SKIP EVIDENCE JOURNEY
EVIDENCE CONTINUITY INFRASTRUCTURE

DIGITAL EVIDENCE
SHOULD NEVER LOSE
ITS CONTEXT.

Digital Witness keeps media, source context, uncertainty, authority decisions, review state and action history connected. An authorised reviewer can reconstruct what the system recorded, preserved, asserted, reviewed and decided.

WORKING DEMONSTRATOR07 LINKED STAGES01 GOVERNED TRANSACTION
PANOPTICON / ONTOLOGYGOVERNED EVIDENCE FIELDDW / 001
PRESERVE MEDIA RETAIN CONTEXT RECORD ACTIONS
CURRENT ARTEFACTBRANCH-LOCAL FOUNDRY DEMONSTRATORSYNTHETIC DATABASELINE / 05 AUG 2026
01ONE ITEM, TRACED

A FILE BECOMES A RECONSTRUCTABLE CASE.

Move through the four states below. Each one adds information that a file store cannot retain on its own.

ONE FRAGMENT / ONE GOVERNED PATHSTEP 01 / 04
CAPTURED MATERIALVIDEO-014REPRESENTED / SYNTHETIC
01Case
02Signal
03Observation
04Captured Item
05Attestation
06Integrity Review
07Governance Audit Event
THE FRAGMENT

A FILE APPEARS.

A video, image or message can be visible without carrying a reliable account of its source, surrounding context or later decisions.

MATERIALVIDEO-014

STATEVISIBLE

CONTEXTDISCONNECTED

DECISION HISTORYUNKNOWN

THE PROBLEM

THE FILE IS ONLY ONE PART OF THE EVIDENCE.

A file can be copied, forwarded, reviewed, restricted or separated from the circumstances that gave it meaning.

Digital Witness models the surrounding context, uncertainty, authority and action history so they remain connected to the material.

READ THE PRODUCT MODEL
DIGITAL WITNESS / EVIDENCE CONTINUITYDW-PT-2026-0314SCROLL / 000%
DWWITNESS / GOVERN
THE DEEP WALKTHROUGH / SCROLL TO TRAVERSE

FOLLOW ONE FRAGMENT
THROUGH EVERY
GOVERNED DECISION.

Move from Case context to persisted audit provenance without silently converting represented state into external truth.

ARCHITECTUREOPERATIONAL MODELPRODUCT STORY

assembly

02WORKING PROOF

WATCH THE SYSTEM DISTINGUISH BLOCKED FROM DENIED.

The canonical synthetic transaction is deliberately small and inspectable: six attempts create five events because an invalid duplicate completion stops before state mutation.

CANONICAL TRANSACTIONTEST-RUN-20260805T1700Z
CASEDW-PT-2026-0314 / PROTECTED TRANSIT
ATTEMPT / 01
PERMITTED

A1 · Attest

The actor is resolved server-side, the permitted edit is applied and a linked Governance Audit Event is generated.

AUDIT EVENT 01
06ATTEMPTED STEPS
05TRANSACTION AUDIT EVENTS
11CASE-WIDE AUDIT RECORDS
16TOTAL VERIFIED ONTOLOGY EVENTS
LONG-TERM THESIS

PREVENTION REQUIRES INSTITUTIONS TO REMEMBER WHAT THEY KNEW—AND WHY THEY ACTED.

Digital material is easy to create, copy, alter and distribute. The evidence system must preserve what was known, what remained uncertain and which decisions followed.

CURRENT SYSTEMStates remain distinct.

Preserved, asserted, reviewed, blocked and denied are explicit in the demonstrator.

EVALUATION QUESTIONCan another reviewer reconstruct the path?

The proposed test measures hidden assumptions and unsupported inference.

LONG-TERM CASEGoverned history can be examined.

Prevention outcomes are a thesis for external evaluation, not a claim of the current artefact.

03CURRENT POSITION

WHAT EXISTS. WHAT COMES NEXT.

The current artefact is a working demonstrator, not a finished police platform. The next value step is external validation and production hardening.

01 / WHY NOW

THE PROBLEM

Digital evidence is fragmented across media, platforms, people, institutions and decision logs.

02 / CURRENT

WHAT EXISTS

A seven-stage ontology, four governed Actions and one verified end-to-end demonstrator transaction.

03 / COMMERCIAL

FIRST USE

A bounded evidence-continuity evaluation that can sit above or alongside existing forensic systems.

04 / MILESTONE

NEXT PROOF

External validation, stronger identity and authority controls, and production hardening.

04FIRST COMMERCIAL ENGAGEMENT

ONE WORKFLOW. ONE CONTROLLED TEST.

Map the workflow, expose hidden inference and test whether another reviewer can reconstruct the recorded path.

01

Discover

Map one existing investigation workflow.

02

Expose inference

Identify evidence states and authority decisions users currently infer.

03

Configure

Build a synthetic or sanitised comparative scenario.

04

Evaluate

Measure reconstruction errors, comprehension failures and task time.

05

Report

Deliver findings and define a restricted proof-of-concept proposal.

BOUNDED EVALUATION

CAN AN UNFAMILIAR REVIEWER RECONSTRUCT THE PATH?

Evaluate the evidence-continuity model against a synthetic or sanitised workflow—without overstating what the demonstrator proves.

INSPECT THE TRANSACTIONDISCUSS AN EVALUATION