Dr Joseph Ikhalia explaining the four layers of the Kenyan regulatory positionDVI 6630 M
The DICB method

One method, from regulation to proof.

The DICB Cyber Compliance Model™ takes any cyber regulation from open-textured legal text to obligations an organisation can run, evidence it can defend, and a record that outlives the people who built it.

Published on SSRN  ·  July 2026  ·  Open to challenge

Governance

Who owns the decision, under what authority, and what the record shows.

Regulation

What the law requires today, what is draft, what is proposed, and the difference.

Evidence

What an organisation can actually demonstrate, to a board, an auditor or a regulator.

The problem

Three problems. Most organisations have all of them.

Regulated organisations have frameworks, tooling and growing supervisory pressure, and still cannot answer simple questions about their own estate. DICB’s founding paper argues that three structural deficits explain a significant part of that gap, and that they are routinely assumed to be solved when they are not.

01

Translation

Which systems does this obligation actually reach, and who decided that?

There is no widely adopted method for taking open-textured regulatory language to a decision at system level. Interpretation happens, but the reasons are rarely kept. The quality of the translation depends on whoever did it.

02

Institutional continuity

Who knows the answer, and what happens when they leave?

The compliance state lives in experienced people and scattered records, not in a form the institution owns and can query. Audit readiness changes when a person resigns, although no control, system or law has changed.

03

Evidence quality

Is this attached, or is it proven?

Evidence is treated as present or missing rather than weighed against the claim it is meant to support. A policy proves that a policy was approved. It does not prove the control operated last week.

The three deficits compound into borrowed confidence: a reported status whose apparent precision exceeds the quality of the mappings and proof beneath it.
The Three Problems of Cyber Regulatory Compliance, Section 6
Requirements

Seven requirements, written so that they can be failed

The paper derives seven design requirements for an adequate discipline and states them as testable criteria, not as claims that DICB has already met them. Any method, including DICB’s own, can be judged against them.

Together they describe what a regulator, an auditor or a board is really asking for when they ask “can you prove that?”

A participant working through the DICB workbook in Nairobi
Working the requirements on a live case, The Social House Nairobi, 31 October 2026.
R1

Reproducible translation

Competent practitioners given the same obligation and context should produce substantially convergent records, with interpretations, assumptions and divergences explicit, reasoned and reviewable.

R2

Obligation-scope granularity

State, ownership, applicability, exceptions and evidence attach to discrete obligation-scope units, not to whole standards. Unknown is a reportable state, never silently converted to compliant.

R3

Institutional residence

The complete demonstrable compliance state is held in an institution-owned, current and queryable form that survives staff transitions and answers supervisory questions directly.

R4

Claim-relative fitness

Evidence is assessed against the proposition it is meant to support or refute: relevance, directness, scope, sufficiency, independence or corroboration, and consistency.

R5

Provenance and integrity

Artefacts meet defined tests for traceability, reviewability, attribution, contemporaneity and enduring integrity before they are admitted to support a conclusion.

R6

Temporal honesty

Reported posture reflects time-based decline, event-triggered invalidation and coverage erosion. Stale, superseded or scope-incomplete evidence cannot keep full credit invisibly.

R7

Examinable and interoperable practice

The method is teachable and practitioners are examinable against it. Records can be kept by hand and represented in machine-readable formats compatible with established standards, including the Open Security Controls Assessment Language (OSCAL).

The model

One model, four layers

DICB’s answer to the three problems is a codification: standards for translation and evidence, a platform for continuity, a method that binds them into daily practice, and a credential that examines people against it. Each layer rests on the one beneath it.

01  FoundationThe problem analysis, published separately from the answer so that it can be tested on its own.

The Three Problems of Cyber Regulatory Compliance

Translation, Institutional Continuity, and Evidence Quality: a position paper towards an evidence-led discipline. Dr Joseph Ikhalia, SSRN, July 2026. It defines the deficits, examines prior art and two public supervisory cases, derives the seven requirements, and names the research that could prove it wrong.

02  Standards and componentsThe parts of the DICB Cyber Compliance Model, each specified as a versioned standard.

Translation, continuity and evidence, specified

Translation · R1, R2

DICB Cyber Regulatory Translation Standard™

The same questions, the same order of reasoning and the same record structure for every obligation. Its output is the Canonical Obligation Record.

