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

# Experiments

> Prepare a controlled agent comparison with a fixed scenario set, version references, and explicit release criteria.

An experiment compares a proposed configuration with a baseline using the same scenarios and scoring rules. Planning that comparison is useful before automated execution becomes available.

<Note>
  Threetone does not currently provide an experiment resource or an A/B experiment runner. Traffic allocation and automated tests remain subject to execution gates. This page describes a review plan, not a dashboard workflow.
</Note>

Start with an existing agent and a narrow hypothesis, such as whether a shorter instruction improves a specific handoff decision.

<Steps>
  <Step title="Record the baseline" id="experiments-baseline">
    Keep the agent, branch, and immutable version IDs in your review notes. Use [versions and drafts](/agents-platform/deploy/versioning) to distinguish a saved proposal from a published snapshot.
  </Step>

  <Step title="Define a controlled change" id="experiments-change">
    Change one instruction or model choice. Prepare the same [test definitions](/agents-platform/monitor/testing), synthetic inputs, and scoring rubric for both candidates.
  </Step>

  <Step title="Set the decision rule" id="experiments-decision">
    Record the required outcomes, unacceptable failures, sample size, and cost or latency measurements needed. Leave result cells unmeasured until comparable execution evidence exists.
  </Step>
</Steps>

Before runtime comparisons are released, review configuration diffs and scenario coverage. Neither successful publication nor a manual prompt review establishes a measured improvement.

<AccordionGroup>
  <Accordion title="Can deployment traffic act as an experiment runner?" id="experiments-faq-traffic">No. The current deployment execution gate applies, and there is no experiment assignment or statistical reporting workflow.</Accordion>
  <Accordion title="Can voice candidates be published for comparison?" id="experiments-faq-voice">Voice and custom-model publication are unavailable. Preserve eligible changes as drafts and record the blocked stage.</Accordion>
</AccordionGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.