Understand the workstream and institutional outcome.
Establish how the work currently operates, where the binding constraint sits, and which institutional capabilities matter to the outcome.
Institutional transformation for the AI state
PathDelta enables governments to diagnose, simulate and redesign institutions around AI to build stronger, more dynamic and more resilient public institutions.
PathDelta combines human institutional diagnosis, senior expert judgement and simulation to trace how AI changes a consequential public workstream — including the second- and third-order effects on authority, capability, dependency and accountability. We model what AI can do alongside how the people, decisions and institutions around it may need to change — and what human capability must remain.
AI capability is advancing faster than many institutions can absorb it. Governments are building strategies, pilots, governance and technical foundations. The harder question now sits underneath adoption: can a specific institution change how work, authority, judgement and accountability actually move around a real AI application?
AI is moving rapidly from experimentation to expenditure. Gartner forecasts worldwide AI spending to reach $2.7 trillion in 2026. Across OECD governments, AI is already used somewhere in 35 of 36 countries, while only 10 of 36 report measuring the financial or non-financial impact of AI use cases. At this scale, deployment decisions require a clear view of intended outcomes, unintended and second- and third-order consequences, lifetime costs, and whether AI is the right route to the public outcome being pursued.
The limiting factor in government AI will increasingly be institutional capacity, not access to technology.
Establish how the work currently operates, where the binding constraint sits, and which institutional capabilities matter to the outcome.
Test AI deployment against redesigned alternatives and compare likely effects on performance, cost, capability and resilience before implementation.
Translate the evidence into an operating design that officials can govern, adapt and retain capability around.
What changes beyond the task — in workflow, authority, accountability and the wider operating system?
Does capability move into government — or out of it?
Which path delivers the strongest outcome: AI now, redesign first and then apply AI, or redesign without AI?
PathDelta combines a senior core with specialist external expertise when a consequential institutional question requires it.
A bounded institutional problem is enough for an initial discussion; PathDelta can help define the scope from there.