Translation · the unit of record

DICB Canonical Obligation Record

One interpreted obligation joined to its source, reasoning, scope, controls, owner, state, exceptions, dependencies and evidence. The atom everything else is built from.

Evidence · R4, R5, R6

DICB TRACE Evidence Standard™

Three linked instruments: the TRACE Evidence Quality Test as the admissibility gate, the Evidence Matrix for claim-relative judgement, and the Evidence Decay Model for temporal honesty.

Evidence · claim fit

DICB Evidence Matrix

Evidence judged against four kinds of claim: control design, implementation, operating effectiveness, and governance and decision. No artefact is strong for every claim.

Evidence · time

DICB Evidence Decay Model

Three ways evidence loses current value: time-based decline, event-triggered invalidation and coverage erosion. No universal half-life.

Continuity · R3

DICB Obligarium™

The Cyber Compliance Digital Twin: a continuously current model of obligations, systems, controls, owners, dependencies, exceptions, decisions and evidence, built to answer a regulator’s question directly.

Measurement

DICB Compliance Benchmark™

Governed, anonymised assessment results accumulated for comparative measurement and future calibration. No benchmark claim is made before sample size and limitations are defined.

03  Professional applicationHow the standards are used, day to day, by a practitioner.

DICB Cyber Compliance Method™

One working sequence that brings the Translation Standard, the TRACE Evidence Standard and institutional continuity into practice: five steps carried from obligation to decision. It is what the DICB Certified Cyber Regulatory Practitioner course teaches and what the examination tests.

04  Professional designationThe examined credential for people who can apply the method.

DICB Certified Cyber Regulatory Practitioner (DICB-CCRP)

Complete the course, pass the examination, become certified. Launching December 2026.

About the programme

Translation

The same questions, in the same order, every time

The Translation Standard does not pretend that legal or engineering judgement is unnecessary. Regulatory language is open-textured by design, and interpretation cannot be removed from the process. What the Standard removes is avoidable inconsistency: it requires the same questions, the same order of reasoning and the same record structure for every obligation, and it keeps the reasons.

The output is the Canonical Obligation Record. One duty per record. Applicability and scope decided with a rationale. An accountable owner by role and by name. A state that is declared only once the evidence position is known, with unknown reportable rather than quietly rounded up to compliant. An exception that explains non-compliance and never converts it.

Records are as coarse as the truth allows and generated rather than rewritten, so an estate of five thousand systems does not need five thousand rules.

Canonical Obligation Record Fourteen fields
  1. 01Authoritative source
  2. 02Requirement
  3. 03Reasoned interpretation
  4. 04Assumptions and uncertainty
  5. 05Applicability logic
  6. 06Affected scope
  7. 07Control mapping
  8. 08Accountable owner
  9. 09Current state
  10. 10Exceptions
  11. 11Dependencies
  12. 12Evidence claims
  13. 13Evidence references
  14. 14Review history
Evidence, part one

Before evidence counts, it has to be admitted

The TRACE Evidence Quality Test is the gate. An artefact enters the evidential record only if it is all five of these.

T

Traceable

The origin, collection path, transformations and relationship to the claim can be followed.

R

Reviewable

A competent reviewer can access, understand and challenge the artefact and the method that produced it.

A

Attributable

Linked to an identifiable, authoritative person, system, process or source.

C

Contemporaneous

Created or captured close enough to the period, event or state it claims to evidence.

E

Enduring

Preserved with integrity, retrievable, and retained for as long as later review requires.

Passing the gate shows that an artefact is well preserved. It does not show that the artefact proves anything. That is a separate test, and keeping the two apart is the point.

Evidence, part two

Judged against the claim, not counted as attachments

The DICB Evidence Matrix asks what the evidence is for. A policy is weak evidence that a control operated yesterday, but direct evidence that a policy was approved. Telemetry may prove technical activity and still be incapable of proving that a risk was knowingly accepted by the person with authority to accept it. Four kinds of claim, four different answers to what counts as strong.

Claim family one

Control design

Was the control appropriately designed for the interpreted obligation and the risk?
Generally stronger

An approved control specification, reviewed architecture, risk analysis, a design decision record.

Commonly insufficient on its own

Policy text with no control design. An unsupported management assertion.

Claim family two

Implementation

Was the designed control deployed across the required scope?
Generally stronger

Direct configuration or inventory data, deployment records, validated system mappings, inspection.

