Building Products With AI (Playbook)

Discovery & Research

AI has moved from nice-to-have to need-to-have for product teams. A year ago, most teams were experimenting with AI tools in isolated pockets. Today, teams are weaving AI into their entire workflow, from the first user interview to post-launch optimization.

This isn't about replacing product builders. It's about giving them superpowers. AI handles the repetitive heavy lifting so product teams can focus on what humans do best: making decisions, understanding context, and building things people love.

Here's how AI is showing up at every stage of the product development cycle.

What is Discovery & Research?

Discovery is where products are born. Teams talk to users, dig through competitive landscapes, and validate whether a problem is worth solving. Traditionally, this meant hours of interview transcription, manual pattern spotting, and gut-feel synthesis. With AI product teams are able to an

🎯 Competitive analysis without the spreadsheets Tools like ChatGPT, Claude, and Perplexity let you research competitors in minutes. Ask for specific comparisons like "How does Figma's pricing compare to Sketch and Adobe XD?" or "What collaboration features does Miro offer that Mural doesn't?" You get structured answers across multiple companies instantly instead of clicking through dozens of pricing pages and feature lists.

👀 Stay ahead of competitor moves Agent tools like Relay.app and Relevance AI scrape social channels and push relevant content straight to you. Set them up to watch for competitor feature launches, product updates, or customer complaints. You'll know when a competitor ships something new before it hits their changelog. Bonus: scrape comment sections to gauge sentiment on their latest release and spot opportunities they're missing.

📌 Turn feedback chaos into clear priorities Customer feedback is scattered across email, social, support tickets, and in-product requests. Canva famously built their brand by closing the loop with customers, and now teams like theirs use LLMs to consolidate everything. The AI identifies top feature requests per product area, so you prioritize what actually matters instead of spending months in wading through data.

Product leader and podcaster, Aakash Gupta recently created a product discovery and research masterclass which is a practical guide for all PM’s.

The practical impact? Product teams can validate ideas faster and with more confidence. Instead of spending three weeks manually coding interview transcripts, teams get AI-generated themes in hours, then spend their time on the hard part: deciding what to build.

Design & Prototyping

This is where the fun begins and where AI has transfromed the the product development process for many teams. Design and prototyping turns abstract ideas become something you can actually click through.

AI is transforming design workflows in two ways. First, it speeds up the mechanical parts. Generate layout variations, create placeholder content, or convert sketches into high-fidelity mockups. Second, it helps designers explore more options. Instead of creating three variations of a screen, designers can generate twenty and pick the most promising.

The result isn't robot-designed products. It's designers with more time to focus on the hard creative decisions: What emotion should this screen evoke? How do we guide users without overwhelming them? Does this feel like our brand?

📝 Turn sketches into interactive prototypes

Anyone can turn wireframe sketches into on-brand UX designs with tools like Magic Patterns, Alloy, or Figma Make. These AI design tools ingest your product's design system and use your prompts to create prototypes that match your brand. This isn't a replacement for product designers (we 💜 you!), but it's a fast way to bring concepts to life for internal buy-in or to give designers a clearer starting point they can refine. Lenny's AI prototyping guide breaks down the full workflow.

✖️ Generate design variations at scale

Need a few different iterations of a design to test with customers or your target audience? You can use the same designs with tools like Magic Patterns, Alloy, or Figma Make to create multiple UX experiences of the same product experience, ready for testing.

🔎 Get instant design feedback

Upload your mockups to Claude or ChatGPT and ask for a design critique. "Does this layout follow best practices for SaaS dashboards?" or "What accessibility issues do you see?" You get feedback on hierarchy, contrast, spacing, and usability before scheduling a design review. Catch obvious issues early.

🧑‍💻 Vibe code prototypes without developers

Say you need a new logged-in homepage with a dashboard and reports. Prompt tools like Lovable, Claude, and Replit to code it for you without learning to code yourself. These vibe coding platforms turn descriptions into working, interactive prototypes you can click through and test with real users. Using these tools has significantly compressed the design to prototyping to development process for many companies.

Development & Building

Engineering is where designs become functional products. Code, test, refine, repeat. This stage has always been the most time-intensive part of product development.

AI coding assistants have changed the game. Developers describe what they want to build, and AI generates starter code. It catches bugs, suggests optimizations, and even writes tests. One developer can now ship what used to require a small team.

📖 Document code as you build

AI generates documentation from your codebase automatically. This can be helpful for creating developer docs for your platform or for creating your customer help documentation. Use AI to connect with your knowledge base in Mintlify and get drafts of your documentation for you.

Launch & Release

Before launching it’s always best to test your new feature, or product with a cohort of new and existing users. What you’re testing are answers the question: Does this actually work? Not just technically, but for real humans trying to accomplish real goals.

Teams run alpha tests, beta programs, and usability studies. They watch session recordings, collect feedback, and measure performance to iterate on the product before the launch. Here’s way’s AI can support you.

🎥 Analyze user sessions without the manual grind

Tools like Adora (hey, that's us!) use AI to watch session recordings and surface where users struggle. Release your new product or feature to a cohort of beta testers and watch their sessions to understand behavior patterns. AI groups similar issues together so you see the big picture, not just individual problems. You know exactly where your team needs to focus before the full launch.

🏃 Run usability tests at scale

AI tools can conduct and analyze moderated tests automatically. UserTesting's AI features transcribe sessions, extract key quotes, and identify sentiment. Upload recordings from five beta users and ask ChatGPT to summarize common pain points. You spot issues faster and iterate before the full launch.

