AI enables faster product development, but introduces new bottlenecks. You’ve heard the reports about how Anthropic writes 90% of their code with Claude, but then there’s the footnote where they mention human review is creating new slowdowns. Every company is experiencing the same thing at the same time. Yes, AI can write code faster, but no, it cannot complete the picture.
As a co product-manager for a company with 6,000 customers and 500,000 users, there are new challenges today which did not exist before. Primarily, moving faster enables more code to be written, but it also means you have to make big decisions much more often. For example, yesterday we introduced our template system for forms which I’m wonderfully proud of. It’s something we always wanted to do for quizzes, but back in 2018 it was impossible to actually build a scalable system.
This time around we can, so that system can handle 10’s of thousands of template for forms and surveys. But, the key decisions made now cannot be undone, and instead of having months to think about it, my co-founder and I had to decide on the fly yesterday. Luckily, having done this before, we knew what to do. The decision we had to make was around categorization of those templates. Because once they are set they become permanent URLs, and I don’t care what anyone says, if you move a ton of pages around later, Google and AI crawlers take a very long time to pick up the new addresses, and you will lose traffic.
So it was a high-stakes decision, and quite honestly the third one of the day. These used to be spread out over long periods of time because development took a long time. But now they happen daily, if not multiple times a day. AI can give you an answer, but it’s the same answer it gives everyone else. This brings you closer and closer to the mean, and gives no one any reason to choose your option over everyone else. This is honestly the biggest risk of AI product management. Letting Claude make every decision means you move closer to what everyone else is doing, and begin to lose your key advantages.
So what allowed us to make the choice we did? Which I believe sets us up to both follow best practices standards, and have a point of view in our product approach that aligns with our customer base? Intuition. I wish there was a better answer, but it truly comes down to a combination of knowing the customer for many years (we were founded in 2014), and also knowing what people are used to (not just your direct competitors, but also what other tools people use and are familiar with), and also knowing what AI is bringing to the table as people spend more and more time with chatbots.
Through all of that, you weave your own pathway, and those are the tasteful choices which give you a differentiator you can stand on.
