How Might We – the underestimated starting point for real innovation
Introduction
AI-powered design processes offer enormous potential – but only if the underlying problem is clearly defined. Without focus, even the best AI ends up generating solutions that miss the mark. In this post, we’ll explore why starting with a “How Might We” question isn’t just a nice method, but essential for developing relevant, user-centered solutions.
“A well-defined design challenge serves as the foundation for problem-solving and AI experimentation.” – Daniel Stump
Why the Way You Frame a Problem Determines Success
Whether you’re optimizing an online shop, designing a digital service, or improving internal workflows – AI can only unfold its full potential if the challenge is framed clearly. In the UX context, we call this a design challenge statement.
Vague intentions like “We want to be better” won’t get anyone far – neither humans nor machines. A strong, well-structured question sounds more like:
How might we support service teams in anticipating customer needs before they arise?
These types of “HMW” questions narrow the focus to a specific, solvable problem, allow for creative interpretation, and can be directly integrated into AI-assisted workflows.
Why This Matters in AI Contexts
Artificial Intelligence generates results based on patterns and context. If the context is unclear, the solutions tend to miss the real needs – whether it’s an irrelevant chatbot flow, a confusing recommendation engine, or an interface that doesn’t make sense.
A clear design challenge helps you:
- Write better prompts for AI tools
- Generate more relevant ideas
- Compare human vs. AI thinking more effectively
Practice Example: Design Jam with Real-World Challenges
In AI design sprints or co-creation formats at Quark or Brutalist Solutions, we like to start with a predefined set of challenges – or we co-create them with our clients. Examples include:
- How might we use AI to improve customer experiences without losing the human touch?
- How might we make e-commerce product recommendations more relevant and less creepy?
- How might we support lifelong learning through AI-powered tools?
The key insight: when participants bring in their own real-world problems, the process becomes instantly applicable – and the value of AI becomes tangible.
Reflection: Framing is Half the Solution
Whether you’re designing for older adults, building a payment system, or supporting customer service teams – the way you frame the problem determines the quality of the outcome.
Ask yourself:
- What exactly do we want to improve – and for whom?
- Which aspect is open enough to invite different ideas?
- How specific can we be without becoming too narrow?
Conclusion
A well-structured design challenge is the foundation for any successful UX process – and even more important when working with AI. Taking time to craft a thoughtful “How Might We” question is an investment in real, meaningful innovation.
What’s Next
🔗 Learn more about our UX/UI design process
🔗 Or discover how Brutalist Solutions integrates AI into Shopify projects
Ready to reframe your challenge together?
We help you define your design challenge in a way that AI can actually support – precise, empathetic, and effective.
