Power Automate sells two ways to license a flow, and the choice turns on who or what runs it. The per user plan licenses a named person to build and run their own flows. The process plan, sometimes called per flow, licenses a single automation to run unattended for an entire department regardless of how many people trigger or benefit from it. The two are not ranked by price. They are matched to a pattern. The expensive error is licensing a high volume, organization wide automation per user across everyone it touches, when one process license on the flow itself would cover the same workload for a fraction of the cost. The opposite error is buying process licenses for individual makers who only run personal productivity flows the per user plan already covers. The correct position separates the personal automation population from the business critical unattended flows and licenses each on the plan built for it.
The two plans license different units. One licenses the maker, the other licenses the flow. The distinction is exactly what makes the breakeven so sharp, because a single high traffic automation can serve hundreds of people on one license or hundreds of licenses, depending entirely on which plan it sits on.
The per user plan licenses a named individual to create and run an unlimited number of their own flows for their own work. It is the right plan for the broad population of makers building personal and team productivity automations: approvals they own, notifications they trigger, small workflows that serve themselves and a few colleagues. For this pattern, per user is both correct and economical.
The process plan licenses a single flow to run, typically unattended and serving an entire department or the whole organization, regardless of how many people trigger or rely on it. It is the right plan for business critical automations: an invoice processing flow, an onboarding workflow, an integration that runs at scale. One license covers the whole audience, which is decisive where that audience is large.
Automation in an estate falls into three patterns, and each one points to a plan. The cost difference between getting it right and getting it wrong widens with the size of the audience the automation serves, which is why the high traffic flows deserve the closest look.
An individual building flows for their own work and a handful of colleagues. Per user is correct and cheap. This is the largest population by headcount, and the right move is to ensure these makers are on per user, not consuming process licenses meant for departmental automations. Buying process plans for personal productivity is paying scale prices for individual use.
A flow that serves a whole team or department, often unattended, triggered by many people or by events. Process licensing on the flow is decisively cheaper than per user across every person it touches. These are the automations where licensing the audience instead of the flow quietly multiplies the cost, and where the largest savings sit when corrected.
Many users already hold limited Power Automate rights bundled with their Microsoft 365 license, sufficient for basic flows within the standard connectors. Before buying any standalone plan, we test whether the existing entitlement already covers the use case. A meaningful share of automation needs are met by rights the estate has already paid for inside M365.
The correct position is reached by inventorying the automation estate, classifying each flow by its pattern, and licensing the high traffic flows on process plans while keeping personal makers on per user and exhausting bundled entitlements first. The savings concentrate in the flows currently licensed the wrong way.
We pull the full flow inventory and the run and trigger telemetry behind it, then classify each automation as a personal maker flow, a departmental automation, or a use case already covered by bundled M365 rights. The high traffic flows serving large audiences on per user licensing surface immediately as the candidates for process licensing, and the personal flows sitting on process plans surface as the reverse correction. This classification is where the entire saving lives, because it identifies exactly which flows are on the wrong plan.
We move each business critical, broadly served automation onto a process license on the flow itself, replacing the stacked per user fees of everyone it touched with a single license. Personal makers stay on per user or fall back to bundled entitlement where it covers them. The result is an automation estate where the heavy, organization wide flows cost one license each rather than scaling with their audience, and where no one holds a standalone plan for work their M365 license already permits. We set a review cadence so flows that grow into departmental scale move to process licensing as they cross the line.
The per user versus process breakeven, the three automation patterns and their correct plan, and the bundled entitlement test that often removes the need for any standalone plan. Sent on request.
We inventory the automation estate, classify every flow by pattern, move the high traffic departmental automations onto process licenses, keep personal makers on per user, and exhaust bundled M365 entitlements first. The result is an automation bill that tracks the work, not the audience.