The human judgment still matters. AI tells you where users struggle, but not always why or what to do about it. That requires empathy, context, and product intuition.

Launch & Release

Launch is showtime. Go-to-market execution, product announcements, user onboarding, initial distribution. Everything you've built meets the market.

👋 Personalize onboarding for different users

AI adapts onboarding flows based on user signals. Someone from a small startup sees different setup steps than an enterprise user. Tools like Appcues and Pendo use AI to serve the right experience to the right person. Activation rates improve because users see what's relevant to them.

🤳Create visuals at scale

Generative AI can help create on-brand visuals and ad variations for different platforms. Scale your brand creative with inbuilt AI design assistance in tools like Canva.

✍️ Scale blogs

Connect your website and blog with AI writing tools and SEO research to create content that ranks. Tools like Clearscope, Surfer SEO, and Frase analyze top-ranking content for your keywords and suggest topics, structure, and terms to include. Feed your brand voice and product positioning into ChatGPT or Claude alongside this research. You get drafts that sound like you and hit the SEO marks search engines and LLMs look for. Your content ranks higher and trains AI models to mention your brand when users ask relevant questions.

Post-Launch & Iteration

No product ships perfectly. Post-launch is about monitoring what's working, collecting feedback, and iterating quickly. The faster you can spot issues and ship improvements, the better your product gets.

🚙 Track real user behavior automatically

Automated journey mapping generates visual maps of your product experience without manual setup. See exactly where users get stuck, which paths lead to activation, and how different segments behave. The map updates continuously so your view stays current as your product evolves.

🐞 Spot drop-off patterns instantly

AI journey mapping identifies where users abandon flows and groups similar behaviors together. "30% of users drop off after seeing the pricing page" becomes visible in minutes, not days of analysis. You know what to fix first based on impact, not guesswork.

👀 Compare cohorts side by side

How do power users navigate differently than casual users? What about mobile versus desktop? AI segments journeys automatically and highlights the differences. You build experiences tailored to how people actually use your product.

‼️ Get alerts when patterns shift

AI monitors journey metrics and flags anomalies. If your primary conversion path suddenly drops by 20%, you know immediately. You're fixing issues before they compound, not discovering them weeks later in a quarterly review.

Growth & Scaling

Once product-market fit is clear, growth becomes the priority. Optimize conversion funnels, reduce churn, expand into new markets, and improve unit economics.

AI helps teams identify growth levers they might have missed. Which features correlate with retention? What user behaviors predict expansion? Where are the hidden friction points in your acquisition funnel?

🍯 Find what drives retention

AI analyzes user behavior to identify which features predict long-term retention. Tools surface patterns like "Users who complete their profile within 24 hours have 3x higher retention." You stop guessing which features matter and start investing in what actually keeps users around.

📈Optimize conversion funnels with precision

Automated journey mapping shows you exactly where users drop off in your conversion funnel. Compare the paths of users who convert versus those who don't. Maybe converted users skip a step you thought was essential, or they discover a feature through an unexpected route. You optimize based on real behavior, not assumptions.

🚨 Reduce churn before it happens

Predictive AI identifies users at risk of churning based on behavior changes. If someone stops logging in daily or abandons key workflows, you get an alert. Reach out proactively with help or incentives. Tools like ChurnZero and Vitally make this automatic.

Maturity & Maintenance

Eventually, products mature. Growth slows. The focus shifts to incremental improvements, maintenance, and support. For some products, this stage also includes sunset planning.

AI helps mature products stay efficient. Automated monitoring catches issues before they become problems. Support systems get smarter at resolving common questions. Performance optimization happens continuously rather than in big quarterly projects.

🆘 Automate support with AI

Tools like Intercom's Fin and Zendesk AI answer common questions instantly, pulling from your help docs and past tickets. Your support team handles complex issues while AI resolves repetitive requests. Customer satisfaction stays high without scaling headcount.

📗 Keep documentation current automatically

AI updates code documentation as your product evolves. Tools like Mintlify and Swimm generate and maintain docs so new developers can onboard without hunting down tribal knowledge. Your documentation never goes stale.

Even maintenance benefits from AI. Bug triage gets smarter. Support resolution improves. Technical debt gets identified and prioritized based on actual impact. This stage is less glamorous than launch, but it's where many products spend most of their lives. AI makes it possible to maintain quality without constantly expanding the team.

AI transformation

Three years ago, AI in product development meant experimental tools used by early adopters. Today, it's table stakes. Teams that don't use AI aren't just slower. They're operating with less insight, less agility, and less capacity to iterate.

But the transformation isn't about the tools. It's about what product teams can now accomplish. With AI handling the repetitive analysis, the manual mapping, and the routine coding, product builders can focus on the work that actually requires human judgment: understanding users, making strategic trade-offs, and building experiences that feel genuinely helpful.

The best AI tools don't feel like AI. They feel like your product development process just got smoother.

That's the future: AI that fits naturally into how teams already work, augmenting human decisions rather than replacing them. Product development used to mean choosing between speed and quality, between data and intuition, between what you wanted to build and what you had resources for. AI removes those trade-offs. You move faster without sacrificing insight. You explore more ideas without burning out your team. You make better decisions because you have better information.

The question isn't whether to use AI in product development. It's how quickly you can integrate it into your workflow before your competitors do.