> For the complete documentation index, see [llms.txt](https://docs.meetalto.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.meetalto.ai/using-meet-alto/playbooks/launch-checklist.md).

# Launch your first workflow

A reusable checklist for publishing a useful workflow and checking its first real results.

Use this checklist for one account and one goal. It is a suggested operating practice, not a promise that every check is enforced automatically.

## Before launch

* [ ] The intended workspace and connected account are selected.
* [ ] The customer-facing promise is specific and accurate.
* [ ] Brand details and relevant knowledge are current and processed.
* [ ] Every link opens the intended destination.
* [ ] The trigger and audience match the use case.
* [ ] Collection prompts explain why the information is needed.
* [ ] Next actions point to the correct list, tool, or sequence.
* [ ] A teammate owns replies, reviews, and handoffs.
* [ ] Usage allowances and required connections are healthy.

## Test the whole path

Use a test person or account you control. Check a normal request, a non-matching request, missing information, and a customer who wants a person. For an agent, run [conversation tests](/using-meet-alto/agents/testing.md). For a campaign, inspect the actual message and downstream result.

Keep fictional information in a test context. Remove test sources or details that could otherwise influence customer replies.

## Enable deliberately

Check the saved configuration and readiness findings. Publish the campaign or sequence, or enable the agent's **Live** control, only when that specific workflow is ready. Start with **Training wheels** or a narrow trigger where appropriate.

## Check the first results

Review early conversations in the Inbox. Confirm the intended messages, contact details, and external outcomes. Compare the actual experience with your original promise.

If something is wrong, pause the affected automation or AI control and inspect what already happened. Pausing AI alone may leave scripted campaigns or sequences active.

Record one improvement for your [weekly review](/using-meet-alto/playbooks/weekly-review.md).