
In my decade as a product strategist, I've witnessed countless shifts in how teams build digital products. But nothing compares to what's happening right now with AI-powered prototyping. As CPO at Basis, where we run design sprints for product-led startups, I've seen firsthand how AI Copilot tools are fundamentally changing the game.
The traditional approach to product development has always frustrated me. Teams spend months building features based on assumptions, only to discover users don't want them. It's expensive, time-consuming, and frankly, a waste of talent. But AI is changing this equation dramatically.
From Representations to Reality
Traditional prototyping typically means creating visual mockups in tools like Figma. While these look real, they're ultimately just representations of ideas. They work in many cases, but they're limited in what they can demonstrate and test.
AI-powered prototyping flips this model entirely. Instead of building representations, we're now creating the real thing.
Working with an AI Copilot directly in our IDE, we can generate functional front-end and sometimes back-end code. The result? A prototype that isn't just a picture of a product—it's an actual working version that users can interact with meaningfully.
This shift matters because it collapses the traditional development cycle. In a recent project with an AI product company, we created a functional prototype in the same time it would have taken to design static mockups in Figma. Even more impressive, the client's development team was able to reuse portions of our prototype code in their production release.
Research at Unprecedented Depth
The transformation begins before a single line of code is written. AI tools enable us to conduct much deeper research around problem spaces than previously possible.
In traditional sprints, research is limited by time constraints and team bandwidth. With AI assistance, we quickly synthesize information from a wide range of sources, becoming temporary subject matter experts in days rather than weeks or months.
This enhanced research capability means we start with ideas that are firmly based in reality rather than best guesses. During a recent project, this approach allowed us to fundamentally rethink information hierarchy and structure within the client's platform. The insights we gained created a holistic understanding of how the application would fit together and its place in the broader marketplace.
This means that even elements that weren't included in the final prototype contributed valuable perspective to the overall solution.
Human Collaboration Remains Essential
Despite these advances, I've found that human collaboration remains the backbone of successful prototyping. The worst mistake teams make is jumping straight to AI-generated solutions without laying proper foundations.
At Basis, we never create in a vacuum. Before spinning up any AI tools, we ensure we have a solid understanding of the team we're working with, their operating environment, and the current state of their product if it exists.
Our design sprints are conducted with clients, not for them. Our first priority is getting holistic alignment on the problem we're solving and buy-in from all stakeholders. The workshop exercises help teams align on solutions together, which means when we do leverage AI tools, they're amplifying a human-centered process rather than replacing it.
This approach ensures the AI serves the business goals rather than becoming a distraction or, worse, leading teams down technically impressive but strategically misaligned paths.
Know When to Use AI (And When Not To)
There's still room to pick your battles when integrating AI tools. Not every prototype benefits from an AI-centric approach.
For example, when working with complex data where calculations need to be precise, defaulting to AI-generated code might not be the best approach. We've found it's sometimes more efficient to present hard-coded values in a visual prototype than to spend valuable time getting calculations to work flawlessly, especially when those calculations won't be part of the permanent solution.
Two effective approaches have emerged in our practice:
First, you can prime the AI Copilot by explicitly explaining that you're building a prototype and want to show hard-coded values rather than implementing actual calculations.
Alternatively, you can fall back on robust design tools like Figma to represent ideas effectively when functional prototypes would create unnecessary complexity.
The key is selecting the right tool based on the specific requirements of the prototype. This measured approach ensures AI enhances rather than complicates the process.
The Future: Design and Development Convergence
Looking ahead, I believe we're witnessing the beginning of a profound shift in how products are created. The traditional bridge between design and development is shrinking rapidly. Soon, there will be complete overlap between these previously distinct disciplines.
It's inevitable that designers and developers will work on the same core deliverables, just expressing their needs to AI Copilots in different ways. This convergence is happening faster than most realize, driven by the accelerating pace of AI tool development.
At Basis, we're preparing for this future by shifting our team's focus to higher levels of abstraction. Instead of getting caught up in implementation details like layout specifications, we're concentrating on strategic questions: How will this prototype impact the business? What specific outcomes do we want to see from user testing?
This approach refocuses everyone on higher-impact work throughout the entire design sprint process. It's a powerful training tool for both our team and clients, teaching them to prioritize the right kinds of problems.
Microsoft's Dragon Copilot: A Case Study in AI-Powered Development
Microsoft's Dragon Copilot for healthcare professionals exemplifies the principles of AI-enhanced rapid prototyping. By focusing on alleviating administrative burdens for clinicians, Microsoft identified a genuine problem that affects healthcare delivery at a fundamental level.
Product teams can learn from Microsoft's focus on specific, measurable outcomes: increased clinician efficiency, reduced burnout, and improved patient care. These clear metrics guided development decisions and provided benchmarks for success.
What's particularly notable is how Microsoft is rolling out Dragon Copilot—starting with a focused geographic release (U.S. and Canada) before expanding to Europe. This measured deployment approach allows for refinement based on real-world feedback, a principle that should be applied to all AI product development.
Getting Started With AI-Enhanced Prototyping
If you're looking to implement AI-enhanced rapid prototyping in your organization, start by focusing on problems, not solutions. Define clear business outcomes before exploring technical approaches.
Invest time in understanding your users and their context deeply. AI can accelerate this research but can't replace genuine human understanding of the problems you're solving (and who you're solving them for).
When selecting AI tools, prioritize those that integrate with your existing workflow rather than requiring complete process overhauls. The best AI implementations enhance your team's capabilities without forcing them to learn entirely new systems.
Most importantly, maintain a collaborative, human-centered approach throughout. The goal isn't to have AI build your product—it's to have AI help your team build better products faster.
The future belongs to teams who use AI as a force multiplier rather than a replacement for human creativity and judgment. By combining the speed and efficiency of AI with the strategic thinking of experienced product teams, we can create better products in less time while delivering experiences that truly resonate with users.