For 20 years, the Double Diamond has helped us crack tough problems, offering structure and clarity when we needed it most. But with AI now reshaping the rules — analyzing at scale, generating faster solutions — this traditional approach is facing pressure. Are we holding onto a framework that no longer fits, or can we evolve it to keep pace with what’s next?

Enter the Stingray Model: a framework designed for the AI era, embracing speed, scale, and non-linear problem-solving. It promises to transform how we think about innovation, breaking away from the sequential nature of traditional design thinking.

But here’s the tension: While AI can analyze thousands of customer support tickets in minutes and generate dozens of solution concepts instantly, it can’t replicate the deep understanding that comes from sitting with users, observing their struggles, and slowly piecing together the why behind their actions. The Stingray Model embraces AI’s speed and scale, but does it risk sacrificing the depth that makes design thinking powerful in the first place?

As someone who has led AI-powered design initiatives, I’ve witnessed firsthand how AI, and especially generative AI, is forcing us to reshape our approach to problem-solving, to collaboration, and to testing. I’ve seen both the promise and the pitfalls: How AI can uncover patterns in user data or even suggest a variety of solutions, but also how teams can get seduced by AI’s speed, rushing to solutions before truly understanding the problem.

I believe the emergence of the Stingray Model reflects this tension, acknowledging that our traditional methods need an upgrade while raising critical questions about what we might lose in the process.

While I should note that I haven’t strictly applied this framework in its formal definition, I have explored fluid design processes that align closely with what’s now being called the Stingray Model. This perspective, combined with my experience in applying the Double Diamond in AI-driven contexts, allows me to examine both approaches critically.

A Diamond Under Pressure

To understand why we’re seeing calls for change, we need to first appreciate what made the Double Diamond so effective. Its genius lies in its simplicity: two consecutive diverge-converge cycles, first to define the right problem, then to create the right solution.

This framework has literally guided billions of dollars in investments and helped teams around the world structure their thinking.

But the world has changed dramatically since the British Design Council introduced this model in 2004. Back then, design teams would:

• Spend weeks conducting in-person interviews and observations
• Gather around walls covered in Post-It notes to find patterns
• Run workshop after workshop to generate ideas
• Create and test prototypes one at a time

Each of these activities was constrained by human cognitive capacity and physical limitations: You could only interview so many users, process so much information, or test so many prototypes in a given timeframe.

I believe that these constraints actually helped by forcing teams to be thoughtful and selective. But in an AI-powered world, many of these constraints disappear. When you can analyze millions of user interactions and test multiple variations of prototypes simultaneously, should your process still follow the same sequential path?

The Technological Shift

When we talk about AI transforming design, we’re not just talking about faster computers or better tools. We’re talking about a fundamental shift in how we understand users, generate solutions, and validate ideas.

The emergence of Large Language Models (LLMs) has introduced capabilities that were almost unimaginable when the Double Diamond was conceived. These models don’t just process information; they can generate creative solutions and synthesize vast amounts of information in ways that complement human creativity rather than replacing it. Beyond just processing more data or generate output, these models can identify complex patterns, predict user needs, and even anticipate problems before they surface in traditional metrics.

This increasingly challenges the linear nature of traditional design thinking: when we can simultaneously explore problems and solutions at scale, the distinct phases of frameworks like the Double Diamond begin to feel artificial.

Let’s take user research: AI can help analyze thousands of user comments, support tickets, or interview transcripts in minutes, identifying patterns and insights that might take human researchers weeks to uncover. It can cluster similar complaints, spot emerging issues, and even predict future pain points based on current trends.

But here’s the critical question: Are we gaining breadth at the expense of depth? While AI excels at finding patterns in large datasets, can it capture the nuanced understanding that comes from in-depth user interviews and observation?

The same capabilities benefit designers: In ideation, AI can generate hundreds of solution concepts based on a prompt, expanding our solution space far beyond what a typical brainstorming session might produce. It can cross-pollinate ideas from different domains, suggest unconventional approaches, and even test basic feasibility in real-time.

But again, we must ask: Are we generating more ideas, or better ideas? Does quantity help or hinder our search for truly innovative solutions?

Enter the Stingray Model

The Stingray Model, developed by Board of Innovation, attempts to address these new realities with three key phases that reflect how AI is changing design: Train, Develop, and Iterate. Each phase reflects a fundamental shift in how we can approach design and innovation in an AI-enabled world.

Train: Setting the Foundation

The training phase represents perhaps the most significant departure from traditional design thinking. Instead of starting with open-ended user research, teams begin by:

1. Defining clear parameters for AI training data

• What historical data is relevant?
• What guardrails do we need?
• What gaps need to be filled?

2. Establishing success metrics that balance:

• User satisfaction
• Technical feasibility
• Business viability
• Ethical considerations

3. Creating initial prompts and guidelines for AI systems

• What types of solutions are we seeking?
• What constraints must be respected?
• What values should guide the AI’s outputs?

This structured approach to data and parameters might seem counterintuitive to designers used to starting with a blank slate. However, it acknowledges a crucial reality: in AI-powered innovation, the quality of your outputs depends heavily on the quality of your inputs.

Develop: Parallel Exploration

Here’s where the Stingray Model really breaks from tradition. Instead of moving sequentially from problem to solution, teams explore both simultaneously. The Stingray Model suggests that with AI assistance, teams can simultaneously:

• Generate and test problem hypotheses based on pattern analysis
• Create and evaluate solution concepts in parallel
• Assess technical feasibility through rapid prototyping
• Validate business assumptions with predictive modeling

This parallel processing capability addresses one of the major criticisms of traditional design thinking: that teams often spend a lot of time understanding a problem only to discover their eventual solutions aren’t technically feasible or economically viable.

