Experiment
Policy Impact Mapper
Rapidly trace the likely knock-on effects of a proposed policy change across actors, services, operations and outcomes.
01
Problem to address
Policy decisions can optimise one variable while shifting workload, risk or unintended consequences elsewhere in the system. Organisational boundaries make these knock-on effects easy to miss.
02
What the evidence tells us
Policy, operations and outcomes are often disconnected. Strategy teams need to understand what actually happened. Services lack the data to show it. Farmers see the behavioural and environmental effects of scheme design on the ground.
03
The proposition to explore
The idea
Rapidly map the plausible knock-on effects of a proposed policy change across actors, services, operations, behaviours and outcomes before implementation.
04
Hypothesis
If policy teams can rapidly see plausible first-, second- and third-order consequences, they can identify operational, behavioural and equity risks before detailed implementation.
05
How it could work
Represent the relevant actors, service steps, incentives and outcomes. AI traces plausible consequence chains and attaches supporting/contradictory evidence where available. Unknowns become explicit research questions.
06
Example: bring it to life
Increase a payment by 30% → attractiveness rises → applications rise → RPA processing and verification demand rise → budget is consumed faster → environmental gain depends on whether uptake occurs on suitable land.
07
AI opportunity
Systems reasoning, causal-hypothesis generation, evidence retrieval and scenario comparison. It maps plausible consequences; it does not prove causality.
08
Potential value
- Whole-system policy conversations
- Earlier visibility of unintended consequences
- Better coordination with operations and service teams
- Clearer research priorities
09
Reduce uncertainty without overcommitting
Smallest meaningful experiment
Map one proposed policy change manually and with the AI mapper. Review with policy, ops, service design and behavioural science experts; record which consequences were genuinely useful.
10
Open questions and risks
What we still need to learn
- Plausible chains can be mistaken for predictions
- Complex maps can overwhelm rather than clarify
- Needs disciplined evidence labels