AI and ML and Bots, Oh My!

Focus on the end, not the means to the end

I was looking over Product Hunt this morning and it seemed like every other product was “X with AI” or a “Bot for Y”. This comment summed things up nicely:

I’ve realized in the tech industry there is a tendency to become really obsessed with the technology itself instead of the problem that needs to be solved. We get caught up in thinking “how can we put AI/ML into our product” or “how can we turn this into a chat bot” instead of starting with the core user problems and identifying the best technologies to solve these problems. In a lot of situations it’s likely not AI/ML or a bot - however we force these into products causing them to be worse: over-engineered and complicated for users.

Dan Grover has a great post titled “Bots won’t replace apps. Better apps will replace apps” where he discusses problems with the recent “bots” hype train. He calls out an example of talking to a chat bot to order a pizza - 73 taps(!) as opposed to opening up a pizza app (only 16 taps). This is an example of focusing on the technology, not the problem.

In contrast, we should aim to build products that focus on the user problem instead of the technology itself. I’m biased as I work on Google Photos but it’s a great example - we have all sorts of amazing ML tech powering parts of our product but do we market it as a “Photo gallery + AI” or call this out to users? Nope. It’s pretty invisible behind the scenes and absolute magic to users when it works.

And that’s the way products should be. We forget that we’re in a world where the user we’re building a product for doesn’t live in Silicon Valley, doesn’t have a CS degree, and hasn’t messed around with TensorFlow for fun over the weekend. Users don’t necessarily care about the means to the end, they just care about the end - having their problems solved in an intuitive and efficient way. Why should they care if you used ML or just a bunch of if statements as long as the product works?

AI/ML is transformative technology that is changing the world. We’re starting to see bots that are actually useful and delightful. I’m personally excited about both of these technologies but I acknowledge that they aren’t perfect for every problem - and that’s ok! We need to focus on the end, and not the means to the end to build good products for users.