CDI-BoK Constitution¶
Status of this Constitution
Javier Forero ratified this Constitution on July 21, 2026. It is normative within the CDI-BoK from v0.4.0. Its institutional authority does not declare CDI a consolidated academic discipline, the CDI-BoK a recognized standard or PULSE a generally validated methodology.
Article 1 — Identity¶
Conversational Decision Intelligence (CDI) is the proposed interdisciplinary domain of practice that studies and designs how people and AI systems collaborate through conversations to turn trustworthy evidence and context into explicit decisions, responsible action and measurable learning.
The CDI-BoK is the public, governed and versioned knowledge system that organizes CDI definitions, principles, models, patterns, evidence, practices and research questions.
PULSE is the ecosystem's framework of practice. Its constitutional meaning remains under its canonical sources; the CDI-BoK may apply, explain and test it, but may not redefine it.
Article 2 — Purpose¶
The CDI-BoK exists to reduce five recurring distances:
- between a business question and the evidence it needs;
- between data and operational context;
- between a conversational answer and an authorized decision;
- between a decision and the action that must change;
- between the observed result and learning in the next cycle.
Its primary design unit is the decision. Conversation is an interface and collaborative process; it is not by itself evidence, criterion, authority or a guarantee of truth.
Article 3 — Scope of authority¶
The CDI-BoK may establish official project language, editorial requirements, candidate practice models and rules for its artifacts. Institutional authority is not external scientific validity.
Authority is interpreted in this order:
- CDI-GOV-001 and ratified governance decisions;
- the PULSE Constitution and canonical sources for PULSE-owned concepts;
- approved CDI-BoK foundational assets;
- derived standards, patterns and guides;
- empirical evidence, cases and demonstrations with declared scope;
- hypotheses, proposals and exploratory material.
A lower-level source may challenge a claim with evidence, but cannot silently change its institutional definition.
Article 4 — Constitutional principles¶
| Principle | Minimum requirement |
|---|---|
| Decision First | Name decision, owner, alternatives, criteria, constraints and subsequent action |
| Evidence and Context | Separate data, calculation, inference, hypothesis, recommendation and unknown |
| Human-in-Control | Keep human authority over purpose, limits, risk, permission, override and consequences |
| Uncertainty Visible | Expose assumptions, quality, sensitivity, disagreement and what would change the conclusion |
| Action Connected | State what behavior, allocation, workflow or threshold changes |
| Learning Closed | Define expectation, result, deviation, possible attribution and adjustment |
| Proportional Complexity | Use the least complexity adequate to decision and risk |
| Traceable Knowledge | Preserve owner, status, version, sources, claims and relations |
| Challengeability | Allow evidence inspection, challenge, escalation and stopping |
| Interface Independence | Select report, dashboard, narrative, conversation, copilot, workflow or agent by decision |
Article 5 — Ecosystem architecture¶
| Layer | Role | It cannot |
|---|---|---|
| CDI | Delimit the proposed domain of practice and research | Claim external consensus through self-definition |
| CDI-BoK | Organize and govern cumulative knowledge | Turn internal repetition into evidence |
| PULSE | Operationalize decisions, action and learning | Be redefined by one interface or case |
| Practice Lab | Test artifacts and generate application evidence | Generalize from a demo without an evaluation design |
Article 6 — Claims discipline¶
Each material claim states, when applicable: class; owner and source; evidence strength and limits; population, unit, period and context; uncertainty and counterevidence; maturity; and the condition for revision. Original Contribution and standard are controlled terms. Authorship does not prove international originality, and publication does not prove institutional recognition.
Article 7 — Ethical boundaries¶
The CDI-BoK does not authorize fabricated certainty or causality; conversational access that bypasses permissions or legitimate purpose; narrative or visual manipulation; open automation of irreversible or high-impact decisions; transfer of accountability to a model, vendor, prompt or agent; nominal human oversight without time, information or power; or automated learning from outcomes without examining luck, confounding and outcome bias.
Article 8 — Evidence and validation¶
A project definition is validated through ratification; an empirical claim through evidence; a recommendation through reasons and conditions; and an intervention through results evaluated with a design proportional to risk.
Claims that a conversational experience improves a decision should compare, when feasible: human or prior-system baseline; process quality and outcome separately; accuracy, calibration, latency and understanding; appropriate trust or reliance; distributional effects, errors and unintended consequences; cost, sustainability and override capacity; attribution and observation horizon.
Article 9 — Evolution and change¶
The Constitution changes through a pull request, impact analysis, an ADR when authority or boundaries change, and explicit owner ratification. Historical versions remain traceable. A constitutional change explains the problem, affected concepts, evidence and counterargument, first- and second-order risks, compatibility or migration, and the result that would justify keeping or reverting it.
Article 10 — Success criterion¶
The CDI-BoK is not evaluated by page count, technological sophistication or textual fluency. It progresses through less ambiguous definitions, better-specified decisions, more trustworthy and traceable evidence, more visible risks and trade-offs, clear action ownership, observable results and learning that changes the next cycle.
The CDI-BoK aspires to become an international reference. That remains an aspiration until sufficient external adoption, review and evidence exist.