Explore the model
The public LARA model

From experience to decision—and back to learning.

LARA preserves what the organisation knows, qualifies it for the situation at hand and connects decisions with what happens next.

LARA Overview — The Decision Intelligence Operating Model A human-accountable, AI-assisted decision sits at the centre, surrounded by the four LARA acts — Learn, Assess, Reason, Act — and the seven LARA dimensions: Context, Lifecycle, Pattern, Signal, Response, Confidence and Traceability. Case memory and organisational learning memory feed the model through humans, AI, domain expertise and business logic. Case memoryCurrent cases anddecisions Organisationallearning memoryTrusted, reusableexperience Humans AI Domain expertise Business logic SignalDetect weak signalsand early indicators PatternUse experiencewithoutassuming it fits LifecycleConsider timing,phases anddependencies ContextMake sense ofthe situation ResponseIdentify optionsand potentialimpacts ConfidenceBe explicit aboutcertainty and risk TraceabilityEverything isattributed andreconstructable LearnCollect andunderstandevidence AssessEvaluate contextand uncertainty ReasonInterpret,synthesise andchallenge ActDecide,implement andreview outcomes DECISION Human accountable. AI-assisted.
LARA Overview — The Decision Intelligence Operating Model A human-accountable, AI-assisted decision sits at the centre, surrounded by the four LARA acts — Learn, Assess, Reason, Act — and the seven LARA dimensions: Context, Lifecycle, Pattern, Signal, Response, Confidence and Traceability. Case memory and organisational learning memory feed the model through humans, AI, domain expertise and business logic. Case memory Organisationallearning memory Humans AI Domain expertise Business logic Signal Pattern Lifecycle Context Response Confidence Traceability Learn Assess Reason Act DECISION

Two enabling roles

  • LARA Steward
  • LARA Engineer

The four responsibilities

Learn — Preserve experience without losing context

Capture evidence, experience and outcomes without losing their context. Keep observation, interpretation and uncertainty distinct.

Assess — Decide what deserves attention here

Test relevance, recency, confidence, contradiction and gaps before reusing past experience.

Reason — Put qualified knowledge to work

Use qualified knowledge to compare options, test assumptions and expose risk. LARA supports reasoning; it does not dictate the answer.

Act — Make accountability and outcomes visible

Preserve the decision, rationale and accepted uncertainty. Connect outcomes back to what was known.

LARA is a connected operating model, not a four-step process.

Knowledge is only useful when its limits are visible.

Seven lenses expose the limits of information before it becomes the basis for action.

The seven LARA lenses and the question each one asks
LensQuestion
ContextUnder what conditions did this arise, and might it apply here?
LifecycleWhen did it become visible, and when could it have mattered?
PatternIs this an isolated event or something that may recur?
SignalWhat observable indication may point to an emerging condition?
ResponseWhat question, test, option or precaution may improve judgement?
ConfidenceHow strong, complete, current and independent is the basis?
TraceabilityCan the origin, reasoning, use and later history be reconstructed?

The lenses do not produce a universal score. They make relevance, uncertainty and provenance visible.

Not every lesson deserves to become organisational truth.

LARA separates information about one situation from learning qualified for wider reuse. A local lesson can remain valuable without becoming a universal rule.

Memory supports judgement. It does not replace it.

LARA is not software, a database, a domain expert or a decision-maker. It makes the basis and limits of a decision more visible; it does not promise the right answer.

Meet the two LARA rolesLearn how LARA may be used