Using Scheduled Agents

Last updated: July 13, 2026

Scheduled Agents run automatically on a recurring cadence. Use them for workflows that should happen every day, every week, or every month without someone manually rerunning a prompt.

Common scheduled workflows

  • Weekly pipeline digest — summarize changes, risks, and next steps every Friday.

  • Daily at-risk deal alert — flag stalled or slipping opportunities each morning.

  • Monthly competitive intelligence report — analyze competitor mentions from recent calls.

  • Quarterly win/loss review — compare won and lost deals from the previous period.

  • Weekly coaching brief — summarize rep activity and coaching opportunities before 1:1s.

Create a Scheduled Agent

  1. Go to Agents → Command Center.

  2. Click Create an agent.

  3. In Step 1: Trigger, choose Scheduled.

  4. Choose the cadence: daily, weekdays, weekly, or monthly.

  5. Choose the day and hour where relevant.

  6. Write the prompt in Step 2: Prompt.

  7. Choose recipients and delivery methods in Step 3: Output.

  8. Test the agent, then save it.

Choosing a cadence

Cadence

Best for

Daily

High-priority operational alerts, such as stale deals or urgent customer risk.

Weekdays

Sales and CS workflows that should run only during the working week.

Weekly

Digest-style summaries for team meetings, pipeline reviews, coaching, and handoffs.

Monthly

Strategic reviews, market summaries, competitive intelligence, and product feedback rollups.

Write the prompt for recurrence

Scheduled prompts should include a time window, clear criteria, and a stable output format.

Every Friday, summarize pipeline changes from the past 7 days.

Include:
- Total open pipeline by stage
- Deals that progressed or regressed
- Top 5 risks with evidence from recent calls or CRM activity
- Recommended next action for each owner

Format the output as an executive summary followed by a table.

Previous runs context

For scheduled agents, you can enable previous-run context. This lets the agent reference recent prior outputs so it can highlight what changed since the last run.

Use previous-run context for trend-aware workflows, such as:

  • Pipeline health week over week.

  • Competitive mentions over time.

  • Recurring risk reports where you want to know whether risks are new, worsening, or resolved.

Leave it off when each run should stand alone.

Output setup

Use generalized output when everyone should receive the same digest. Use individual output when each user should receive an analysis scoped to their own calls, deals, or accounts.

Troubleshooting

The report is too long

Ask for a maximum number of items, a short executive summary, or a table with specific columns.

The report is too vague

Add evidence requirements such as call quotes, dates, deal names, CRM fields, or specific risk criteria.

The wrong people receive the output

Review Step 3 and confirm whether the agent is using generalized or individual output. Also confirm Slack and email delivery settings for each recipient.