AI-Augmented Development: Improve AI assistance for software design, prioritization, and maintenance

Last week I released my guide on hiring AI-Augmented Engineers. Most of my network already knew that they'd need to be thinking about how to do this, which is great to see, and the feedback was super positive. But hiring engineers who operate at that level is only half the problem solved - and if your company isn't set up to get the most out of these people, it has the potential to create even more issues. Most processes - even by typical fast startup standards - were designed for a world where building software was expensive and slow. Sprints, PRDs, long term roadmaps. They exist to prevent the problem of wasted developer time. But that world is gone. An engineer working with AI agents can ship in days what used to take a team weeks, and the bottleneck has completely shifted to product judgment and organisational decision speed - not whether you have the engineering capacity to build. In the new ways of working, startups who are slow to adapt will begin to look painfully corporate by comparison to those that do - and this is going to be a far more challenging period of adaption to go through that simply tweaking who and how you hire. So I wrote a 2nd guide examining the changes winning teams will make - the product-engineering model, conversations about team size and leadership that many are afraid to talk about, and a practical plan to begin moving in the right direction. Admittedly, this one is more predictive than the last - not everything in here is playing out in real time (yet). But I have never been more sure that the companies moving now, through the uncertainty, will be the ones thriving in a year. Those waiting for a more certain, 'safer' path be carved by someone else will regret they didn't begin heading in the right direction sooner. And that right direction can be read in full below.