Innovation comes with a risk of failure because, by definition, it requires us to try things that haven’t been done before. That means risk is part of the job. But as product leaders, we’re also accountable for ROI, customer satisfaction, and business outcomes. So the question becomes: How do you innovate in a way that’s both responsible and effective - and maybe even fun?

I’ve found it helpful to use both inside-out and outside-in styles of thinking.

Inside-out thinking: Feed your curiosity engine

Inside-out thinking is all about discovering what’s possible. For me, that means carving out dedicated time to learn—whether that’s testing new tools, experimenting with AI models and prompts, and getting hands-on with emerging technology.

I had JPMorgan’s Payments executive director, Peter Bailey, on my podcast Hard Calls, and he shared something that stuck with me: He blocks 10–15 hours a week just for learning—podcasts, courses, side projects. It helps him stay energized and curious about what’s next, instead of feeling overwhelmed or threatened by it.

Outside-in thinking: Obsess over the customer problem

Outside-in thinking starts with the question: What’s the customer problem I’m trying to solve? Is the solution a pain killer or a vitamin, is this something customers NEED or want?

No one buys a feature for the feature’s sake—they’re buying a result. A better way to work. A simpler process. A faster outcome. Value.

This mindset is the antidote to the feature factory. It’s how we avoid building something no one needs. It forces me and my team to get specific about value, not just functionality.

Whether we use the working backwards framework, or JTBD framework, it is critical that we focus on customer value and driving outcomes. AI for AI does not provide ROI or customer value. This is not innovation, even if it it’s cutting edge.

The magic happens at the intersection

When inside-out and outside-in thinking meet, that’s where the most powerful innovation happens.

Once I have an idea I believe in,I share it early by prototyping and I bring customers in through a modern discovery process: using experimentation, feedback loops and betas. I want to find the issues early and continue to pivot until we have product market fit and adoption.

This earns trust. It sharpens our ideas. And more often than not, it gets us to value faster.

Try this

Here are four ways to put this thinking into action:

  1. Block inside-out time. Add learning time to your calendar—every week. Treat learning like a measurable KPI or outcome. Make sure this is hands-on work, not just training.
  2. Write the outside-in problem. Get yourself into the POV of the customer and write out the problems they face. If you can’t finish “Our users need to ___ so they can ___,” you’re not ready to build. If you can’t quantify the outcome for your buyers and users, you should not build it
  3. Share your idea early. Use generative AI to prototype. Want help? My colleague, Dave Killeen, talks about this in an episode of The Vibe PM podcast.
  4. Measure what matters. Look at agent and feature adoption and retention, not just delivery. Who’s using the new AI flow? Where’s the friction? Are outcomes improving? Tools like Pendo’s Agent Analytics help connect usage with results.

Keep yourself (and your team) accountable with these questions

  • Inside-out: What did we learn this week that changes how we’ll build?
  • Outside-in: What outcome will the customer see, and how will we prove it?
  • Intersection: Which customers are co-piloting with us, and what have we changed because of them?

If you’re trying to balance the pace of AI innovation with the responsibility of delivering meaningful customer outcomes, this mindset helps you move fast—but with intention.

Want more stories like this? Check out the latest episodes of the Hard Calls podcast to hear from product leaders who’ve learned by doing (and failing, and trying again).