# State-Action-State Trajectory Data

> State-action-state trajectory data from MARFI connects before state, human action, and after state for AI agent training workflows.

Canonical URL: https://insight.marfi.app/state-action-trajectory-data

State-action-state data describes what the environment looked like, what a person did, and what changed after that action. MARFI assembles those examples from consented work sessions so AI teams can train and evaluate agents against grounded workflow demonstrations.

## Audience

AI engineers and data teams preparing demonstrations for task agents.

## Search Intent

Understand state-action-state trajectory datasets for AI agent training.

## A trajectory is a supervised slice of work.

Each step ties an observed action to nearby state. For visual tiers, before and after keyframes help models reason about what changed. For telemetry-only tiers, app, title, URL, timing, and system context still provide a grounded sequence.

## Real demonstrations reduce brittle automation.

Agents trained only on simplified examples often fail in messy production workflows. MARFI helps teams collect proprietary demonstrations that include the real interfaces, interruptions, and workarounds their users face every day.

- Provider-neutral JSONL
- Session boundaries
- Action labels
- Optional visual keyframes

## Trajectory capture follows plan and consent gates.

The Insight plan does not collect screenshots. Trajectory unlocks visual replay and training exports, but each capture tier still depends on explicit user consent and workspace controls.

## FAQ

### What is a state-action-state example?

It is a record containing the state before an action, the action itself, and the state after the action. MARFI exports this as trajectory data for AI training workflows.

### Do all trajectories include screenshots?

No. Screenshots only exist when the tenant is on Trajectory and the user opted into screenshot capture.

## Related Pages

- [AI Training Data From Real Workflows](https://insight.marfi.app/ai-training-data): MARFI Trajectory converts consented workplace sessions into state-action-state training data for AI agents and workflow models.
- [What Is State-Action-State Trajectory Data?](https://insight.marfi.app/answers/what-is-state-action-trajectory-data): State-action-state trajectory data records what the environment looked like before an action, what the user did, and what changed after.
- [AI Teams](https://insight.marfi.app/solutions/ai-teams): MARFI Trajectory helps AI teams collect consented workflow demonstrations and export state-action-state JSONL for agent training.
- [Pricing](https://insight.marfi.app/pricing): MARFI Insight AI pricing is per sensor: Insight at $15 per sensor per month, Trajectory at $99, and Enclave as custom self-hosted annual licensing.