Ok what about if you're tasked with adding AI to your own product? First things first, start with the problem, not the technology. Before you add AI to your product, ask yourself: What user problem will this solve? The best AI features address specific pain points that couldn't be solved well without AI. For example, instead of saying "Let's add AI recommendations," start with the problem of "Users are having a tough time discovering content"

Set realistic expectations

Be honest about what AI can and can't do right now. Many teams get in trouble by promising perfect AI that never makes mistakes. Instead, think about how to design features that are still valuable even when the AI isn't perfect. This might mean showing confidence scores, giving users ways to correct mistakes, or designing the interface to make AI limitations clear.

Consider the full user experience

AI features don't exist in isolation. Think about how they fit into your users' workflow. A great AI feature that takes users out of their flow might actually hurt the overall experience. Think about when and how the AI should appear, how much control users need, and what happens when they want to override or adjust what the AI is doing.

Plan for feedback loops

The best AI features get better over time. Design ways to collect feedback on AI performance from the start. This could be explicit feedback (thumbs up/down buttons) or implicit signals (did the user accept or ignore the AI's suggestion?). Make sure your team has a process for reviewing this feedback and improving the model.

Build trust gradually

Users need time to build trust with AI features. Start with low-risk applications where mistakes won't frustrate users, then gradually expand as you earn their confidence. For example, you might start with AI that suggests actions rather than taking them automatically, then add options for automation as users get comfortable.

Address ethical concerns early

Every AI feature comes with ethical questions. Who benefits from this feature? Could it harm or exclude certain users? How transparent should we be about how it works? These aren't just nice-to-have considerations – they're essential product decisions that affect adoption and trust. Make ethics part of your product discussions from day one.

Remember that you're not trying to add AI everywhere – you're just trying to user problems in the best way possible. Sometimes that means sophisticated AI, sometimes it means a simple non-AI solution, and often it means a thoughtful combination of both. The products that win aren't the ones with the most advanced AI; they're the ones that use AI to deliver experiences that truly help their users succeed.