- Why do automations break after they go live?
- Three ways to look after automations once they're built
- What should an automation maintenance checklist include?
- What happens when an automation breaks?
- What to agree before the build ends
- Where I fit when you're planning who looks after your automations
- Frequently asked questions
Someone has to own every automation after the build: a named person on your team, the consultant on call, or regular check-ups, ideally agreed before the project ends. Automations break when apps update, connections expire or form fields change, so you'll need error alerts and written notes either way. I'm Lilach Bullock, I build automations in Make.com, Zapier and n8n, and I'm one option if you'd like help planning the handover.
Last checked: October 2026.
Automations feel finished on the day they go live. They aren't. They're more like a car: they run nicely until something small changes, and then they stop at the least convenient moment.
That's not a sign of bad work. It's the nature of tools that depend on other tools. The question isn't whether something will break, it's who notices, who fixes it and how quickly. This guide covers why automations break, the three common ways to look after them, and what to agree before your consultant hands over.
Why do automations break after they go live?
Most breakages come from a short list of causes, and almost none of them are anyone's fault:
- App updates. One of the connected tools changes how it works, renames a field or retires an old connection.
- Expired connections. Someone changes a password, leaves the business, or a login simply times out, and the automation loses access.
- Changed forms and fields. A team member adds a question to a form or renames a column in a spreadsheet, and the workflow can't find what it expects.
- Usage limits. A busy week pushes the number of runs or requests past what a plan or connected app allows.
- AI output drift. An AI step starts producing slightly different wording or format, and the next step can't read it.
- Odd data. A blank field, an emoji in a name or a duplicate contact that nobody planned for.
If you'd like to see how many of these could already be lurking in your set-up, a checklist for auditing your automations is a calm way to review them.
Three ways to look after automations once they're built
Small businesses usually settle on one of three set-ups, or a mix. None is right for everyone.
| Set-up | Suits | Watch out for |
|---|---|---|
| An in-house owner | Teams with someone organised who's been shown how each workflow works | Everything resting on one person, with no cover for holidays |
| The consultant on call | Businesses with complex or customer-facing workflows and no technical person | Slow fixes if the consultant is busy, and you knowing less over time |
| Regular check-ups | Stable set-ups that rarely change | Problems that happen between check-ups going unnoticed |
A common, sensible mix is an in-house owner for day-to-day alerts, with the consultant available for anything bigger. Whatever you choose, write it down and make sure everyone involved knows.
What should an automation maintenance checklist include?
Whoever looks after your automations, these are the basics. If you only do a few, start at the top.
- Error alerts to a named person. Failures should land with someone on your team who'll act on them, not in an inbox nobody reads.
- Run history reviews. A quick look at recent runs every week or two shows failures, retries and anything running far more or less often than usual.
- Up-to-date documentation. For each workflow: what it does, its trigger, the apps it touches and who owns it.
- A test record. A known test entry you can push through after any change, so you can see it works end to end.
- Owner accounts. Every automation and connection set up under business-owned logins, not a freelancer's personal account or a former employee's email.
- A change log. A simple note of what changed, when and why, so you can trace a problem back to its cause.
- Checks on AI steps. Someone glances at a sample of AI output now and then, especially anything that reaches customers.
What happens when an automation breaks?
With the basics in place, a breakage is usually a short, boring fix. The alert arrives, the owner opens the run history, sees which step failed and why, and either fixes it, such as reconnecting an app, or passes it on with the details.
Without them, the story's different. Leads sit unrouted, customers don't get their confirmation, or a report quietly shows the wrong numbers. Lead handling is where this tends to hurt most, which is why this walkthrough on routing leads automatically builds the safety checks in from the start.
A good rule: anything customer-facing should have a fallback. If the automation fails, the enquiry should still land somewhere a human will see it.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
What to agree before the build ends
The best time to sort out maintenance is before the consultant finishes, while everyone still remembers how things fit together. Agree these points in writing:
- Who owns each workflow day to day, by name
- Who receives error alerts, and who covers when that person's away
- How your owner will be trained, and whether there's a recorded walkthrough
- Where the documentation lives and who keeps it current
- What counts as a quick fix your team handles, and what goes back to the consultant
- How to reach the consultant later, and roughly how quickly they can respond
- When the first check-up will happen, usually a few weeks after launch
If a consultant is reluctant to talk about any of this, that's worth noticing. A good one wants you to be able to cope without them.
Where I fit when you're planning who looks after your automations
I build automations in Make.com, Zapier and n8n for small businesses and marketing teams worldwide, working remotely, and I plan the handover from day one: accounts you own, alerts to your person, written notes and a recorded walkthrough. I'm a good fit if you'd like someone on your team to own the workflows confidently, with me there for the bigger changes.
I'm not the right fit if you need round-the-clock monitoring or an IT department to run your systems for you. If the first description sounds right, here's how I hand over the automations I build.
Frequently asked questions
How often should automations be checked?
A quick look at error alerts as they come in, a glance at run history every week or two, and a fuller review every few months or after any big change to your tools.
Why did my Zapier automation suddenly stop working?
Usually an expired connection, a changed form or spreadsheet field, or an update to one of the connected apps. The run history will show which step failed.
Who should own the automation accounts?
Your business. Set them up under a business-owned login so you don't lose access when a freelancer or team member moves on.
Do AI automations need more maintenance than normal ones?
A little more. AI output can change format or tone over time, so someone should check a sample regularly, especially where it reaches customers.
What should I do if my automation consultant disappears?
If the accounts are yours and the workflows are written down, another builder can pick them up. That's why ownership and notes matter from day one.
