🧠 Pause & Reflect with ChatGPT

Great experiments aren’t just about accuracy — they’re about learning fast. Use ChatGPT to run a 4Ls review: What did you like, learn, lack, and long for?

Most of us rush from idea to execution — especially when speed is the goal. But the best behavioural designers don’t just move fast — they pause and reflect.

Here’s how to turn ChatGPT into your reflection partner, using three proven frameworks:

1️⃣ Steelman your ideas with Six Thinking Hats

Before you move from insight to intervention, pause to strengthen your design from every angle using Edward de Bono’s Six Thinking Hats.

Ask ChatGPT to guide you through:

🔴 Red Hat (Gut Feelings) – What are your initial gut feelings about this idea? Does anything feel off or exciting before you dive into the data?
White Hat (Evidence) – What existing evidence supports or challenges this idea? How will you test if it works in practice?
🟡 Yellow Hat (Strengths) – What strengths or positive aspects make this idea worth pursuing?
Black Hat (Risks) – What are the potential risks, barriers, or unintended consequences?
🟢 Green Hat (Creativity) – What creative or alternative directions could this idea take? How could you push the boundaries further?
🔵 Blue Hat (Execution) – What are the first practical steps to bring this idea to life? What should happen next?

Pro Tip: Ask ChatGPT to walk you through all six hats after you summarise your intervention in three bullet points — it helps spot gaps faster.

Before you summarise your idea, share these Six Thinking Hats with ChatGPT — then prompt:

"Help me stress-test my behaviour change idea using the Six Thinking Hats. First, ask me to summarise my idea, then guide me through each hat."

2️⃣ Anticipate unintended consequences with IN CASE

Even well-intentioned designs can trigger unexpected backfires to alienating the very people you want to help. Before launching, use ChatGPT to scan for risks and side effects with the INCASE framework, developed by the UK Cabinet Office Behavioural Science Team in 2021.

This behavioural approach helps anticipate unintended consequences by exploring:

🔎 Intended – If this intervention succeeds, can the system or process handle the success, or will it break under pressure?
👥 Non-target – How might this intervention affect people outside your target audience?
🔄 Compensatory – Could people offset the desired behaviour with something negative or counterproductive?
🌊 Additional – What new behaviours (helpful or harmful) might emerge as a side effect?
🪧 Signalling – What unspoken messages does this intervention send — and do they align with your intentions?
😟 Emotional – What feelings might this intervention trigger — and how could those feelings shape future behaviour?

Citation: Emery, A., Molière, L., Lang, P., Nicolson, M., & Prince, E. (2021). In CASE: A behavioural approach to anticipating unintended consequences.

Share the INCASE framework with ChatGPT, then say:
"Please help me apply the INCASE framework to my behaviour change idea. First, ask me to describe my intervention, then walk me through each question."

3️⃣ Reflections with the 4Ls

Every test is a learning opportunity — but only if you pause to reflect.

✅ Liked – What worked well and should be repeated?
💡 Learned – What surprising insights or unexpected patterns emerged?
❌ Lacked – What was missing or caused friction that could be improved?
✨ Longed for – What do you wish you’d included or tested differently?

After your experiment, ask ChatGPT to guide you through a 4Ls reflection. Use this prompt:

"Help me run a 4Ls reflection on my recent behaviour change experiment. Please ask me these four questions and help me capture key insights."