Run a rule on historical conversations

Manual Run lets you apply an automation rule to existing conversations — not just future events. Use it for one-off backfills (tag every existing returns conversation), clean-up sweeps (close ancient Done conversations), or bulk re-application after the rule logic changes.

The flow is two-step on purpose: Simulate first, then Execute. Automation actions can be irreversible, and you want a chance to review the matches before pulling the trigger.

Before you start

• A working automation rule, ideally one you’ve already built and tested on a few live events
• A clear answer to “what happens if this runs against the wrong conversations?” — if any action is destructive (sending messages, closing cases, firing webhooks), be especially careful

Steps

1. Give the rule a manual trigger. Go to Settings → Automation rules and open the rule. Its trigger must be Run by hand; only rules with that trigger can be simulated and executed against existing conversations.

2. Narrow the run with the rule’s conditions. There is no separate scope picker: Simulate evaluates the rule’s conditions against every conversation in the workspace (spam excluded). To limit the run, add conditions to the rule, for example Conversation status is Active and Minutes since the last message (from the customer, an agent or the Agent Stack) is more than 20160 (14 days).

Timestamp conditions are available as age fields, like Minutes since the last message (from the customer, an agent or the Agent Stack), or date fields, like Date of the last message (from the customer, an agent or the Agent Stack). Age values are in minutes; date values use the date picker.

Make the conditions as narrow as what you mean to backfill. Broader conditions mean more matches and more risk if the rule is wrong.

3. Simulate. In the Rules tab, click Simulate this rule (the play icon) on the rule’s row, then confirm Simulate. Atender evaluates the rule’s conditions against every conversation. No actions run. You are taken to the Manual executions tab, where the run shows its Evaluated and Matched counts.

4. Review the matches. Click View matches on the run to open Matched Conversations, and spot-check a handful. Confirm they’re the kind of thing you actually want the rule to act on. If anything looks off, return to the rule and tighten conditions, then simulate again.

5. Execute. Once satisfied, click Execute on the run and confirm. The rule’s actions run against every match in the simulated list. If more than 20,000 conversations matched, the confirmation tells you the run holds only the first 20,000, by conversation id.

6. Watch the run. The Manual executions tab lists each run with its rule, status, and Executed and Errors counts.

Verify it worked

After the run completes, sample a few of the matched conversations and confirm the action’s effect:

• Tag added → tag is visible on the conversation
• Status changed → status reflects the new value
• Message sent → outbound message is in the timeline (and has the expected X-Email-Template header for Email Studio templates)
• Webhook fired → check the downstream system received the payload

Notes

• Manual Run ignores the rule’s Limit how often it runs and Only during opening hours settings. It runs on every match when you click Execute, regardless of opening hours or how often the rule ran before. Those settings only govern runs that an event starts.
• Long backfills run asynchronously. You can leave the page; the run continues in the background. The Manual executions tab shows the final result.

Troubleshooting

• Symptom: Simulate shows zero matches. Fix: The conditions are too narrow. Loosen them and simulate again. If Simulate this rule doesn’t appear on the rule’s row, the rule’s trigger isn’t Run by hand.

• Symptom: Execute fires actions on more conversations than Simulate showed. Fix: This shouldn’t happen. If it does, contact support with the rule ID and run timestamp.

See also

• Simulate a rule before enabling
• Debug a rule using execution history