6 mindset tips for building an AI product strategy

I’ve been working with ML & AI for 7+ years and have shipped products ranging from expert systems, to statistical models, to LLM-powered features, and convolution-powered recommendation engines.

I’m working on a keynote for the AI for Product Leaders summit and wanted to share some of the content that’s not going to fit. So here’s 6 mental models for building an AI product strategy:

1. Start with Strategic Problems: Don’t do AI for the sake of AI. Identify the most impactful customer problems or business goals where AI can make a difference, and focus there first. Align every AI initiative with clear, measurable outcomes (e.g. increased retention, efficiency, revenue).

2. Leverage What Makes You Unique: Use your proprietary data, user insights, and domain expertise to fuel your AI. This might mean fine-tuning models on your data or crafting AI features that use knowledge only you have. This creates a moat that competitors who use generic AI can’t easily cross​.

3. Reimagine the User Experience: Think beyond bolt-on features. Design AI-first workflows that fundamentally improve how users accomplish tasks. Even if it’s just one part of your product to start, create an experience that makes users say “Wow, that was so much easier (or better) with AI!” Be the product that defines the new standard for your category.

4. Iterate and Learn Fast: Treat AI features as living, learning components. Ship improvements often, use feedback and data to get better, and stay agile with the rapidly evolving AI tech.

Your goal is to learn faster than your competitors — that learning will translate into a better product, quicker. In the AI era, the agile, experiment-driven teams will outpace the rest​.

5. Build Trust and Be Responsible: Differentiate not just on what your AI does, but how it does it. Be transparent, handle data with care, ensure quality and fairness, and keep the user in control.

Make your AI features something your customers can rely on and feel good about using. Trust is hard to earn, but once you have it, it’s a powerful competitive advantage.

6. Empower Your Team: Equip your product and engineering teams with the skills and mindset to work with AI.

Don’t make the mistake of limiting AI product work to a single team, while ML requires some unique skills and workflows to succeed, every PM should be considering LLMs as a tool in their toolkit and fostering a culture that experiments with AI.

When your whole team understands the possibilities and limitations of AI, you’ll execute these strategies much more effectively. We don’t all need to be PhDs, but invest in growing your AI literacy as an organization.

If you’d like to join see the keynote, I’ll also be delivering it as a free Maven lightning talk in March which I’ll share more about here soon — so be sure to subscribe.