How EOS LUX Works

EOS LUX — See the unseen.

What EOS LUX Is

Source-anchored intelligence for contested records.

EOS LUX is British Black Light's source-anchored evidence-intelligence layer for contested records, fragmented communications, omissions, provenance gaps and high-consequence decision files. It turns messy records into a structured first map: what was said, what is missing, where the source trail is reliable or incomplete, how power and pressure appear to move, and what a responsible reviewer should examine first.

EOS LUX is a necessary evidence, intelligence and control layer built for the AI age — a way for organisations to regain control of high-volume records, synthetic content risk and unstable information before decisions, allegations or public narratives harden.

The Three Intelligence Layers

The Three Intelligence Layers

01

The Smoking Gun

The visible source anchor: a quote, document, timestamp, contradiction, event or decision point that belongs in human review. A source anchor is not automatically a finding.

02

The Silencer

The missing explanation, absent authority, selective forwarding, unexplained gap, omitted rationale, or silent window where the record would normally be expected to speak. The Silencer is not proof of concealment. It is an absence signal.

03

Behavioural Intelligence

Record-derived indicators of role, pressure, influence, escalation, avoidance, alignment, tone shift and power movement, drawn only from the communications provided. Behavioural intelligence is not clinical diagnosis or motive determination.

How EOS LUX Works — Before / During / After

How EOS LUX Works — Before / During / After

Before

Before

Before: the record boundary is defined, scope is agreed, and data-handling terms — including local pseudonymisation of identifying details — are confirmed before any analysis begins.

During

During

During: the record is structured into chronology, source anchors, omissions and behavioural signals, with uncertainty registered rather than resolved by assumption.

After

After

After: a structured evidence pack is delivered for qualified human review, with follow-up questions attached rather than conclusions.

Pseudonymisation / rehydrationReal identifiers stay local; analysis sees surrogates; pack returns; local map restores for authorised review.REAL IDENTIFIERSstay localANALYSISsees surrogatesPACK RETURNShuman reviewLOCAL MAPrestores for authorised review

Pseudonymisation / rehydration

Pseudonymisation, not anonymisation.

Before: the record boundary is defined, scope is agreed, and data-handling terms — including local pseudonymisation of identifying details — are confirmed before any analysis begins.

Real identifiers stay local; analysis sees surrogates; pack returns; local map restores for authorised review.

The 13-step method

The 13-step method

  1. Start with a defined record boundary
  2. Preserve source structure
  3. Ingest and parse the record
  4. Build chronology
  5. Map actors, roles and decision points
  6. Identify source anchors
  7. Map omissions and silent windows
  8. Test contradictions and provenance gaps
  9. Add behavioural intelligence
  10. Register uncertainty
  11. Produce an evidence pack
  12. Export, review and decide
  13. Responsible-use boundary
Workflow13 steps
Before: the record boundary is defined, scope is agreed, and data-handling terms — including local pseudonymisation of identifying details — are confirmed before any analysis begins. During: the record is structured into chronology, source anchors, omissions and behavioural signals, with uncertainty registered rather than resolved by assumption. After: a structured evidence pack is delivered for qualified human review, with follow-up questions attached rather than conclusions.

Illustrative Evidence Card

Illustrative Evidence Card

Evidence-card anatomy

Fictional, anonymised

Source anchor

Internal message, 14:32, “confirmed with the client separately — no need to loop in finance on this one.”

Why this belongs in review

The message references a decision made outside the documented approval process.

Known limitation

The referenced “separate confirmation” does not appear elsewhere in the provided record set.

Human follow-up

Was the finance team's exclusion from this decision authorised, and by whom?

Public boundary

This is not a finding, allegation, legal opinion, or determination of wrongdoing.

This illustrative evidence card is fictional, anonymised and does not describe any real matter.

Human Review Requirement

Human Review Requirement

Every EOS LUX output is reviewed by a qualified human before any action, allegation, or decision relies on it. EOS LUX does not determine truth, guilt, intent, fraud, liability, authenticity, admissibility or legal outcome. All outputs require responsible human judgment before use.