Intent
GoalConstraintSuccess criteria
The purpose behind the work
LIVING CONTEXT / HUMAN-DIRECTED AI
Cognitive TwinContext that
moves with the work.
A living model of intent, decisions, evidence, and current reality.
Explore the model◎
01 / INTENTPreserve the why
≋
02 / EVIDENCEGround the claims
Cognitive
TwinWORKING UNDERSTANDING
TwinWORKING UNDERSTANDING
◇
03 / DECISIONSKeep the rationale
⌘
04 / ARCHITECTUREConnect dependencies
↻
05 / CURRENT REALITYReconcile change
Persistent contextTraceable relationshipsHuman validation
01 / THE MODEL
Keep the relationships.
Carry the reasoning.
A history becomes useful when its pieces stay connected.
Evidence
SourceObservationClaim
What supports an understanding
Decisions
OptionsTrade-offRationale
Why a path was chosen
Architecture
ComponentInterfaceDependency
What a change could affect
Current reality
Previous stateNew evidenceRevised state
A revisable view of what is known
02 / THE REVIEW
Evidence enters.
Judgment stays human.
Confidence, permission, and completion are separate states.
01
Observe
Source + freshness
02
Reconcile
Agreement + uncertainty
03
Validate
Human review + approval
APPROVAL BOUNDARY04
Carry forward
Decision + confirming evidence
03 / THE CHANGE
One new fact.
A visible chain of effects.
Fictional example: a workshop venue becomes unavailable.
PREVIOUS UNDERSTANDINGBefore
IntentRun an accessible workshop
VenuePreferred venue available
DecisionRecommend that venue
NEW EVIDENCE
The venue is no longer available.
REVISED UNDERSTANDINGAfter review
IntentRun an accessible workshop
VenueAlternatives need checking
DecisionReopen the recommendation
BOOKING PAUSED
Human approval required
RECHECK THE DEPENDENCIES
AvailabilityTimingCostAccessibility
Preserve the intent. Supersede the assumption. Keep the next action bounded.
COGNITIVE TWIN
The context evolves.
The why stays connected.
Justin Johnsen
AI architect · hands-on builderI connect AI architecture, practical implementation, and human judgment.
Back to the model