Make vs Power Automate: Credits vs Seats and Bots
Make bills $9/month billed annually for 10,000 credits with unlimited users; Power Automate bills $15 per user per month and $150 per unattended bot. Fix a workload, convert both to the same unit, and the answer flips depending on whether you already pay Microsoft.
Here is the answer before the working: if your company does not already pay Microsoft, Make is almost always the cheaper meter for cloud-to-cloud automation, and it is not close. Make's entry paid plan is $9 a month billed annually. One Power Automate Premium seat is $15 a month paid yearly, and you buy one per person. If your company does already run on Microsoft 365, the comparison inverts, because part of what you would be buying from Make you may already own — and the only honest way to settle that is to read your own license agreement, not an article.
The reason these two are compared badly is that they do not bill the same kind of thing. Make bills for work done. Power Automate bills for people, and then separately for robots. Until you convert both to the same unit, "which is cheaper" has no answer.
Make's meter: the pricing page states, verbatim, "Each module action in your scenario, like adding a Google Sheet row or fetching Gmail account data, counts as one credit." You buy credits per month, per organisation: the plans read on 8 September 2026 are priced by credit tier, and headcount is not a line item on any of them. Confirm your own user allowance with Make before you treat seats as free. Power Automate's meter: Power Automate Premium is "$15.00 user/month, paid yearly." Power Automate Process is "$150.00 bot/month, paid yearly." Premium is priced per user. Process is priced per bot. Figures read from both vendors' pricing pages on September 8, 2026.
The two meters, side by side
This is the table the rest of the article rests on. Read the middle column before you read any price.
| Make | Power Automate | |
|---|---|---|
| What you buy | Credits per month | User seats, plus bots for unattended work |
| What consumes one unit | One module action inside a scenario | Nothing — a seat or bot is a flat monthly fee |
| Cost grows with | Volume of steps executed | Number of people licensed, and number of unattended bots |
| Cost of adding a person | Not a priced line item — plans are sold by credit tier, not by seat | $15/month paid yearly, per Premium user |
| Cost of adding volume | Rises with every extra module action | $0 until you hit a service limit |
| Unattended automation | Not a separate line item on the pricing page | Process, $150/bot/month; Hosted Process, $215/bot/month |
| Free option | Free plan, up to 1,000 credits/mo, 2 active scenarios | 30-day trial, per the pricing page |
What a credit actually costs
Make publishes a price grid by credits per month and plan tier. These are the monthly-billed prices in USD:
| Credits/mo | Core | Pro | Teams |
|---|---|---|---|
| 10,000 | $10 | $18 | $34 |
| 40,000 | $34 | $62 | $116 |
| 150,000 | $116 | $180 | $343 |
| 300,000 | $214 | $315 | $600 |
| 1,000,000 | not offered | $880 | $1,647 |
Core stops at 300,000 credits. Above that the cheapest plan available is Pro. That is a hard cliff worth knowing before you build a high-volume scenario on Core and then discover the next step up is a different tier, not a bigger bucket.
Annual billing is advertised as "Save 15% or more." At the published tiers, check it yourself. At 150,000 credits, Core drops from $116 to $99: $116 − $99 = $17, and $17 ÷ $116 = 14.7%. Pro drops from $180 to $153, which is exactly 15%. Teams drops from $343 to $291, or 15.2%. But at 10,000 credits, Core drops from $10 to $9 — 10% — and Pro from $18 to $16, or 11.1%. The 15% figure holds at volume and does not hold at the entry tier. Nothing there is wrong; "or more" is just doing more work at the top of the grid than at the bottom.
Unit price falls as you scale. Take Pro billed annually across three of the published tiers: $16 for 10,000 credits is $1.60 per thousand; $153 for 150,000 is $1.02 per thousand ($153 ÷ 150); $748 for 1,000,000 is $0.75 per thousand. Roughly half the unit price at a hundred times the volume.
Two limits shape the free plan into an evaluation tool rather than a production one: 1,000 credits per month, 2 active scenarios, a 15-minute minimum interval between runs, 5 minutes maximum execution, 512MB of data transfer and 7-day log storage. Core is where 1-minute scheduling and unlimited active scenarios start. Data transfer scales with your plan — the page states "5 GB of data transfers per 10,000 monthly credits" — so a file-heavy workload has a second meter to watch that is not the credit count.
What a seat and a bot actually cost
Power Automate publishes three paid SKUs, all priced paid yearly:
| SKU | Price | Unit | What the page says it covers |
|---|---|---|---|
| Premium | $15.00 | per user/month | Cloud flows and desktop flows in attended mode |
| Process | $150.00 | per bot/month | Cloud flows and desktop flows in unattended mode |
| Hosted Process | $215.00 | per bot/month | Unattended automation with a virtual machine managed by Microsoft and hosted on Azure infrastructure |
Premium is per user. Process and Hosted Process are per bot. That single distinction generates most of the confusion in this comparison. Attended desktop automation is a user capability — a person is at the machine, and you license the person. Unattended desktop automation is a machine capability — nobody is watching, and you license the robot. The $65 difference between Process and Hosted Process ($215 − $150) is what Microsoft charges to also supply and manage the Windows virtual machine the bot runs on, rather than you supplying it.
Price one workload both ways
Fix a workload and run it through both meters. Count your own module actions — Make's page defines a credit as a module action, and only your scenario blueprint tells you how many actions a run performs.
Workload A, small team, cloud only. Three automations, each running 300 times a month, each run performing 10 module actions. That is 3 × 300 × 10 = 9,000 credits a month, which fits the 10,000-credit tier. Make Core, billed annually: $9 a month, or $108 a year. Now price the same work in Power Automate for six people who need it: 6 × $15 = $90 a month, or $1,080 a year. Ten times the cost, for the same three automations.
