Back at the start of product time, a lot of upfront work happened. Research, PRDs, user interviews, annual plans. The goal was to make sure we built the right product, the thing that customers would buy, or even the thing we could get done in time. Building was expensive.
The cost assumptions were different. Lean was invented because building full production ready solutions was still expensive. You had to make sure you knew what you were building before you started, because it cost so much to build it. This is part of the goal of agile, to reduce upfront cost burden, get started and minimise the risk of high sunk costs by being able to pivot mid-project.
This is even worse for things like movies, or games, where building costs $200m+. That’s not something you take a risk on. This is why we are drowning in sequels, remakes and sure things™ guaranteed by brand loyalty. In regular software we can make things faster so there is more creativity.
AI and other productivity tools have evolved. Cloud has reduced capex cost, and tools like copilot, replit, and cursor are cutting chunks off the time to prototype, maybe even to product. You can create a simple app in a matter of a few hours (if you know what you’re doing). Creation keeps getting cheaper, and the economies for scale continue to grow for every manufacturing process, from plastic widgets to software systems. So what will this mean for product people?
Product is not about de-risking development cost anymore. Product is about de-risking attention deficit. We can now churn out so much product so fast, that the waste is not our engineering effort, but our customers’ effort to consume what we produce. This means product really comes into its own. We have to focus on generating customer value even more so than in the pre-AI era.
The AI economy is about managing attention (even more than before).
AI also competes with solutions selling. SaaS needs to think. Why would I compromise on someone else’s form-and-database application when I can have an AI build a custom solution just for me in a few hours with Lovable or Base44? So many platform vendors start with a base product that relies on professional services to drive a solution for a few hundred thousand dollars. Why do we need this when a business user can spin up a perfectly customised react web app and scalable backend in a few hours with Claude? Sure, you get a lot of process and documentation with the consultancy, but how often does anyone really use that 60 page engagement report?
In this world, good product and process taste, and some means to choose and control the tide of options vying for attention, becomes the new mission of the product person. We’re not about avoiding going down the wrong path, we’re about making sure we close the irrelevant paths as fast as possible before everyone else is lured down them.