> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pavoai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Engagement models

> Two ways to work with Pavo: run the improvement loop yourself, or have a Forward-Deployed Scientist run it with you.

Pavo finds the highest-value change to your production ML/DS surfaces, proves it against a real benchmark, and ships it. You can run that loop yourself, or have us run it with you.

The two are **distinct and complementary**: same platform underneath, same modules, same data isolation. The only difference is **who drives.**

## Self-serve

*Designed for teams with data-science bandwidth who want to own and run the improvement loop themselves.*

Your team uses the Pavo platform to build system knowledge, generate and test approaches in parallel, and ship what moves the metric.

## Forward-Deployed Scientist

*Designed for teams who want a metric win proven before they staff it.*

Everything in self-serve, **plus** a Pavo scientist who embeds with your team and runs the loop with you until a metric moves, then hands it back so your team runs it.

<CardGroup cols={2}>
  <Card title="Self-serve" icon="user-gear" href="/engagement/self-serve">
    Your team runs the loop on the Pavo platform.
  </Card>

  <Card title="Forward-Deployed Scientist" icon="users-gear" href="/engagement/fds">
    A Pavo scientist runs the loop with you, then hands it back.
  </Card>
</CardGroup>

## Shared foundation

<Info>
  Both modes run on the same platform and the same two modules, [System Knowledge](/system-knowledge/overview) and [Agent Studio](/agent-studio/overview). Both keep your code, data, and learnings inside your own environment: a per-client Google service account, SOC-II/ISO-compliant, deployed in your cloud or ours. Nothing you choose here changes your security posture.
</Info>

## Side by side

|                     | Self-serve               | Forward-Deployed Scientist                 |
| ------------------- | ------------------------ | ------------------------------------------ |
| Who drives the loop | Your team                | A Pavo scientist → your team               |
| Best when           | You have DS/ML bandwidth | You want a win proven first                |
| Time to first win   | Your cadence             | \~8 weeks (single metric)                  |
| You end with        | A loop your team runs    | A moved metric **and** a team that runs it |
| Where data lives    | Your environment         | Your environment                           |

Not sure which fits? [Start a conversation](https://pavoai.com), we'll help you pick.