What Ongoing Automation Maintenance Actually Costs
What ongoing AI automation maintenance actually costs: tool and AI usage fees, fair upkeep, why run cost should sit below build cost, and retainer red flags.
You paid for the build. Now there is a monthly number sitting next to it, and you want to know if it is fair. The honest answer: ongoing cost has two parts, and together they should be small. There is the tool and AI usage cost, the platforms the automation runs on, which is predictable and modest. And there is upkeep, the watching and fixing, which should be light. Both together should sit well below what you paid to build the thing. When the monthly number rivals the build, something is wrong.
We will name the parts in the open. If a quote bundles “ongoing” into one fat figure with no breakdown, that is the first red flag, and this post will show you the rest.
The Bottom Line
- Ongoing cost has two parts: tool and AI usage (Make.com, n8n, your AI spend) plus upkeep. Both should be small.
- Run cost should sit well below the build cost. A maintenance figure that rivals the build every month is a red flag.
- A fair retainer names a deliverable: monitoring, fixes, a set number of new automations. “Ongoing optimisation” that ships nothing is a subscription to hope.
- When it is done-for-you, the upkeep is the builder’s job, and you should still own every line on accounts in your name.
What Are The Two Parts Of A Run Cost?
Ongoing cost splits cleanly into two buckets, and a fair quote shows both. First, the platforms the automation runs on: tool subscriptions and AI usage. Second, the upkeep: the monitoring and fixing that keeps it alive. The first is predictable and modest. The second should be light. Lumped together with no breakdown, you cannot tell which one you are overpaying for.
The reason this matters is simple. Tool costs are mostly out of anyone’s hands, set by Make.com or your AI provider. Upkeep is a service someone charges for. When a quote hides both inside one number, you lose the ability to push back on the part that is actually negotiable. Ask for the split. A builder who has nothing to hide will give it to you on the spot.
What Do Tool And AI Usage Costs Cover?
This is the smaller, more predictable half. Tool and AI usage covers the platforms the automation actually runs on: a Make.com or n8n subscription to host the scenarios, and your AI usage for any step that calls a model. For most single automations this is a modest monthly figure, often tens of dollars, scaling with how often the flow runs and how much it does.
Make.com charges by operations, roughly how many steps fire per month. A flow that runs nightly costs little. One that fires on every inbound email costs more, but still rarely much. n8n can be cheaper at volume if you self-host, which is one of the trade-offs we cover in Make.com vs n8n. AI usage is metered too: a few cents per call, adding up only if the automation runs thousands of times a month.
The point is predictability. These costs track usage, not a builder’s margin. You can estimate them before you sign, and they should never surprise you. If someone quotes a vague “platform fee” with no link to actual usage, ask what sits underneath it.
What Should Maintenance Actually Cost?
Upkeep should be light, and that is the headline. Real maintenance is monitoring that catches a flow when it stops, renewing credentials before they expire, adapting flows when a platform changes, and testing after every change. That is steady, low-touch work, not a second build. For a single automation or a small stack, it should be a small slice of the build cost per month, not a figure that competes with it.
The honest framing: most of the cost of an automation is building it. Once it runs, keeping it alive is cheap, because nothing is being created, only watched and occasionally patched. The work spikes only when a third party changes something underneath you, and even then it is a fix, not a rebuild. We cover what that work involves, and why automations rot without it, in how to maintain automations.
So the rule of thumb is a ratio, not a dollar amount. If upkeep is a small fraction of the build per month, that tracks. If it is half the build every month, you are paying build money to run something that is already built. Ask what that figure produces.
What Are The Red Flags In A Maintenance Quote?
A few patterns tell you a maintenance quote is padded. Spot these and push back, or walk. The clearest one: a retainer that rivals the build cost every single month. You paid once to build it. Paying nearly that again every month to run it makes no sense, because running is the cheap half. That number is margin, not maintenance.
Three more to watch for.
A retainer with no named deliverable. “Ongoing support” or “ongoing optimisation” with nothing shipped each month is a subscription to hope. A fair retainer names what it produces: monitoring with alerts, fixes when something breaks, a set number of new automations. If it names nothing, it should not exist.