Workload B, mid-size, cloud only. Six automations, each running 1,000 times a month, each run performing 25 module actions: 6 × 1,000 × 25 = 150,000 credits. Make Core billed annually is $99 a month, or $1,188 a year. Twenty Power Automate Premium seats: 20 × $15 = $300 a month, or $3,600 a year. Make is still about a third of the cost — and if you need roles and shared templates, Make Teams at 150,000 credits is $291 a month billed annually, which is $3,492 a year and now roughly at parity with 20 seats.
Workload C, the same as B but two processes run unattended overnight against a desktop application. Two Process bots: 2 × $150 = $300 a month, or $3,600 a year, on top of the 20 seats. Total $600 a month. Make's pricing page lists no equivalent product, so this is not a like-for-like comparison — it is a capability Make's published plans do not price at all.
The crossover, in seats
Because Make charges by credit tier rather than by seat and Microsoft charges per user, the break-even is just division. Take Make's annual-billed price and divide by $15:
| Make plan (billed annually) | Monthly price | Equivalent Premium seats |
|---|---|---|
| Core, 10,000 credits | $9 | $9 ÷ $15 = 0.6 |
| Core, 150,000 credits | $99 | $99 ÷ $15 = 6.6 |
| Pro, 150,000 credits | $153 | $153 ÷ $15 = 10.2 |
| Pro, 1,000,000 credits | $748 | $748 ÷ $15 = 49.9 |
| Teams, 1,000,000 credits | $1,400 | $1,400 ÷ $15 = 93.3 |
Read the last row plainly: a million module actions a month with teams and roles on Make costs about what 93 Power Automate Premium seats cost. If you have more than 93 people who each need a Premium license, Make's top published tier is cheaper than licensing them — before you count a single bot.
The Microsoft 365 question, stated honestly
Everything above prices Power Automate at list. For most companies already on Microsoft 365, list price is the wrong number, because some automation capability arrives with seats they are already paying for. That changes the marginal cost of adding Power Automate, sometimes to near zero for basic cloud flows.
What is included with which Microsoft 365 SKU is not stated on the Power Automate pricing page, and this article does not guess at it. Getting that wrong by one line item is worth thousands of dollars a year in either direction, so treat it as a question to answer against your own agreement rather than a table to copy.
Three questions to put to your license reseller or Microsoft account team, in writing, naming your exact SKU: First, which Power Automate capabilities does our current subscription include, and does that cover cloud flows using standard connectors only, or premium connectors too? Second, does anything in our current subscription permit desktop flows, and if so is it attended only? Third, does anything we already own permit unattended running, or does that always require a Process or Hosted Process bot at $150 or $215 a month? Check the answers against Microsoft's own licensing documentation at Microsoft Learn's Power Platform licensing pages and the current Power Platform Licensing Guide, which is the document that governs, not the marketing page.
Where each one loses
Make gets expensive when scenarios are long. Because every module action is a credit, cost scales with the number of steps, not the number of jobs. A 40-step scenario running 2,000 times a month is 40 × 2,000 = 80,000 credits before you have automated anything else. The same 2,000 jobs handled in 8 steps is 16,000 credits. The lever is scenario design: fewer, fatter modules, aggregators instead of per-item loops, and filters placed early so branches exit before they spend. On Make, refactoring a scenario is a cost reduction, which is an unusual and useful property.
Power Automate gets expensive when you need unattended bots. A single Process bot is $150 a month, or $1,800 a year, before it runs once. Month against month, that $150 is more than Make's entire Core plan at 150,000 credits billed monthly ($116) and less than Core at 300,000 credits ($214). If your automation genuinely needs to drive a Windows application overnight with nobody logged in, that cost is unavoidable and Make does not offer a competing product. If it does not, you are paying robot prices for API calls.
The differences that are not about price
Where the work runs is the first divide. Make's published plans describe cloud scenarios: schedules, connectors, and HTTP calls executed on Make's infrastructure. Power Automate's plans explicitly describe both cloud flows and desktop flows, which run on a Windows machine — yours, in the case of Process, or a Microsoft-managed VM on Azure, in the case of Hosted Process. If the system you need to automate has no API and only a screen, that difference decides the tool on its own, and price never enters the discussion.
The second divide is people. Make's plan tiers separate credits and features, not headcount. Core adds unlimited active scenarios and 1-minute scheduling. Pro adds priority execution, custom variables and full-text log search. Teams adds teams, roles and shareable templates. Enterprise is custom. Power Automate Premium is priced per user, so the same feature set costs more as adoption spreads. If your goal is to get automation into the hands of forty people in operations, those two models pull in opposite directions.
The third is what surrounds the tool. Each Power Automate SKU carries Dataverse storage entitlements and other bundled allowances; the exact figures are on Microsoft's licensing pages and are worth reading before you assume the seat price is the whole bill. Make prices data transfer instead, proportional to credits. Neither is better; they are different things being metered, and the one that bites you is the one matching your workload's shape.
How to decide in one sitting
Open your list of intended automations and count two numbers. First, total module actions per month: for each automation, steps per run × runs per month, added together. That number picks your Make tier straight off the grid above. Second, the number of people who would need a license and the number of processes that must run with nobody logged in. Those two numbers give you a Power Automate figure: (people × $15) + (unattended processes × $150), per month, paid yearly.
Then apply the one question that overrides both: are you already paying for Microsoft 365, and what does that license already give you? Answer it from your own agreement. If the answer is "nothing relevant," the arithmetic above stands and Make wins on cost for cloud work by a wide margin. If the answer is "quite a lot," Power Automate's marginal cost may be small enough that the comparison is no longer about money at all — and at that point you are choosing between a credit meter you control by writing better scenarios, and a seat meter you control by deciding who gets a license.