Reconstruct AI-assisted decisions into defensible timelines.
The record shows what happened across AI, human, and system actors. AegisTrace separates what is KNOWN, what is INFERRED, and what remains UNKNOWN without overstating the evidence.
Directly observed logs, hashes, timestamps, and actor records.
Transitions derived from adjacent evidence with the derivation kept visible.
Missing rationale, absent records, and uncaptured context stay explicit.
Reconstruction Timeline
Review how available records become a sequenced timeline of AI, human, and system activity under uncertainty.
AI output observed: transfer request for $50,000 to an external account.
System record attached a policy threshold reference for transfers above $10,000.
Manual approval record observed with justification label "Emergency Liquidity Event".
Sequence reconstructed from available records. Full operator rationale was not captured.
Reconstructed from surrounding evidence, not directly observed.
This timeline shows how scattered evidence can be reconstructed into a reviewable sequence. It does not determine correctness, compliance, or liability.
What We Do
Three steps to turn scattered evidence into defensible timelines.
Ingest what you have
Upload structured or messy logs—AI traces, tickets, approvals, screenshots, policy PDFs. No SDK required to start.
Reconstruct decision paths
We reconstruct a timeline of AI outputs, human interventions, and system events. Every entry stays marked Known, Inferred, or Unknown.
Produce reviewable records
Organize timelines, neutral reconstruction notes, and linked evidence hashes for review. Traceability, not automated judgment.
Built for Risk, Legal, and Compliance Teams
When an AI-assisted decision is questioned, you need to reconstruct what happened using incomplete logs. AegisTrace is purpose-built for that challenge.
Chief Risk Officers
Build defensible evidence chains for AI-assisted decisions when incidents are questioned.
General Counsel & Legal
Reconstruct decision timelines with timestamped evidence for regulatory inquiries and litigation support.
Compliance Officers
Prepare review-ready documentation showing what happened, what was known, and where gaps exist.
Audit & Security Teams
Investigate AI decision incidents with forensic-grade reconstruction and explicit uncertainty markers.
What We Are — What We Are Not
Clear boundaries matter. We reconstruct records for review, not outcomes for judgment.
What We Are
- Evidence-first reconstruction from logs, records, and system artifacts
- Forensic decision path reconstruction
- Traceability support for legal, risk, and audit review
- Transparent about gaps (Known/Inferred/Unknown)
- Starting point: upload what you already have
What We Are Not
- Compliance determination or violation detection
- Risk scoring or pass/fail certification
- Prevention or real-time blocking system
- Guarantee of safety or regulatory approval
- Requiring perfect logs or full AI instrumentation
Explicit Uncertainty
Every reconstruction keeps the evidence state visible. Missing records stay marked as missing instead of being smoothed away.
Directly observed in the record. Example: logged action, actor ID, recorded timestamp.
Reconstructed from adjacent evidence. Example: a transition derived from multiple partial records.
Missing or uncaptured evidence. Example: human rationale absent from logs.
Reviewers can see the limits of the record immediately: what was observed, what was reconstructed, and what is still absent.
Start With What You Have
We don’t require perfect logging. We work with partial data and mark gaps as unknown.
No SDK or agent required
Upload what you have. We reconstruct the decision path and explicitly mark missing evidence.
Forensic Event Feed
Every AI and human decision event captured with cryptographic proof. Immutable audit trail for regulatory investigation and review.
Important: AegisTrace provides evidence and analysis support. It does not determine compliance, violations, or liability. Demonstration visuals and data.
Explore linked scenarios and incident reconstructions
Browse cross-domain scenarios that demonstrate decision timelines, evidentiary artifacts, and explicitly marked uncertainty—without breaking the narrative flow.
Frequently Asked Questions
Common questions about how AegisTrace works and what it delivers.
Decision Event Schema
Illustrative data structure for evidentiary capture. Demonstrates how events are cryptographically recorded for review.
Important: AegisTrace provides evidence and analysis support. It does not determine compliance, violations, or liability. Demonstration visuals and data.
{
"event_id": "evt_2026_001_au_nsw",
"timestamp": "2026-01-13T21:15:42.123Z",
"agent_id": "agent_legal_counsel_01",
"jurisdiction": "AU-NSW",
"regulatory_scope": [
"APRA CPS 230",
"Privacy Act 1988"
],
"event_type": "AI_EVENT_OBSERVED",
"decision_stage": "RECONSTRUCTION_COMPLETE",
"chain_hash": "sha256:a7f5c8d9e2b4f1a8c6d3e9b2f4a1c8d6e3b9f2a4c1d8e6b3f9a2c4d1e8b6f3a9",
"previous_hash": "sha256:b8f6d9e1a3c5f2b9d7e4f1a9c7d4e2b5f3a1c9d7e4b2f5a3c1d9e7b4f2a5c3d1",
"data_provenance": "CUSTOMER",
"hashed_reasoning": "sha256:c9g7e1b4d6f3a2c9e8b5f1d7a4c2e9b6f3d1a8c7e5b2f9d4a1c8e6b3f2d9a5c7",
"observed_action": "Document classification performed",
"human_action_present": false,
"policy_reference_ids": [
"APRA-CPS-230-3.1",
"PRIVACY-1988-5.2"
],
"policy_context_available": true
}Cryptographic fingerprint linking events in sequence
Reference to prior event creating unbroken chain
Category of observed event (AI action, human intervention, etc.)
Current reconstruction phase of the decision path
SHA-256 hash preserving AI deliberation for later review
Human-readable description of action recorded
Tracks origin of data (customer, third-party, hybrid)
Links to applicable policy contexts for investigative review