> ## 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 4: Build the agent self-improvement loop

> Use workflow outcomes to improve onboarding and extend guidance into continual product adoption.

Autoplay does not stop helping when initial onboarding is complete. The workflows your team defines can also represent recurring tasks, advanced capabilities, newly released features, and role-specific product journeys.

This turns onboarding into continual adoption. Autoplay can help users keep learning the platform based on what they have already completed and what is relevant next.

## Continue from initial value to deeper adoption

For the reporting product, the learning journey may continue after the first report:

| Stage                  | Workflow the user can learn               |
| ---------------------- | ----------------------------------------- |
| Initial onboarding     | Connect a data source                     |
| First value            | Publish and share the first report        |
| Continued adoption     | Schedule a recurring report               |
| Advanced adoption      | Configure alerts and automations          |
| Role-specific learning | Manage teammates or workspace permissions |

Completed workflows remain part of the user's context. Autoplay can avoid repeating what they already know and focus support on unfinished or newly relevant outcomes.

## Learn from observed outcomes

The same data used to guide an individual user also helps the product team improve the overall onboarding experience:

* Which workflows users complete or abandon
* Where different segments tend to stall
* Which proactive offers users accept or dismiss
* Whether guided users reach value faster than a control group
* Which advanced workflows lead to deeper adoption

Teams can use these outcomes to refine workflow definitions, segment rules, exploration thresholds, cooldown periods, and trigger copy.

<Frame caption="Observed outcomes feed the next iteration of onboarding and continual adoption.">
  <img src="https://mintcdn.com/autoplayai/nWg8NTuzdnpgOvaA/images/proactive-onboarding/self-improvement-loop.svg?fit=max&auto=format&n=nWg8NTuzdnpgOvaA&q=85&s=76cf1fc9f5aeffaeabe73b4f96de261f" alt="Workflow outcomes help teams refine objectives, segments, proactive settings, and future guidance." width="900" height="360" data-path="images/proactive-onboarding/self-improvement-loop.svg" />
</Frame>

## Where the agent self-improvement loop is heading

Today, your team reviews these signals and updates the configuration with Autoplay. The longer-term direction is a more self-improving agent that can recommend changes based on observed outcomes, while keeping product teams in control of what is deployed.

That future loop can help answer questions such as:

* Should this segment receive a different onboarding objective?
* Is the exploration gate offering help too early or too late?
* Which nudge performs better for stalled users?
* Which workflow should become the next adoption goal?

This is the progression from a fixed onboarding sequence to an agent that keeps learning how users adopt the product and helps the team improve the experience over time.

<CardGroup cols={2}>
  <Card title="See the technical setup" icon="code" href="/quickstart">
    Follow the engineering quickstart to connect the required parts.
  </Card>

  <Card title="Talk to Autoplay" icon="comments" href="https://www.autoplay.ai">
    Discuss which user workflows and segments you want Autoplay to support.
  </Card>
</CardGroup>