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

# Provenance

> Population-level interaction dynamics at enterprise scale.

Provenance is the satellite view. It maps the geometry of an entire organisation across C, B, and I: where activation is happening, at what intensity, and how the population is shifting.

<Frame caption="Provenance. The field, at population scale.">
  <img src="https://mintcdn.com/blankstate/_qJUwEIy0R37Mjeb/images/site/provenance-og.png?fit=max&auto=format&n=_qJUwEIy0R37Mjeb&q=85&s=b0630ad1ad9c4d6c7fa414857a44f541" alt="Provenance. Population field." width="1200" height="627" data-path="images/site/provenance-og.png" />
</Frame>

## What Provenance shows

* **Population Signature map**: spectral projection of all active participants. Clustering and drift visible without exposing individual identities.
* **Protocol activation heatmap**: which Protocols are firing, on which surfaces, at what fidelity.
* **Cohort comparison**: compare defined cohorts (new vs. returning, human vs. agent, region A vs. region B) on any Protocol axis.
* **Drift alerts**: thresholds fire when a cluster centroid moves beyond the baseline.
* **Gravity**: what-if scenario planning on top of the live population. Perturb a Protocol threshold, a cohort mix, or a Blueprint version and see how the population geometry moves before you ship the change.

## Gravity

Gravity is the counterfactual layer inside Provenance. It runs projections on the current population Signature (no new Dives required) and shows how the map would deform under a specified change.

Typical uses:

* Rehearse a Blueprint upgrade before rollout.
* Model the impact of a policy change on a specific cohort.
* Stress-test a threshold against historical drift.

Gravity produces a differential view against the live baseline. Nothing is committed until you promote the scenario.

<Frame caption="Gravity. The field under a different framing.">
  <img src="https://mintcdn.com/blankstate/_qJUwEIy0R37Mjeb/images/site/provenance-field.png?fit=max&auto=format&n=_qJUwEIy0R37Mjeb&q=85&s=7e6e3cae573ee1fae72620f2f4970e15" alt="Gravity. A shifted population field." width="613" height="310" data-path="images/site/provenance-field.png" />
</Frame>

## Privacy model

Provenance operates on aggregate Signature data. Individual actant IDs are never displayed in the population view. Per-actant access is a separate grant and produces an audit log entry.

## Attribution-resistant aggregation

Population insights flow upward through Protocols that structurally cannot identify individuals. By construction, not by policy.

## Setup

1. Publish a Blueprint via Provenance (**Deploy > Provenance** in Atlas)
2. Attach it to a live Stage or API integration
3. Define a Signature baseline (first N interactions, or a pre-loaded reference)
4. Set drift alert thresholds per Protocol
5. Provenance updates continuously as interactions arrive

## Alert delivery

Atlas dashboard, webhook (HMAC-SHA256 signed), or email.