How I use AI to create product prototypes
AI is not a replacement for engineers, but it can be very useful to product managers testing hypotheses and product ideas.
Turning a promising idea into a product prototype can be a slow, resource-intensive process. Coordinating schedules, getting design mockups, and pulling engineers for early prototypes takes time and resources we often don't have.
AI tools are changing this. I'm using them to significantly speed up my workflow, particularly in the ideation and prototyping phases. Technical PMs stand to gain a lot from emerging AI tools, but even non-technical PMs can benefit immensely.
Here’s a practical look at how it works.
Rapid prototyping with AI
When it's time to visualize an idea, here are my go-to AI-powered approaches:
Direct Idea-to-Prototype: If I have a clear concept defined, I turn to tools like Lovable or Bolt, platforms that take a prompt and create a full-fledged application. I feed them a product description using natural language, and they generate a functional web app prototype. From there, I iterate directly using simple prompts: "Change the button color," "Add a user profile section," etc.
Hybrid Approach for Complex Edits: Sometimes, the natural language prompts hit a limit. If I need finer control or more complex code changes than Lovable understands, I export the project (often to GitHub). Then, using VS Code with Gemini Code Assist, I make the necessary code adjustments quickly before bringing it back into Lovable for further no-code iteration. This blends AI speed with precise developer control.
Figma-First for Complex Designs: For ideas needing a more intricate UI/UX foundation, I start in Figma. I'll create the core design there, potentially using Figma's own AI features to accelerate wireframing or user flow generation. Once the design structure is solid, I import it into Lovable to instantly add functionality and interactivity, creating a working prototype without manual coding.
Enhanced ideation with AI
Before prototyping, robust ideation is key. AI accelerates my research and planning significantly:
AI-Powered Market Research: I use the Deep Research features of Gemini, ChatGPT, or Grok to fast-track market analysis. I often start with my core "Jobs-to-be-Done" (JTBD) description as input. The AI helps me perform competitor research, identify potential market gaps or opportunities, and even generate a first draft of a Product Requirements Document (PRD).
Connecting Research to Prototypes: The insights gathered from AI-driven research directly feed into the prototyping tools mentioned above. This creates a tight loop: faster research informs faster prototype creation, allowing for quicker hypothesis testing. Sometimes, using advanced reasoning models like OpenAI o3 and Gemini 2.5 Pro can help better bridge the gap between research and prototype.
Team Collaboration Integration: I also leverage AI within existing workflows. For instance, using Slack integrations, I can capture key points from team brainstorming sessions and feed them directly into an AI model (like Gemini 2.5 Pro) to automatically generate a structured research plan. The results then guide the prototype creation.
Why this AI-assisted approach matters
The primary benefit is “speed to validation.” Using these AI tools, I can often test an idea with a functional prototype in hours or days, not weeks. This bypasses the traditional bottleneck of coordinating designers and engineers for early-stage explorations. If it’s valid, we can move on to allocating resources to create a full-fledged product. If it doesn’t work out, we have invalidated a hypothesis at a fraction of the time it would have normally required and can iterate or move on to the next hypothesis.
This agility is particularly crucial for:
Startups: Where resources and runway are tight.
Small Teams: Lacking dedicated design or engineering bandwidth for speculative work.
Innovation Hubs: Needing to rapidly test multiple concepts.
AI isn't replacing your engineers (yet)
Let's be clear: the prototypes generated by tools like Lovable and Bolt are fantastic for validation and iteration, but they typically aren't production-ready. They often lack the polish, scalability, security hardening, and seamless integration capabilities required for a live product.
You still need your skilled engineering team to take a validated concept and build it properly for scale and integration into your existing tech stack. AI accelerates the “what” and “why”; engineers are crucial for the robust “how.”
Looking forward
The integration of AI into product workflows is just beginning. I envision a future where:
AI assistants actively monitor team discussions (e.g., in Slack or Teams) and proactively suggest relevant ideas, market trends, or potential risks.
AI automatically analyzes customer feedback, support tickets, and usage data, surfacing actionable insights and trends for the product team without manual deep dives.
You will still be in full control. But by embracing AI tools now for specific tasks like ideation and prototyping, you can significantly boost your productivity and get valuable feedback much faster. Start experimenting—the efficiency gains are real.