Skip to content
⌘K
InboxAsk a human

Scheduled Agent Automations

Updated by Vaishnavi

Use scheduled Enterpret Agent automations for daily reports, weekly digests, monthly summaries, Voice of Customer reports, and account-health updates. A run can send a report to Slack or email, take supported actions in connected tools, or do both.

For one HTML page you revisit and refresh, use Live Artifacts. Use an automation for recurring delivery. Capabilities depend on the automation type, available tools and Skills, connected accounts, and saved permissions; actions requiring interactive approval may not run unattended.

Set up a daily report

  1. Open Automations and create a scheduled automation, using a template if useful.
  2. Describe the product area or accounts, sources, segment, reporting window, comparison period, and output format.
  3. Choose Daily, a run time, and timezone. Choose weekly or monthly for less frequent reporting.
  4. Choose a Slack channel, Slack DM, or email destination. Add reference files, connectors, and Skills as needed; review permissions.
  5. Review and choose Create. Confirm the saved status. Creation enables future runs; use Run now where offered to check output and delivery, review run history, and pause future runs while making changes if needed.

Example brief: “Prepare a daily report on [product area] for [team]. Cover the last 24 hours, compare with the previous 24 hours, and highlight new complaints, feature requests, and significant changes. Include supporting feedback links and representative quotes. Deliver the report to [Slack channel or email recipients].” A prompt alone does not activate recurring delivery; configure the schedule and destination.

Create a scheduled automation

Describe the job → Set the schedule → Configure delivery and tools → Review → Create → Run now where offered.

The walkthrough below uses a mobile-onboarding PM’s weekly digest: what changed, emerging or persistent themes, representative quotes, and delivery to the team’s Slack channel every Monday morning.

1. Describe what to analyze

Write the brief as you would in an Agent Session. Include sources, filters, scope, time window, comparisons, evidence, and output. Agent can research available feedback in your Context Graph; check its resulting counts, quotes, and links rather than assuming every claim is cited.

Generate a weekly Voice of Customer digest for the mobile onboarding product area. Cover the past 7 days against the prior 7 days, and check the past 4 weeks for persistent themes. For each significant theme, give the volume count and week-over-week delta, classify it as emerging, trending, or persistent, and include two or three verbatim quotes with source links. Lead with what changed, not what exists. Prefer tables over prose.

Use the refine action, where available, for help tightening scope, comparisons, and output structure. Review and adjust its draft before continuing.

Attach reference files

Add a rubric, report template, or other reference that every run should read.

Entry pointWhat to do
Automation editorIn Instructions for action, select Attach files, or type / and select the attachment action. Choose your files; they appear above the prompt.
Enterpret Agent chatUpload the file and ask Agent to create a new automation that uses it. Identify the intended file if the chat contains several. Review the saved automation’s instructions and attached files.
Automation instructions with Attach files highlighted in amber and an example Review rubric.pdf file above the prompt.
Select Attach files in Instructions for action. Added files appear above the prompt. Example file.

Explain how to use each file. Click a saved file to preview it where supported; other supported types may download. In edit mode, use the file’s remove control to replace or remove it, then Save. Review the saved file list before relying on future runs. Chat-based creation can attach selected uploads; to change files on an existing automation, use its editor.

RequirementLimit or supported format
File countUp to 10 files per automation. If a selection would exceed the limit, select a smaller batch; that batch is not added automatically up to the remaining capacity.
File sizeEach file must be smaller than 10 MiB (10,485,760 bytes). A file exactly at that size is rejected; the uploader’s message calls this “10 MB.”
Documents and dataPDF, DOCX, XLSX, CSV, TSV, JSON, JSONL/NDJSON, Markdown, plain text, HTML, and GZ.
ImagesPNG, JPEG/JPG, and WebP.

Adding a file with the same full filename replaces that file in the draft; Save applies the replacement. If several selected files share a name, only the last one is kept. Reference files are inputs to the job, separate from generated report attachments.

Save file changes before Continue with Agent if you want to keep them; the confirmation explains that unsaved file changes do not carry over. If connecting Slack takes you away before new files are saved, follow the warning and attach them again when you return.

2. Set the schedule

Choose the frequency, day, time, and timezone. For the weekly example, use Monday at 8:00 AM in the team’s timezone. A scheduled time makes work eligible to start; it is not a delivery deadline. Allow time for execution, delivery, and review. Check run history for delays or failures.

3. Choose what happens with the result

ResultConfigure
ReportA Slack channel, Slack DM, or email destination. For the example, choose the team’s Slack channel.
Connected actionThe supported ticket, record-update, or messaging tools for the configured account, with saved permissions.
BothConfigure delivery and the connected action separately.

Connecting a provider does not grant every action. Account access and unattended permissions still apply.

4. Add connectors and Skills when needed

  • Default context: available Context Graph feedback and built-in analytical Skills.
  • Connectors: external context or supported actions in tools such as Slack, Notion, Linear, and PostHog.
  • Skills: reusable analysis methods. Select an available Skill and test that the result follows the intended method and scope.

Select the automation’s account and tool permissions. Interactive Session approvals do not automatically authorize background work.

5. Review before creating

Check the prompt, reference files, schedule, timezone, accounts, tool permissions, and delivery. The scheduled editor has no separate execution test before saving.