Iterate: Continuous Refinement

The final phase recognizes that with AI-powered systems, the line between development and deployment blurs. Solutions need constant refinement based on real-world usage. This means:

• Using AI simulations to test variations
• Validating with real users continuously
• Optimizing technical performance
• Refining the business model based on actual usage

This approach recognizes that with AI-powered systems, the line between development and deployment is increasingly blurred. For one, because solutions increasingly need to be tested with real data, but also continuously be refined based on real-world usage.

Evolution, Not Revolution

The Stingray Model sure presents exciting possibilities for design in an AI-enabled world, but it’s not without its challenges. While it embraces AI’s capabilities, it also makes some big assumptions:

First, it assumes teams have the AI maturity to implement it effectively. This isn’t just about having data scientists on staff — it’s about understanding how to:

• Define meaningful success metrics for AI systems
• Create effective training data and prompts
• Balance AI insights with human judgment
• Maintain ethical oversight and accountability

Second, the model’s emphasis on speed and parallelism could lead to superficial understanding. When you can generate and test solutions instantly, it’s tempting to skip the deep problem exploration that often leads to breakthrough insights. AI may helps us move faster, but sometimes we need to move slower to move better.

Finally, the Stingray Model’s non-linear nature could lead to a fragmented process if not implemented thoughtfully. Teams might find it challenging to align on priorities, track progress, or ensure accountability when phases like problem discovery and solution development overlap. Without clear coordination, the framework could feel chaotic rather than fluid. But then again, when is the Double Diamond ever implemented smoothly?

From Framework to Practise

Perhaps the most significant shift isn’t even in any particular process, but in AIs potential to transform our everyday design practice. This transformation might turn out to be most evident in the evolution of our tools.

Traditional design tools are built around artboards that present layouts in a more or less static manner. Prototyping tools aim to bridge the gap between showing a static experience with a more interactive, albeit not dynamic, experience. Both only really work well for linear, non-generative flows.

Tomorrow’s tools will need to enable more dynamic, state-based design systems where components respond to (mock) context and user behavior. This shift will challenge fundamental assumptions about how we design, document and specify design work.

Consider a modern component library: components might already have dozens of states today, but they are selected by the designer for the specific case. In the future, they’ll have to be able to respond to various types of user input, and adapt to different contexts, in order to effectively define and showcase AI-driven experiences. Traditional documentation methods, with static mockups for each state, become impractical. Instead, we’ll move toward designing systems and rules rather than specific instances.

More significantly, these tools are evolving from enabling collaboration between different designers, or between designers and developers, to enabling collaboration with AI. Design tools will need to evolve into co-pilots, into co-designers, if you will. They won’t just automate tasks — they will understand design intent and help with design decisions.

This challenges not just the Double Diamond, but any design process we try to define. When a tool can grasp the problem you’re trying to solve, generate multiple solution variants, and simultaneously test them against guidelines, requirements, and constraints, the traditional boundary between Ideation and Validation in the Double Diamond becomes just as blurred as it does between Develop and Iterate in the Stingray. What this demands is a more fluid, non-linear process than the distinct phases of either framework.

Practical Steps Forward

Regardless of how exactly tools and processes evolve, teams can already adapt their ways of working to a future where AI-driven products become the new norm. Here are some practical tips to prepare:

• Invest in training
  Teams must be equipped with the skills to use AI effectively. This includes training designers in prompt engineering, AI and general data literacy, and understanding the strengths and limitations of various AI tools and models.
• Maintain human oversight
  While AI can enhance efficiency, human judgment is critical to ensure ethical considerations, cultural sensitivity, and (most importantly) alignment with user needs. Teams should establish clear processes for reviewing AI outputs.
• Start small
  Before overhauling your entire process, test the Stingray or any other model on smaller projects or phases. Use these experiments to refine workflows and identify any gaps in processes or capabilities.
• Foster collaboration
  Any AI-fit design process will require closer collaboration between designers, data scientists, and engineers. Build cross-disciplinary trust and establish shared goals early in the process.
• Avoid over-reliance (or over-focus) on AI
  Balance AI-driven insights with human creativity and intuition. AI is a tool, not a replacement for deep user empathy or creative problem-solving.

Personal Take

The future of design thinking isn’t about choosing between the Double Diamond and the Stingray Model — it’s about evolving beyond frameworks altogether. We need to become more fluid, more adaptive, and more thoughtful about how we combine human and artificial intelligence.

The Double Diamond gave us crucial principles: the importance of problem exploration, the value of divergent thinking, the need for validation. The Stingray Model shows us how these principles might evolve in an AI-powered world. But the real challenge isn’t adopting a new framework — it’s developing the judgment to know when to move fast with AI and when to slow down for deeper human understanding.

Teams could potentially also get the best of both worlds by combining the two frameworks. The Double Diamond can provide the structure and human-centered rigor needed to define the right problem, while the Stingray Model can accelerate exploration and refinement once a direction is established.

Ultimately, it’s about finding the right balance — leveraging AI where it adds value while preserving the human-centered principles that make design meaningful. The key is to stay thoughtful, adaptable, and intentional about how we integrate AI into our processes. Evolving beyond frameworks also means evolving our craft — evolving our processes, tools, mindset and collaboration — all the while ensuring our work remains grounded in solving real human problems.