THE DISHA INTELLIGENCE ENGINE → HUMAN-CAPITAL DECISION INTELLIGENCE
Turn Human-Capital Intelligence Into Better Decisions.
Bring evidence, constraints, scenarios and trade-offs together before deciding what to do next.
What should the organization, institution or government do next — given evidence, constraints, scenarios and objectives?
Signature: SENSE → FRAME → SIMULATE → COMPARE → DECIDE → EXECUTE → LEARN
One system, six questions — you are here:
Signature explorer — a governed decision object
The Decision Cockpit: uncertainty stays visible
Objective, owner, evidence, options, constraints, dependencies, scenarios, risks, approvals, chosen action and review date — one card, fully traceable.
objective
Close the rotation-rostering capability gap before the two-site expansion
owner
VP Operations (human decision-maker)
evidence
Workforce tower: 8 evidenced experts, 5 in one region · adjacency route from planners verified in Skills
options
Hire externally · Redeploy 2 planners via adjacency · Automate part of the scheduling work
constraints
Budget cycle closes in 6 weeks · partner contracts standard terms
scenarios
Expansion proceeds / slows — capability impact differs per lever
risks
Regional concentration persists under hire-only option
approvals
Ops board review · works council consultation (illustrative)
chosen
Pending — decided by the accountable owner, not by AI
review
Review date set at decision time; outcome evidence feeds back into the engine
Framing Wizard
An ambiguous question becomes a governed decision object
Ten steps — click through them. Each step produces a field of the decision card above, not a paragraph of prose.
Step 1 — Objective
What outcome, in whose words? Ambiguity is the enemy — the wizard forces one sentence.
Trade-off Matrix — try it
Your weights. Your trade-offs. No hidden best answer.
Set what matters to you and watch the comparison change. The matrix never outputs a single 'best decision' score — it makes your own priorities comparable.
Weight each dimension (0–10)
Sensitivity: margins under 5 points are ties — inspect the underlying values instead of trusting the order.
Comparable options — under your current weighting
Raw dimension values stay inspectable — Hire externally: capability 7, time 6, cost 4, risk 6, evidence 5 · Redeploy planners (adjacency): capability 6, time 7, cost 8, risk 8, evidence 7 · Automate part of the work: capability 8, time 3, cost 5, risk 4, evidence 3. Normative weightings are yours by design; the AI never disguises them as objectivity.
Brief generator & the learning loop
From evidence to action — and back into intelligence
Executive brief — every claim source-linked
- · Decision statement
- · Context
- · Evidence
- · Options
- · Scenarios
- · Trade-offs
- · Uncertainties
- · Stakeholder implications
- · Proposed actions
- · Owners
- · Measures
- · Review date
Explainability is built in: evidence, assumptions, model version, uncertainty, counter-evidence, dependencies and missing information — plus “What would change this?” sensitivity on every claim.
Action & learning loop
Outcome evidence flows back into Talent, Skills, Learning, Career and Workforce — the ecosystem closes where it began.
Institution & government mode: programme and policy decisions carry explicit jurisdiction, authority, evidence and implementation metadata — political choices stay separate from descriptive evidence and scenarios.
Calculators
Run the numbers — as scenarios, never verdicts
Calculator 1
Scenario Calculator
Hire, develop, redeploy, partner, automate, redesign work or change learning capacity — capability impact, effort, time, dependencies, evidence confidence. Never an unsupported single best-decision score.
Calculator 2
Trade-off Matrix
Your dimensions, your weights — options compared with sensitivity shown. Normative choices stay yours, not the AI's.
Calculator 3
Brief Generator
Decision statement, context, evidence, options, scenarios, trade-offs, uncertainties, owners, measures and review date — every claim linked to its source.
Every decision traces backward to evidence and forward to accountable action and review — with options comparable, and normative choices never disguised as objective AI answers.
Ask the AI Decision Advisor
Supports human judgement — it does not replace decision authority.
- Help me frame this decision.
- What evidence is missing?
- Generate options — including opposing ones.
- Stress-test this scenario. What would change this conclusion?
Grounding rules
- Answers cite the authorized records behind them — or say the evidence does not exist.
- Uncertainty is shown, never smoothed away.
- Recommendations are options for a human to judge — never silent decisions.
- Every answer stays within your purpose and access level.
In the live product these conversations stream from your authorized data. Here, the questions show exactly what the analyst is built to answer.
Interconnections
Where human-capital decision intelligence hands off next
Conceptual navigation across one intelligence system — links carry your context; none of this implies a causal or predictive pipeline.
Technical architecture — how this page is built
Connected systems
Decision object store, evidence & provenance graph, scenario engine, rules/policy layer, workflow orchestration, AI reasoning, approvals, audit trail, action tracker, outcome-learning analytics.
Core objects
Decision, objective, owner, option, constraint, scenario, trade-off, uncertainty, approval, action, metric, review, audit record.
Governance — identical on every Intelligence page
How this system stays honest
These five statement types never blend: everywhere in the engine, you can tell what is measured, what is derived, what is projected and what is advised.
Statement types
Observed, inference, forecast, scenario and recommendation are visually distinguished — never blended.
Illustrative by default
All data, people and outcomes on these pages are labeled illustrative composites.
No fabrication
Never fabricate people, employers, credentials, outcomes or statistics.
Explainable AI
AI explanations expose the evidence and the uncertainty behind them.
Authorized use
Data is used only within its authorized purpose and access level.
Human authority
AI supports judgement; it never replaces human decision authority.
Human-Capital Decision Intelligence adds: Decision authority is explicit and human-controlled. High-impact actions require authorization and review. Evidence, model version, owner and review date are recorded — and people affected by a decision have correction and appeal paths.
The engine behind the pages
Six pages, one interaction contract
Rule 1
Context handoff
Moving between the six pages carries your question and context with you — you never restart from zero.
Rule 2
Shared vocabulary
Skill, evidence, proficiency, readiness and scenario mean the same thing on every page.
Rule 3
Common evidence layer
Every claim carries source, timestamp, provenance, confidence and access level — everywhere.
Rule 4
AI layer boundaries
The AI summarizes, explains, compares and simulates — without inventing facts or silently converting recommendations into decisions.
Rule 5
Progressive disclosure
Overview first; evidence, assumptions and audit trails open on demand.
Journey · Recruit
Journey · Develop
Journey · Move
Journey · Transform
Journey · Institution
Illustrative journeys — conceptual paths across the engine, not claims about how any decision is made.