halt your daily work. → Code

bank customer, investorSoftware company revenue long depended on the number of paying users, so investors watch how much agents reduce the required number of seats. → Money

A common misconception

MythIf I pay per seat, I can use AI without limits.

In realityIn most AI subscriptions, behind the per-person fee there is a usage limit, credit system, or throttling. Those who work a lot with agents can easily hit the cap or pay a surcharge.

Limits and risks

You also pay for seats that no one uses, so it is worth regularly reviewing who really needs access.
The per-person price does not show how much value you get: you pay the same amount for a colleague who works with it eight hours a day as for one who logs in once a month.
Sharing access among multiple people is prohibited by most contracts and is also risky from a security perspective because it is unclear who did what.
In the age of AI agents, pricing changes quickly, so check what happens with overage usage before signing a contract.

Similar, but not the same

usage-based pricing – There, the bill is determined not by the number of users, but by actual consumption (for example, processed tokens or the number of queries).
outcome-based pricing – There, you pay only for completed, successful tasks, such as every resolved customer service inquiry, not for access.
flat-rate pricing – There, the entire company pays a single fixed amount regardless of how many people use the software.

Related concepts

Broader

software pricing, subscription model, software as a service

Narrower

named-user license, concurrent user license

On the edge

usage-based pricing, outcome-based pricing, token, AI-agent, digital workforce

A brief history

Per-user licensing originates from the world of enterprise software, where programs were licensed per machine or per user. It became mass-market with software rented from the cloud: Salesforce, founded in 1999, made monthly, per-user subscriptions the basis of its business model, and most of the industry adopted this.

Where things stand

By the end of 2026, this remains one of the most common pricing models for corporate software, but it is under strong pressure due to AI agents. More providers are shifting to hybrid models, adding credit limits, consumption-based, or outcome-based elements alongside per-user fees.

Status: 2026 November

Also known as

per-seat pricing · seat-based pricing · per-user pricing · per-user pricing · per-head pricing · seat-based pricing · user license · per-seat licenc

Sources

Wikipedia: Software as a service

This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.