Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.usecustory.com/llms.txt

Use this file to discover all available pages before exploring further.

Custory AI can help you build and refine automations directly from the automation editor. This is useful when you know the outcome you want, but do not want to manually configure every trigger, action, and integration detail from a blank screen.

​
What automation AI mode can do

In automation mode, Custory AI can help edit:
  • Trigger choice
  • Action choice
  • Integration-backed properties
  • Message and issue-writing instructions
  • Overall automation structure
The automation chat is not a separate brainstorming tool. It is tied to the actual automation you are editing.

​
Save first, then edit with AI

Custory requires the automation to be saved before AI can edit it in place. That matters because AI needs a real automation record to work against. If the automation is not saved yet, the editor will ask you to save it first before opening automation chat.

​
How to open automation chat

Open the automation in the editor, then use the Custory AI entry point from the workflow header. Custory saves the current draft first, then opens AI in automation mode with the current:
  • Workspace
  • Journey context where relevant
  • Automation ID

​
What AI has access to

In automation mode, AI can work from:
  • The current automation structure
  • Connected integrations
  • Workspace context
  • Journey context where relevant
If required integrations are missing, the chat can point you toward connecting them first.

​
Best use cases

​
Start from the goal

Example prompts:
  • I want a weekly Slack summary of the most important changes in this journey.
  • When a high-priority opportunity appears, create a Linear issue and tell the team in Slack.
  • When a GitHub PR merges, update this journey only if the change affects customer experience.
This is the highest-leverage way to use automation chat because the AI can choose and refine the structure around a clear operational outcome.

​
Refine an existing automation

Good examples:
  • Switch this trigger to something better for the goal.
  • Refine this automation so it is less noisy.
  • Update this to create Jira issues instead of Slack messages.

​
Improve AI instructions

Good examples:
  • Make the Slack update more concise and decision-oriented.
  • Rewrite the issue instructions so engineers get clearer customer context.

​
Suggested automation chat prompts

Custory surfaces automation-oriented suggestions such as:
  • Add Slack action
  • Switch trigger
  • Add Jira action
  • Add Linear action
  • Refine this automation
These are useful shortcuts when the team wants to make an existing automation better instead of starting from zero.

​
What AI is best at here

Automation AI is strongest when the job is:
  • Translating intent into workflow structure
  • Tightening message or issue-writing instructions
  • Converting a manual habit into a repeatable flow
It is less useful when the team has not yet decided what the workflow should accomplish.
  1. Start with the smallest useful automation
  2. Save it
  3. Open automation chat
  4. Ask AI to improve one part at a time
  5. Review the result before activating
That sequence is usually better than asking AI to invent a large multi-step workflow all at once.

​
Common mistakes

​
Asking AI for automation before the goal is clear

If the team cannot explain what the workflow should accomplish, AI usually produces something technically plausible but operationally weak.

​
Skipping the review step

AI can help configure the workflow, but the team should still confirm:
  • Trigger sensitivity
  • Integration routing
  • Target journey or step
  • Message quality

​
Using AI to compensate for stale data

If the journey structure, statuses, or priorities are weak, AI will have less to work from. Clean the context before trying to automate it.