Experiment
DREAM BIG
WHAT IF
Farmers could see the potential combined impact of different actions before deciding what to do?
Outcome Explorer
Help farmers explore what different combinations of actions could potentially achieve on their land.
01
Problem to address
Farmers can be asked to choose from schemes, actions or interventions without a clear, contextual picture of what different combinations could achieve on their particular land. Individual options can be understandable in isolation while the cumulative environmental effect, interactions, trade-offs and uncertainty remain hard to see. This makes it difficult to move from “which option can I apply for?” to “which combination of actions is most likely to help me achieve the outcomes I care about?”
02
What the evidence tells us
Suitability is highly place-specific: an intervention that works in one location may not work in another, and farmers want to understand whether an action will genuinely work on their land before committing. Policy teams also need to know where interventions should be targeted for the greatest environmental return. Existing concepts such as Consequence Preview, Collective Intelligence for Landscapes and Outcome Learning Loop point to the same gap from different perspectives. The specific interactive Outcome Explorer is a DREAM BIG proposition built from this synthesis.
03
The proposition to explore
The idea
Give farmers an interactive way to explore what different combinations of actions could potentially achieve on their land, using the best available evidence while making uncertainty visible.
04
Hypothesis
If farmers can see the potential combined outcomes, trade-offs and evidence behind different action packages in the context of their own land, they will be better able to choose interventions that fit both their business circumstances and intended environmental outcomes.
05
How it could work
Start with what Defra and the farmer know about the holding: land characteristics, existing agreements, environmental context, previous interventions and relevant local evidence. Let the farmer explore combinations of actions rather than a flat menu of scheme options. Show the outcomes each combination may contribute to, where actions reinforce or conflict with each other, what funding may be available, and how confident the underlying evidence is. The experience should support exploration rather than tell the farmer there is one algorithmically “correct” answer.
06
Example: bring it to life
An upland farmer is considering peat restoration, changes to grazing and new habitat actions. Instead of reading three separate guidance pages, they open their holding and explore a combination. The service shows: “Together, these actions could improve water retention and habitat connectivity across this part of the holding. Evidence for peat restoration here is strong; evidence for this grazing change on your soil type is moderate.” The farmer removes one action, adds a hedgerow option and immediately sees how the potential outcome picture changes, including cost, commitment and uncertainty.
07
AI opportunity
AI could make complex evidence explorable in plain language, compare combinations, retrieve evidence from similar contexts and explain trade-offs. More advanced versions might use optimisation or simulation models. AI must not manufacture causal certainty: environmental outcome estimates need appropriate scientific models, provenance and explicit uncertainty.
08
Potential value
- More informed farmer choice
- A clearer connection between actions and environmental outcomes
- Potentially better combinations of interventions rather than isolated scheme uptake
- Greater transparency about evidence and uncertainty
- A bridge between personalised services, policy targeting and landscape-level outcomes
09
Reduce uncertainty without overcommitting
Smallest meaningful experiment
Choose one environmental outcome, one geography and a small set of 3–4 existing actions with credible evidence. Build a manually curated interactive prototype that lets farmers compare combinations and shows the evidence and uncertainty behind them. Test whether farmers understand the trade-offs better, whether the explorer changes the questions they ask or options they consider, and where they still need expert advice.
10
Open questions and risks
What we still need to learn
- Combined environmental effects may be difficult to attribute or model
- Poor evidence could create false confidence
- Farm business priorities may conflict with environmental optimisation
- Recommendations must remain explainable and contestable
- The service must distinguish evidence, modelled estimates and speculation clearly