Automation and AI

Trigger and Action

Also called If this then that, event and action

The event that starts an automation, and the step it performs in response — the two halves of every rule.

Quick facts: Trigger and Action

Category
Automation and AI
Also called
If this then that, event and action
Level
Beginner
Affects
Automation reliability, duplicate records, response speed
Where to see it
n8n, Make, Zapier, Google Tag Manager, CRM automation rules
In this article4
  1. How triggers and actions work
  2. Why triggers and actions matter
  3. Where they go wrong
  4. How to act on it

How triggers and actions work

Every automation, in every tool, is built from the same pair. Something happens, and something else follows. The trigger is the event being watched for: a form submitted, an order paid, an email received, a date reached, a row added to a sheet, a stage changed in the CRM. The action is what the platform then does: create a record, send a message, update a field, post a notification, call another system.

Triggers are watched in one of two ways, and the difference shows up in real life. Some are pushed to you the instant they happen, because the source system announces the event. Others are polled, meaning the platform checks the source periodically and notices the change on its next look. Pushed triggers feel immediate; polled ones arrive on a schedule, which matters when a customer is waiting for a reply.

Between the trigger and the action sit the parts that make automations usable: conditions that decide whether to continue, formatting that reshapes the data, and lookups that check whether this record already exists. The pair is the skeleton, not the whole animal.

Why triggers and actions matter

Naming the two halves is what turns a vague wish into something buildable. “Follow up with leads faster” cannot be built. “When the contact form is submitted, create a CRM record, send the acknowledgement, and notify the sales channel” can be built by anyone, and can be tested by anyone.

It also shows you what is missing. If you cannot name a trigger, there is no event to hang the automation on, and something upstream has to change first — a spreadsheet nobody updates or a phone call that leaves no record cannot start anything. Many automation projects fail here rather than in the building.

Where they go wrong

Firing more than once is the classic problem. A form submitted twice, a page reloaded, a record edited three times in a minute — each fires the trigger again, and you get duplicate leads, duplicate emails and an annoyed customer. Any action that reaches a person or writes a record needs a check for whether it has already run.

The second is choosing a trigger that is too broad. “Any change to a deal” fires on every trivial edit; “deal stage changed to won” fires when you meant it. Narrow triggers are easier to reason about and cheaper to run.

The third is silence. When an action fails — a wrong format, an expired connection, a required field left empty — many workflows simply stop with no visible sign. Nothing looks broken, and the lead was never created.

How to act on it

Write each automation as one plain sentence before you touch a tool: when this happens, do that, unless this. If the sentence needs three unlesses, it is more than one automation. Prefer the most specific trigger the source system offers, and add a condition immediately after it so the run stops early when it should.

Protect anything that contacts a person or creates a record: look up existing entries first, and use the source system’s own identifier rather than matching on a name someone might have typed differently. Then test the ugly cases deliberately — the same form sent twice, an empty phone field, an unusual character in a name — because those are what real traffic sends, especially from mobile users typing quickly. This is the smallest unit inside any workflow automation platform, and how well you choose it decides how much error handling you will need later. If several of these pairs are already chained together and you have lost track of what fires what, that is the moment to map and rebuild the workflow properly.

Do and do not

Do

  • Write the rule as one sentence before building it
  • Pick the most specific trigger the source system offers
  • Check for an existing record before creating one

Do not

  • Let a person-facing action run without a duplicate check
  • Trigger on any change when you mean one change
  • Assume a polled trigger will fire immediately

Questions people ask about this

Why does my automation run twice for the same lead?

Usually because the trigger fired twice. A visitor submitting the form again, refreshing the confirmation page, or a record being edited soon after creation will each look like a new event. The fix is a check before the action: look up whether a record with that email or reference already exists, and stop if it does.

Why is there a delay before my automation runs?

Because the trigger is probably polled rather than pushed. A polled trigger means the platform checks the source system on a schedule and notices the change on its next look, rather than being told the moment it happens. If speed matters, look for a webhook or a native instant trigger from the source system instead.

Can one trigger have several actions?

Yes, and most useful automations do. A single form submission can create a CRM record, send an acknowledgement, notify a team channel and add a row to a reporting sheet. Keep them in a sensible order, decide what should happen if one step fails partway through, and make sure a failure is visible to someone.

Related terms

Found this useful?

Share it, or ask an AI to summarise it

Back to the glossary

Knowing the term is the easy part

Applying it to your own site and budget is the work. Book a call and I will tell you what actually applies to you.