> ## Documentation Index
> Fetch the complete documentation index at: https://developers.autoplay.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Step 2: Define the onboarding objective
> Define what each user segment should learn, what successful and unsuccessful progress look like, and how you will measure the result.
Live activity tells Autoplay what a user is doing. The onboarding objective explains what that user should learn and which product outcome they should reach.
Your team defines the workflows that lead users to value. Autoplay then recognizes progress against those workflows automatically in the background. You describe the objective and the evidence; Autoplay keeps comparing live behavior with it.
## Define what a successful onboarding outcome looks like to the agent
Your team does this by defining the workflows each user should complete. Each workflow represents something the user should learn or accomplish, while its progress and completion signals tell the agent what success looks like.
Imagine your product helps teams publish reports from connected data. The objective is not for a new customer to visit the integrations page. It is for them to learn enough of the product to reach a useful outcome:
1. Connect a data source and verify that it works.
2. Create the first report using that data.
3. Share the report with a teammate.
4. Later, learn how to automate the report each week.
Each meaningful outcome becomes a workflow Autoplay can understand. Together, these workflows describe how a customer reaches initial value and continues adopting the product.
## Define the objective for each segment
Different users may need different onboarding. An administrator may need to configure the workspace, while an analyst needs to connect data and publish a first report.
Segments let your team decide:
* Which users should learn a workflow
* Which workflows are required for a particular role, plan, lifecycle stage, or use case
* Which success metric matters for that audience
* Which onboarding experience or message they should receive
Autoplay currently supports role-aware workflow plans so guidance stays relevant to the user's responsibilities. Your analytics platform can provide broader cohorts and segments for evaluating the experience.
## Define how to detect workflow progress
For every workflow, tell Autoplay what successful and unsuccessful progress look like.
| Workflow evidence | Example |
| --------------------- | ------------------------------------------------------------- |
| Started | The user opens integration settings |
| Progressing | The user selects and configures a provider |
| Successful completion | The connection is created and verified |
| Unsuccessful attempt | The user submits the setup but no verified connection appears |
Autoplay uses these signals to recognize whether the user is progressing, completed the workflow, appears stalled, or returned to unfinished work. This recognition happens automatically through the Autoplay service and MCP context. Your team does not need to inspect every session manually.
Once Autoplay sees the success signal, it marks the workflow complete and stops offering help for a task the user already knows how to do.
## Compare onboarding experiences
Segments also make onboarding measurable. Your team can use an experimentation tool such as Amplitude Experiment to compare different onboarding approaches for similar audiences.
For example, one segment can receive the current onboarding experience while another receives proactive help from Autoplay. You can then compare workflow completion, time to first value, feature adoption, or churn between the groups.
Autoplay supplies the workflow progress and guidance layer. Your analytics and experimentation tools remain the place where your team defines the test and evaluates its business impact.
Configure when Autoplay should offer help and how the user stays in control.