Harnessing Generative AI for Behaviour Change: My Theory of Change Prompting Model
The rise of generative AI has opened up new possibilities—not just for productivity, but for tackling some of the most pressing challenges of our time. Over the past months, I’ve been experimenting with ways to align AI with behaviour change for climate action and decarbonisation. The result is a framework I call the Generative AI – Theory of Change Prompting Model.
This model helps me (and can help others) design and deliver projects with environmental and social benefits. But it’s flexible enough to apply in any setting where you want to move from ideas → action → impact.
The Model in Brief
At its heart, the model blends two concepts:
Theory of Change (ToC): a structured pathway linking inputs → activities → outputs → outcomes → impact.
AI Prompting Styles: six ways of engaging AI—clarifying, exploratory, analytical, roleplay, storytelling, and iterative refinement.
When combined, these provide a clear roadmap for how to use AI not just for quick answers, but to support deeper thinking, planning, and communication across a project lifecycle.
How It Works
Here’s a snapshot of how prompting styles map onto the stages of the Theory of Change:
Clarifying & Framing: Define problems and baselines.
Exploratory: Generate multiple options and perspectives.
Analytical: Compare, critique, and prioritise.
Scenario / Roleplay: Test assumptions and empathise with different stakeholders.
Creative & Storytelling: Connect emotionally and communicate impact.
Iterative Refinement: Polish and adapt until it fits the context.
For example, if you were designing a decarbonisation program:
Inputs: AI helps clarify the emissions baseline using real data.
Activities: AI generates 10 potential strategies for reducing carbon across operations.
Outputs: AI drafts your action plan or stakeholder briefing.
Outcomes: AI roleplays resistant stakeholders to refine engagement approaches.
Impact: AI helps craft stories that illustrate long-term climate and community benefits.
Why It Matters
Many AI conversations focus on efficiency and productivity. This model shifts the lens towards impact—using AI not just to save time, but to co-create solutions with people, accelerate behaviour change, and support systemic transformation.
In the case of decarbonisation, it provides a practical way to combine data-driven insights, creative engagement, and iterative planning to deliver climate strategies that are more robust, inclusive, and actionable.
Best Practices I’ve Learned
Don’t settle for the first response—refine iteratively.
Mix prompting styles as your project evolves.
Always ground outputs in evidence and lived experience.
Use AI as a co-creator, not a replacement for human insight.
Closing
I’ve licensed the model under Creative Commons so it’s free for anyone to use and adapt. My hope is that it helps more people tap into AI as a tool for meaningful climate action and beyond.