You will be able to list the common causes of automation failure and recognise each one in a run history.
Mei Ling's enquiry flow ran without trouble for eleven weeks. Then, one Monday, a parent called to ask why nobody had replied to her enquiry from the previous Thursday, and when Mei Ling opened the sheet, the last row was from Wednesday afternoon. Since then, five enquiries had come in through the form, and none had reached the sheet, sent a confirmation or created a reminder. The automation tool had not emailed her. Nothing looked wrong until someone complained.
The cause turned out to be small. On Wednesday she had changed the password on the centre's email and sheets account, as the IT contractor had advised, and the automation tool's connection to the sheet had stopped working.
Plenty of automations fail this way, quietly, with nothing on screen to say so. This lesson covers the usual causes, and how to recognise each one in your tool's run history.
Every app you connect to an automation tool is linked by a connection, a stored permission that lets the tool act as you. Connections stop working for ordinary reasons. You change your password. Someone in IT resets access or turns on a new security setting. The app expires the permission after a period of time. The person whose account was used leaves the organisation and their account is closed.
When a connection stops working, every step that uses it fails. Depending on the tool and the step, the run may stop with an error or the trigger may stop firing at all, so there are no runs to fail. In run history this shows up as errors mentioning authentication, authorisation, permission, an expired token or "reconnect", or as a sudden gap where runs used to be.
This is why lesson 3.3, What a tool can see once you connect your accounts, suggested dedicated accounts. A shared enquiries account whose password changes on a known schedule is easier to manage than a personal account that changes whenever its owner feels like it.
A flow maps fields by name: a sheet column called "Phone", a folder called "Invoices 2025", a form question called "Child's level". Rename any of these and the mapping can break, even though the flow has worked for months.
Some tools fail with an error saying a column or field cannot be found. Others carry on and write nothing into the renamed column, or write into the wrong place, and report success. The second kind is worse, because the run history looks healthy. You only notice when you look at the data itself.
Lesson 4.1, Plan the record that ties form, sheet and email together, suggested protecting header rows for this reason. The same applies to form questions and folder names. In run history, look for "not found" errors, or for runs that succeeded but whose output has blanks where values should be.
Real people type things you did not plan for. An empty field where you assumed there would always be a value. A message several thousand characters long. An emoji in a name. A phone number with letters in it. An attachment in a format the next step cannot read.
Each of these can stop a step. A formatter expecting a date fails on "next Tuesday". An email step fails because the recipient field is empty. An AI step returns something your filter does not recognise. In run history these show as errors on one particular run, while the runs before and after work fine. That pattern, one run failing among many good ones, almost always means unusual input. Lesson 2.3, Format dates, names and numbers on the way through, covered setting defaults for blanks, which prevents many of these.
The last group of failures comes from outside your flow entirely.
Apps limit how many requests they accept in a given time, often called rate limits. A burst of activity, such as a hundred form responses after a school holiday promotion, can hit them, and some runs fail or wait. Apps and automation tools also have outages, when they are simply not working for a while. And if your automation plan has a usage allowance, running out of it can pause your flows until the next billing period, as lesson 3.1, Zapier, Make and n8n: three ways to build the same flow, warned.
Some tools retry automatically after these failures, and some notify you, but you cannot assume either. In run history, look for errors mentioning limits, quotas, timeouts or "service unavailable", and check your account's usage page.
All four causes share a problem. The person who built the flow has usually stopped looking at it. It worked for weeks, so nobody opens the run history. Notifications from the automation tool, if they exist, go to an inbox nobody reads, or get filtered as bulk mail. And a trigger that stops firing produces no errors at all, only an absence.
So the first job is simply to look. Open the run history of every flow you have built in this course, including the test flows from lesson 3.4, and read through it. The activity below asks you to write down every error so far, its cause, and how you found out about it, which will show you how many you would have missed.
Open the run history of your flows and write down every error so far, with its cause and how you found out.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).