Commonly insufficient on its own

Generic screenshots. A project plan. A statement that deployment occurred.

Claim family three

Operating effectiveness

Did the control operate consistently during the relevant period?
Generally stronger

Continuous or repeated telemetry, sufficiently scoped testing, independent test results, exception data.

Commonly insufficient on its own

A single point-in-time artefact for a period claim. A policy. An uncorroborated attestation.

Claim family four

Governance and decision

Was ownership, approval, risk acceptance or escalation validly exercised?
Generally stronger

An authorised workflow record, a signed decision, contemporaneous minutes, a responsibility map, a reasoned exception record.

Commonly insufficient on its own

Technical telemetry. An informal email. An undocumented verbal approval.

A participant reviewing evidence in the DICB workbook
Evidence reviewed against the claim it has to carry. Nairobi, 31 October 2026.
Evidence, part three

Three ways evidence goes stale

The DICB Evidence Decay Model refuses a single universal half-life. How long evidence stays good depends on the claim, the artefact, how fast the estate changes and what has happened since it was captured.

Time-based decline

Credit falls progressively unless renewed or corroborated. A penetration test becomes less representative as releases and threat conditions accumulate.

Event-triggered invalidation

A material release, a control redesign, a certificate expiry or a supplier change breaks the link between the artefact and the live state. The artefact stays historically authentic. It stops proving the present.

Coverage erosion

Fresh telemetry covers one region, platform or supplier while the claim covers all of them. Coverage credit falls, and the uncovered population becomes explicit instead of invisible.

Practice

Five steps, carried from obligation to decision

The DICB Cyber Compliance Method brings the standards into one working sequence. The compliance state is declared at the end, after the evidence, and not before.

STEP 01

Assess the obligation

Isolate the source, extract one duty per record, interpret it with reasons, and decide applicability and scope.

STEP 02

Link it to cyber risk

Connect the obligation to the risks it exists to reduce and the systems that carry them, and set priority.

STEP 03

Implement controls

Map controls, assign an accountable owner by role and by name, and govern exceptions explicitly.

STEP 04

Gather evidence

Admit artefacts through TRACE, judge them against the claim, and record what is missing as missing.

STEP 05

Navigate governance

Declare the state, report it without changing its colour, and secure a sign-off that can be defended later.

Two participants working through the DICB workbook together in Nairobi
The sequence worked on a live artificial intelligence use case, The Social House Nairobi, 31 October 2026.
Taught in the room

Nairobi, 31 October 2026

Thirty professionals from banking, law, technology, insurance and government carried one use case through the sequence in a morning: assessed the risk, built a register entry, screened it against Kenyan law and wrote a one-page briefing their leadership could act on. Four working papers each, and a method to take home.

Read about the workshop

Stated limitations

What the method does not yet claim

A discipline that asks organisations to be honest about their evidence has to be honest about itself. The founding paper states its limits in its own text. The components have substantial prior art in financial audit, records management, digital forensics, organisational memory and continuous auditing; the claimed contribution is their unification at obligation level into one teachable discipline, not the invention of each part.

The empirical base is limited, and the continuity proposition in particular needs direct testing. The Evidence Matrix and the Decay Model are architectural, not calibrated, and fixed scores will not be published merely because software can calculate them.

The author holds a commercial interest in the codification, which is why the problem analysis is published separately from the answer and the discipline remains open to alternative codifications that meet the same requirements.

Seven research streams are specified to disprove, refine or constrain the central claims, with both positive and negative results to be published. The decisive test is not the elegance of the vocabulary. It is whether different practitioners translate obligations more consistently, and whether organisations can answer a regulator in minutes without heroics.

Read the paper on SSRN

A control is not what the policy says should happen. It is what you can prove happened.
Dr Joseph Ikhalia
Two ways in

Learn it, or bring it into your organisation

For practitioners

DICB Certified Cyber Regulatory Practitioner

A self-paced course of about five hours, with a workbook, built on a real regulation of your choice. Complete the course, pass the examination, become certified. Launching December 2026.

See the programmes

For organisations

Compliance services built on the method

Cyber regulatory compliance services, artificial intelligence use case assessments, executive briefings and private cohorts, all run on the same standards described on this page. Independence is part of the method: DICB does not certify governance it built.

Bring DICB into your organisation