Tool costs marked up. The Make.com and n8n subscriptions and AI usage should be billed at cost, or paid directly by you on your own accounts. A builder adding a fat margin on top of a platform fee you could pay yourself is charging you to be a middleman.
Bundled with lock-in. If the only way to keep the automation running is to keep paying a retainer, because the whole thing lives on someone else’s accounts, that is not maintenance. That is rent. The test is one question: if I stop the retainer tomorrow, does the automation keep running? It should. For how this plays out across the whole industry, see how agencies price automation.
Whose Job Is The Upkeep?
When it is done properly, the upkeep is the builder’s job, not another task on your plate. That is the position most builders avoid saying out loud, because it commits them to work after the invoice clears. You own the system. Someone else watches it, catches the rot, renews the credentials, and fixes the breaks before they cost you. That is what the monthly figure should buy.
None of that costs you ownership, and this is the part that separates a fair retainer from a leash. You still own every line. The Make.com and n8n scenarios, the logic, the credentials, all of it sits on accounts in your name, with no lock-in to the people maintaining it. If you stopped paying for upkeep tomorrow, the automation keeps running. You would lose the watching, not the asset. That is the difference between buying a service and renting a tool.
So the reframe is this. Budget for run and maintain, not just build, but expect both to be small. The build is the expensive part. Keeping it alive is cheap when someone competent is doing it, and it should never feel like paying for the build twice.
Frequently Asked Questions
How Much Should Ongoing Maintenance Cost Per Month?
A modest figure that sits well below the build cost. It covers tool subscriptions like Make.com or n8n, your AI usage, and light upkeep: monitoring, credential renewals, and the odd fix. For a single automation it can be tens of dollars in tools plus a small upkeep slice. If the monthly number rivals what you paid to build it, that is a red flag worth questioning.
Why Is The Run Cost So Much Lower Than The Build Cost?
Because building is the expensive part and running is not. The build is where the hours go: mapping your tools, wiring integrations, handling edge cases, testing. Once it runs, nothing new is being created. The platforms charge by usage, and upkeep is just watching and the occasional patch when an API changes. Cheap to run is the normal, healthy state, not a discount.
What’s A Fair Thing To Pay A Retainer For?
A retainer should name a deliverable. Fair ones cover monitoring with alerts when a flow stops, fixes when a platform changes something underneath you, credential renewals before they lapse, and often a set number of new automations each month. You are paying for things shipped and a system kept healthy. “Ongoing optimisation” with nothing to point to each month is not a deliverable.
Do I Have To Pay Maintenance Forever To Keep My Automations?
No, and that is the test for any builder. A proper build sits on accounts in your name, so the scenarios keep running whether or not someone maintains them. Maintenance buys you the watching and the fixing, not the right to keep the asset. If stopping the retainer means the whole thing vanishes, that is lock-in, and it is a different product than the one you should be buying.
The build is the expensive half. Running and maintaining an automation should be cheap: a modest monthly figure for the tools it runs on, plus light upkeep that sits well below what you paid to build it. If a maintenance quote rivals the build every month, names no deliverable, or only works while you keep paying, those are the red flags to push on. With an AIOS, you own every line on accounts in your name, the upkeep is ours, and the monthly number stays small and honest. If you want a straight breakdown of what your automations should cost to run, no padding and no lock-in, Get In Touch.
Sam co-founded Echelon AI Solutions and leads transformation strategy, client engagements and growth. He has built and operated businesses across marketing and AI education, and has guided companies in retail, trades, hospitality and professional services through operational change. His focus is making AI earn its place through measurable business performance.
More In Cost & Pricing
See all Cost & Pricing →
What AI Automation Costs In Australia (And Payback)
What a custom AI automation build really costs in Australia, the four things that drive the price, and the payback maths to run before you sign.
Read itHow Much Does AI Automation Cost For An Australian Business?
AI automation cost in Australia tracks your scope, not a sticker price. The four drivers that set the number, cheap vs built-to-last, and how to judge payback.
Read it
What Actually Drives The Price Of An Automation Build
The four things that drive the price of an AI automation build: integrations, data quality, edge cases and human review, so you read a quote like an operator.
Read it