Source-governed scenario package
Controls AI drift by forcing every downstream product to trace back to an approved source layer.
AI capstone case study
CADE shows what managed AI production looks like under real constraints.
Human design authority. AI production speed. Source-truth discipline. Execution evidence.
Framework Combined Arms Decision Exercise
A consequence-based staff training framework built around fixed decision rhythm, controller discipline, and learning capture.
AI role Production engine, not design authority
AI accelerated research, drafting, and artifact iteration while human governance controlled purpose, standards, and approval.
Product move Unified Controller Package
The runtime artifact consolidates execution logic, role aids, decision prompts, and review tools for independent facilitation.
Featured Case Study
View allAI Operating Decisions
View allCADE is not presented as a generic AI project. The decision records show how AI was governed: where it accelerated production, where it was constrained, and where human judgment stayed authoritative.
Use AI as the production engine, not the design authority CADE required fast production of research, planning artifacts, turn content, and controller materials, but the training effect depended on human judgment about purpose, boundaries,... Read decision Change CADE only when evidence justifies changing the system A reusable training framework needs stability. If every execution creates ad hoc changes, the core training effect erodes and future adopters cannot tell which parts are canonical. Read decision Make the Controller Package the runtime center of gravity A CADE event succeeds or fails in execution. Controllers need to find prompts, role guidance, adjudication aids, timing cues, and review structure under pressure without relying on... Read decision Use deterministic adjudication bands tied to observable behavior CADE depends on consequences that feel credible to participants and repeatable to controllers. If outcomes depend too heavily on individual controller judgment, different controlle... Read decision Anchor CADE artifacts to OPORD-quality source truth CADE produces multiple downstream artifacts: scenario materials, turn prompts, controller aids, decision products, briefings, and review structures. Without a governing source laye... Read decision Use plain-language and visual-forward delivery CADE may be executed with mixed-language audiences and uneven doctrinal familiarity. Dense phrasing and text-heavy products can slow comprehension, disrupt timing, and create avoid... Read decision
CADE Product Architecture
Full case studyCADE is presented as a system, not a document bundle. Each module shows a practical AI operator skill: source control, workflow design, adjudication structure, runtime usability, and evidence capture.
Turn engine
Turns training intent into repeatable workflow logic that AI can support without changing the exercise purpose.
Adjudication model
Converts qualitative staff behavior into structured decision support without outsourcing judgment to vague impressions.
Controller delivery package
Designs AI-produced artifacts for real human use under time pressure.
Learning capture layer
Closes the loop between AI-assisted production, live execution, and evidence-driven iteration.
How A Turn Works
Read mechanicsCADE uses a fixed decision rhythm. The value is not improvisation; it is repeatable pressure, visible tradeoffs, and consequences that carry forward.
Establishes the operational update and decision context for the turn.
Allows the staff to resolve essential uncertainty before deliberation begins.
Forces cross-functional coordination and tradeoff discussion inside a time box.
Requires explicit commitment instead of deferral or open-ended discussion.
Turns staff reasoning into a commander-facing recommendation.
Applies consequences and sets conditions for the next turn.
Evidence Over Polish
The site keeps sensitive details generalized for now, but the framework is still presented around observable execution signals rather than vague claims.
Confirmed across all three executions. S1, S2, S3, S4, XO, Medical, Engineer, Fires, and additional staff positions operated inside the framework without designer support during turns.
Confirmed on first execution. Identified fragmented artifacts as a failure point — led directly to the unified Controller Package architecture in subsequent runs.
Confirmed. Resource depletion, casualties, timing constraints, and degraded options from earlier turns visibly shaped staff behavior in later turns.
Confirmed. Protected AAR windows held across all sessions. Execution observations fed directly into the v3.0 redesign.
Where CADE Fits
CADE is intentionally scoped. It is strongest when the goal is decision behavior under pressure, not when the event requires technical simulation fidelity or formal doctrinal certification.
- Training time is compressed and setup overhead must stay low.
- The primary objective is decision behavior under pressure.
- The staff needs to practice cross-functional integration, not isolated section work.
- Language friction or uneven proficiency is present or anticipated.
- The event owner needs a repeatable format that can be facilitated from a Controller Package.
What This Site Is Built To Demonstrate
Framing ambiguous requirements · governing AI output · preserving source truth · designing human-usable artifacts · testing against execution behavior · iterating from evidence instead of polish
Talk About CADE
Reach out to discuss CADE, AI-assisted training design, or AI operations in high-consequence domains.
vincent.taijeron@gmail.com