Companies of all shapes and sizes descended on Mill Valley, California this week to discuss and learn from each other about using artificial intelligence at VentureBeat’s inaugural Transform 2018 conference. Everyone had a story to tell, pitches to make, and insights to share. While the general consensus was positive — everyone wants to say they are using cutting-edge and transformative technology — there was also some cautionary language, if you listened carefully.

Of course, AI isn’t a one-size-fits-all strategy and everyone has their own definition for the term. Some see it as a broad umbrella that encompasses everything from simple machine learning to robot consciousness, others rely on characterizations specific to their industry, and everyone else just uses it to generate hype.

That’s why it was refreshing to hear how AI solutions aren’t necessarily the best, or simply aren’t always the most cost-effective. If you’re not taking ROI into account when jumping on the AI bandwagon, you’ve already lost.

There was certainly a ton of hype at Transform. But there were also meaningful results, whether you agree that AI is the correct descriptor or not.

Do your research. Figure out what your peers are doing. Be wary and listen carefully — AI can sink your business just as quickly as it can transform it.

Maybe you can leverage something as simple as a common algorithm, but that doesn’t necessarily mean you need a dedicated department for algorithms. Maybe you aren’t developing something as complex as self-driving cars, but you can still use computer vision for your onboarding process.

It’s OK if your company isn’t going all-in on AI. It’s not OK if your company is blindly adopting AI.

Whether it’s AI or social or mobile or whatever the latest trend, just make sure your teams are doing their due diligence.

ProBeat is a column in which Emil rants about whatever crosses him that week.