Business Strategy
How to Price an AI Automation Retainer for a Client
How to price an AI automation retainer: what a retainer actually has to cover beyond the automation itself, the Break-Fix Floor that sets your minimum, and the Watch-Build-Grow ladder for structuring three tiers instead of one flat number.
Founder, AI Tools and Training Club · August 29, 2026 · 9 min read
The short version
- An AI automation retainer isn't priced on what the automation costs to run - it's priced on what breaks without you, what gets missed without monitoring, and what you build next. Pricing to your own tool costs is the single fastest way to underprice the work.
- Set a Break-Fix Floor first: the minimum monthly number that covers your own time if the client's automations run quietly and nothing breaks all month. Never sign a retainer under that floor, no matter how much the client pushes back.
- Structure the offer as a ladder - Watch, Build, Grow - instead of one flat monthly fee. Each tier adds a specific, nameable thing on top of the last, so upgrading is an easy yes instead of a renegotiation.
The short answer
Price an AI automation retainer on what it protects and what it adds each month, not on what the underlying automation platform or API calls cost you to run. Start by setting a Break-Fix Floor - the minimum that covers your own time if you spend the whole month just keeping things working and nothing breaks - then build a Watch-Build-Grow ladder on top of it so a client can see exactly what a higher tier buys. A retainer priced to your tool costs is underpriced by definition, because the client isn't paying for API calls. They're paying for the automation working when they're not thinking about it.
What a retainer actually has to cover
A one-off automation build and a retainer are different products, even when the automation itself is identical. The build is a project with an end date. The retainer is an ongoing promise that the thing keeps working, and 'keeps working' is doing more than it sounds like.
- Monitoring - noticing when an automation silently stops running, before the client notices a lead never got followed up with.
- Break-fix - fixing it when a connected app changes its API, a login expires, or a field gets renamed upstream and the automation starts failing quietly.
- Small changes - the client's process shifts slightly and the automation needs a tweak, not a rebuild.
- Reporting - telling the client what actually ran, what it caught, and what it's worth, because an automation nobody sees the results of gets cancelled at renewal.
The mistake: pricing to your own tool costs
The instinct when pricing anything AI-related is to start from what it costs you to run - the automation platform's subscription, the API usage, maybe an hour of your time. That number is almost always small, and if you price a margin on top of it, you end up charging for infrastructure instead of the outcome. The client isn't buying API calls. They're buying the fact that the process runs, on its own, without landing back on their desk.
This mistake compounds on a retainer specifically, because it repeats every month. Underpricing a one-off project costs you once. Underpricing a retainer means you're locked into a number that doesn't cover your real time the moment something breaks in month three, and renegotiating a retainer upward is a harder conversation than pricing it right on day one.
The Break-Fix Floor
The Break-Fix Floor is the minimum monthly retainer price you'll accept, calculated from your own time, before you add a single dollar of value-based pricing on top. It answers one question: if this client's automations run perfectly all month and nothing breaks, is the number still worth having them as a client?
Calculate it from three inputs: your own hourly rate (or the rate you'd want if you outsourced the monitoring), the realistic hours you'll spend in a quiet month just checking things are running and answering the client's questions, and a buffer for the one bad month a quarter where something actually breaks. Add those together and that's your floor. Anything you charge below it is a client you're subsidizing, not a client you're serving.
| Input | What it captures | Why it matters |
|---|---|---|
| Your baseline hourly rate | What an hour of your time is actually worth | A retainer priced without this defaults to whatever feels reasonable, not what your time costs |
| Quiet-month hours | Checking dashboards, answering a Slack message, confirming a run completed | This is real time even when nothing breaks - it's easy to forget to price it |
| Bad-month buffer | One real break-fix incident roughly every quarter, averaged into the monthly number | Without this, the one month something actually breaks turns the retainer into a loss |
Building your Break-Fix Floor
The Watch-Build-Grow ladder
Once the floor is set, structure the offer as three tiers instead of one flat number, so a client can see exactly what moving up buys them instead of asking you to just discount the one price you quoted.
- Watch - monitoring and break-fix only, priced at or just above your Break-Fix Floor. The automation keeps running, gets fixed if it breaks, and the client gets a monthly note confirming it ran.
- Build - Watch, plus one new automation or meaningful expansion added each month. This is where most of the real value-based pricing lives, because you're shipping something new the client can point to.
- Grow - Build, plus reporting that ties the automation's output to a number the client already tracks - hours saved, leads followed up, invoices sent on time - so the retainer's value is visible at renewal instead of assumed.
When to raise the price
Raise a retainer once you have a real track record to point to, not on a fixed schedule. Ninety days in, you'll know the actual maintenance load - how often something broke, how many small changes you made, whether the client's usage grew past what you originally scoped. That real data is a much stronger basis for a price change than 'it's been a year.'
The easiest way to raise a retainer without it feeling like a renegotiation is to tie the increase to something that changed - the client added three new workflows since you started, or usage grew past the volume the original price assumed. A price increase tied to a specific, visible change is a conversation. A price increase tied to time passing is an argument.
How I coach members through this
When a member tells me they're losing money on a retainer, the first thing I ask is whether they ever calculated a Break-Fix Floor or just picked a number that felt fair to the client. Almost always it's the second one, and almost always the number was set by imagining a perfect month where nothing goes wrong. [How to write business proposals with AI](/blog/how-to-write-business-proposals-with-ai) covers the same discipline for one-off project pricing - the Scope-Price-Proof framework there and the Break-Fix Floor here are solving the same problem in two different sales motions.
[How to automate your business with n8n](/blog/how-to-automate-your-business-with-n8n) is usually the build members are pricing when this question comes up - if you're structuring your first automation retainer around an n8n build, price the maintenance load honestly before you quote the number, not after the first thing breaks.
Frequently asked questions
How much should I charge for an AI automation retainer?
There's no single number that applies across clients or automations, and repeating one here would be a guess dressed up as an answer. Start from your own Break-Fix Floor - your time cost for monitoring and fixing, plus a bad-month buffer - then add value-based pricing on top for what the automation is actually worth to that specific client.
Should I ever do an AI automation retainer without a break-fix component?
No. Even a simple automation depends on connected apps that change their APIs, rename fields, or expire logins without warning. A retainer that only covers new builds and ignores maintenance will have the existing automation quietly fail while you're both looking at the next project.
What's the difference between a retainer and just billing hourly for fixes?
Hourly billing for fixes means the client only pays when something visibly breaks, which gives you no incentive to monitor proactively and gives them no way to budget for it. A retainer bundles monitoring, break-fix, and incremental work into one predictable number, which is what makes the ongoing relationship worth having for both sides.
How do I raise the price on an existing retainer without losing the client?
Tie the increase to something specific that changed - more workflows added, usage that grew past what the original price assumed - and say so plainly. A client can argue with 'it's been a year.' It's much harder to argue with a documented increase in what you're actually maintaining.
Can AI help me calculate the right retainer price?
AI can help you organize the inputs - your hourly rate, estimated hours, a bad-month buffer - into a clean number, the same way [how to use AI for pricing](/blog/how-to-use-ai-for-pricing) describes for general pricing decisions. It can't tell you what your own time is worth or how often things will actually break for a specific client. Those inputs still have to come from you.