6. Create and check the saved automation

Choose Create and confirm its saved status, cadence, files, connections, permissions, and destination. Future runs are enabled. After creation, the creator can use Run now where offered. Check scope, evidence, quotes, links, format, output, and delivery; adjust and test again when needed. Access and service availability can still affect execution and delivery.

Start from a template

Tell Agent what you want, such as “Create a weekly Voice of Customer report.” It shows matching templates and guides setup with your choices. Browse available templates in your workspace’s Automations page.

PatternUse it forAudience
Product Area VoCOne product area: what changed; emerging, trending, or persistent themes; quotes and a driver hypothesis for each.Product-area PM
Account Health MonitoringA defined roster, such as top accounts by ARR, a CSM book, or a segment. Names the roster, accounts getting louder or quieter, risks, and wins.CS and post-sales leaders
Org-Wide VoC ReportA five-minute company-wide readout of what worsened, improved, or appeared across feedback.Product, support, and CX leaders
Ticket triageFind new high-priority feedback issues and file or tag them in a connected tracker such as Linear or Jira before the day starts.Support and triage owners
CRM account flagsFlag rising feedback risk and update the connected CRM account record where its owner works.CSMs and account owners

Ticket triage and CRM flags need authorized connector actions rather than only report delivery. Treat the catalog as a starting point: adapt scope to your data and available tools, confirm the saved setup, and test it.

Read the result

Reports

Report partRequest and verify
Executive snapshotThe main change, defined counts and period comparisons, representative quotes, and emerging/trending/persistent classifications supported by evidence.
Deep diveSupporting feedback and hypotheses about change. Treat causal explanations and confidence as analysis to investigate, not proof.
CitationsOpen count and quote links; check records, filters, window, wording, and attribution. Verify unsupported claims and incomplete coverage.
PDF, where supportedRequest the format, then open the exported file and check content and source links. File access and linked-record access can differ.

Report structure depends on the prompt, Skills, and output settings. Every report is not guaranteed to contain every section or citation.

Connected actions

Inspect the result in the connected tool: a filed or tagged ticket, CRM update, or message. A configured report may also be delivered. Actions use the connector account’s access and are attributable to the automation creator. Check run and delivery results to confirm completion.

Manage your automations

The Automations list shows names and owners across the workspace; filter to your own to find what you created.

Control or stateMeaning
ActiveEnabled for configured triggers. Check run history and delivery for a particular run’s completion.
Paused / ResumeDisables future scheduled runs while retaining setup. Resume enables the configured cadence; do not assume missed runs replay.
EditWhere permitted, change prompt, reference files, cadence, delivery, connectors, or Skills for future runs. Use Run now where offered to check a saved change.
DeleteStops future runs and removes the automation from the active list. It does not promise permanent deletion of all historical run data.
Run historyInspect when each run started, its completion status, and output. Start here for missing delivery before rechecking the prompt or destination.

Duplicate an automation or stop a run

Choose Duplicate in the actions menu where offered. Review the separate copy’s prompt, schedule, delivery, files, Skills, connectors, and permissions, and connect its intended accounts. Creating a scheduled copy enables future runs; test with Run now where offered and pause while adjusting if needed.

For running work, open run history and choose Stop where available. The owner needs permission to run Agent work. Stopping a run does not pause future triggers; pause the automation too when needed. Pausing alone may let already accepted or running work finish.

Who can do what

  • Owner: controls the scheduled prompt, cadence, delivery, files, connectors, and Skills; can use Run now where available. Account access and Agent permissions still apply.
  • Applicable administrators: can pause an active automation or delete it. A non-owner administrator can pause it, but only its owner can re-enable it through normal workspace controls.
  • Separate accounts: admin access does not grant the owner’s connector accounts or saved-configuration editing. Use Duplicate where offered for your own setup.

Troubleshooting

First check Active status and run history after the scheduled time. Queued/running work may need time; failures need investigation. There is no fixed report-arrival deadline.

ProblemWhat to check or do
Run failedRead the recorded error; check connections, permissions, scope, and destination. Correct the identified issue and test again. Do not assume all failures are transient or caused by prompt size.
Report missingCheck the saved channel/person/email destination and delivery status. For a Slack channel, ensure the Enterpret Slack app has been added, then test.
Connector lost accessIn Connectors, check status and reconnect or re-authorize with your own account. OAuth can lapse; an external admin can revoke access; your access can narrow. Confirm target Slack membership or Linear project visibility, then test.
Truncated output or narrow coverageInspect the run for output/runtime limits, tool failures, or scope problems. Focus on one product area, roster, or window; request representative quotes rather than all. Split wider coverage across scoped automations. Use Run now where offered and verify coverage and delivery.
Unsupported or oversized fileUse a supported format and a file smaller than 10 MiB. Review which files were accepted before saving.
Too many filesRemove files or select fewer so the automation has at most 10; retry the selection.
Files could not uploadIf the message says nothing was saved, remove the failed files or try saving again. Check the saved automation’s file list; a successful upload alone does not prove all settings saved.
Files could not loadUse Retry where offered, or reload the page as instructed. Resolve file-loading errors before editing attachments.
The message Couldn’t load this automation’s files with the Retry control outlined in amber.
If files cannot load, select Retry where offered.

Still blocked? Contact support@enterpret.com if the cause is unclear, failures continue after narrowing scope and confirming access, or focused output still truncates. Include the automation name, expected versus actual behavior, run time, destination, and connector involved.

For another reporting setup, return to Scheduled Agent Automations.