Founder narrative
Why We Built Duplios
Governments do not fail because people stop caring.
They fail because decisions become harder every year.
The decision problem
Information is fragmented.
Evidence arrives too late.
Institutional knowledge disappears when experienced people leave.
Technology has promised to solve these problems, yet too often it has produced more dashboards, more reports and more complexity.
We believed there was a better way.
Not Artificial Intelligence replacing people.
Not automation replacing accountability.
But technology helping public servants make better decisions.
That belief became Duplios.
Fragmented government evidence
Public decisions depend on budgets, legislation, operational data, institutional responsibilities and lived service pressures. That evidence usually lives in different systems, owned by different teams, refreshed on different cycles.
When leaders must decide quickly, the risk is not only missing information — it is acting on partial information as if it were complete.
Disappearing institutional memory
Governments lose capability when experienced officials retire, rotate or move on. The lessons of previous decisions — what worked, what failed, what assumptions were wrong — often leave with them.
Duplios treats institutional knowledge as something that must be structured, governed and steward-reviewed — not as something that silently evaporates between administrations.
Beyond dashboards and disconnected reports
Dashboards can show status. Reports can summarise a point in time. Neither, on its own, is a decision environment.
A serious decision needs evidence, options, assumptions, authorities, implementation pathways and monitoring — held together so a recommendation can be explained and challenged.
Why human control matters
Everything we build follows one simple principle:
- Humans remain accountable.
- Evidence remains visible.
- Every recommendation remains auditable.
Duplios supports judgement. It does not replace elected representatives, officials, analysts or subject-matter experts.
Why transparent uncertainty matters
Historical, partial, proxy and conflicting evidence must be labelled rather than converted into false certainty. Leaders deserve to know what is known, what is assumed and what still blocks a defensible answer.
Why evidence should remain visible
A recommendation without an evidence trail cannot be defended in cabinet, council, court or public scrutiny. Duplios is built so the basis of a recommendation can be reconstructed.
Implementation and monitoring are part of decision quality
A decision is not finished when it is approved. Implementation plans, monitoring indicators and outcome comparison determine whether the recommendation was sound — and what should change next time.
Learning must be steward-reviewed
Potential lessons are reviewed before they are allowed to influence future recommendations. Learning is never silent. Unreviewed outcomes do not become hidden policy.
Artificial Intelligence is a technique. Decision Intelligence is the product.
Artificial Intelligence is one of the techniques we use. Decision Intelligence is the product we are building.
Helping governments make better decisions.
Helping governments make better decisions
Continue
Read The Journey of Duplios, explore the Operational Capability Programme, or review Trust and Security.