Creative Lineage Demo

Creative Lineage: From AI Exploration to Verifiable Execution

See how PFC turns AI governance into decision traceability, audit trail evidence, and execution verification without granting authority to exploration.

Request API Key

Use this walkthrough to see how verifiable AI decisions become audit-ready records.

From Proposal to Proof

Proposal → Governance → Execution → Proof

Creative Lineage: From AI Exploration to Verifiable Execution

Start with the execution boundary demo to see how PFC governs actions before execution.

The execution boundary governs the action. Creative lineage verifies the recorded path from exploration to the bound governance decision.

PFC can explore multiple possible actions without granting any of them authority. It evaluates each candidate under live policy, authority, and state conditions, and only an explicitly bound admissible action receives an execution receipt. The system also cryptographically links the governed decision to the signed exploration artifact from which it was derived.

Open developer guide See execution-boundary demo How it works

This page shows how AI decision governance becomes visible as an audit trail: PFC records decision traceability from creative exploration through governed execution and final verification. Teams use this kind of AI governance flow to improve auditability, compliance posture, trust, and execution verification under real policy conditions. Want the implementation path? See the API and learn how PFC governs decisions.

Flow Visualization

The chain starts with a goal, moves through candidate evaluation, reaches explicit binding, then ends with an authoritative execution receipt plus independently verifiable exploration lineage.

1

Exploration

Session `sess_req_creative_demo_001` begins with the goal to resolve account fraud case safely.

2

Candidate Evaluation

Freeze account, limit transactions, notify user only, and monitor silently are labeled and evaluated.

3

Explicit Binding

Only candidate `c1` is both admissible and explicitly bound. The others remain blocked or review-only.

4

Execution Receipt

The execution receipt is the only authoritative artifact. Exploration remains non-binding.

5

Creative Provenance

The signed receipt carries creative provenance and points to the exact signed exploration artifact hash.

6

Exploration Verification

The exploration artifact can be verified independently without becoming executable authority.

Demo Story

Goal Input

Business-readable scenario: fraud response for a high-risk account.

{
  "request_id": "req_creative_demo_001",
  "operation_mode": "explore_then_evaluate",
  "goal": "Resolve account fraud case safely",
  "context": {
    "account_id": "123",
    "risk_flags": ["velocity_spike", "device_change"]
  },
  "generation_config": {"max_candidates": 4},
  "evaluation_config": {"bind_candidate_id": "c1"}
}

Exploration Session

Operation mode `explore_then_evaluate` returns metrics, a selected candidate, and a signed exploration artifact.

{
  "session_id": "sess_req_creative_demo_001",
  "operation_mode": "explore_then_evaluate",
  "goal": "Resolve account fraud case safely",
  "metrics": {
    "total_candidates": 4,
    "execution_candidates": 2,
    "admissible_candidates": 1,
    "blocked_candidates": 3,
    "bound_candidates": 1
  },
  "selected_candidate": "c1",
  "exploration_artifact": {
    "exploration_id": "exp_sess_req_creative_demo_001",
    "artifact_type": "exploration",
    "binding_status": "non_binding",
    "artifact_hash": "exp_hash_4f17d7",
    "signature": "sig_exploration_9u1"
  }
}

Candidate Set

The candidate cards make the distinction between non-binding exploration and authoritative execution obvious.

c1 · execution_candidate

Temporarily freeze account

governance_result: allow

admissible: true

requires_human_review: false

Explicitly bound and issued an execution receipt.

c2 · execution_candidate

Limit transactions

governance_result: block

admissible: false

requires_human_review: true

c3 · plausible

Notify user only

governance_result: block

admissible: false

requires_human_review: true

c4 · speculative

Monitor silently

governance_result: block

admissible: false

requires_human_review: true

Verifiable AI Decision Artifacts

Each block below is evidence of how an AI decision was governed, linked, and independently verified. Together they create an audit-ready record for execution verification, decision traceability, and AI compliance evidence. If you want the control point first, return to the execution boundary demo.

Artifacts and Verification

Proof of Lineage

One hash links exploration to execution

This is the lineage-verification point. The receipt carries a direct cryptographic pointer to the signed exploration artifact associated with the bound governance decision.

Receipt creative_provenance.exploration_artifact_hash = exp_hash_4f17d7
Exploration Artifact artifact_hash = exp_hash_4f17d7

These hashes match. The recorded governed decision is cryptographically linked to this signed exploration artifact.

Execution Receipt

Only an explicitly bound admissible action receives an execution receipt.

{
  "decision_id": "dec_4b91",
  "receipt_hash": "sha256:ab71e6c2",
  "signature": "ed25519:sig_receipt_9k2",
  "creative_provenance": {
    "lineage_type": "explored_then_bound",
    "exploration_id": "exp_sess_req_creative_demo_001",
    "session_id": "sess_req_creative_demo_001",
    "candidate_id": "c1",
    "origin_operation_mode": "explore_then_evaluate",
    "binding_status_at_origin": "non_binding",
    "exploration_artifact_hash": "exp_hash_4f17d7"
  }
}

Exploration Artifact

This is a different artifact class: non-binding exploration, not execution authority.

{
  "exploration_id": "exp_sess_req_creative_demo_001",
  "artifact_type": "exploration",
  "binding_status": "non_binding",
  "artifact_hash": "exp_hash_4f17d7",
  "signature": "sig_exploration_9u1"
}

Verification Result

Exploration artifact verification stays separate from receipt verification.

{
  "verified": true,
  "artifact_type": "exploration",
  "binding_status": "non_binding",
  "verification_status": "valid"
}

Why this matters

Creative lineage gives teams more than a model output. It provides verifiable evidence of the recorded exploration, the governed decision, and the cryptographic linkage between them. This makes high-stakes decisions explainable under real policy, authority, and execution conditions.

Audit

Prove what was considered and what was executed.

Compliance

Show how an allowed action was derived under policy.

Incident review

Reconstruct the recorded lineage that led to a governed decision.

Operator trust

Clearly separate exploration from authority at every step.

Output filtering vs governed creativity with proof

Output filtering

  • tries to control what the model says
  • acts before or around generation
  • does not govern execution
  • cannot prove how a possibility became an action

Governed creativity with proof

  • allows exploration without granting authority
  • evaluates candidates under real execution conditions
  • binds only explicitly allowed actions
  • verifies cryptographic lineage from exploration artifact to execution receipt

Most systems try to filter outputs. PFC preserves verifiable lineage from explored possibilities to the recorded governed decision.

Next Step

Turn AI Decisions Into Verifiable Records

Within a PFC-governed execution path, decisions can produce signed artifacts that downstream systems can independently verify. Explore the runtime path in the execution boundary demo, or see the API for the implementation details behind verifiable AI.

Run the Demo Explore the API Browse resources