Marty Cagan’s ‘new standard’ for Product in the age of AI
Yesterday I joined a talk by Marty Cagan on the ‘new standard’ for product management in the age of AI. Marty’s main point was that if you’re not practising proper product management right now – if you’re following the project model, or you’re a feature factory, and you’re not focused on strategy and outcomes – you’re going to have even more problems providing value to users than you do currently.
He said that the product owner role hasn’t been proper product management for a while now, and if you’re just gathering requirements, writing tickets and managing a backlog, your work is distant from the responsibility of achieving outcomes. I agree: in interviews I’m keen to find out if people want me to be a backlog jockey or whether I’ve got power to shape a strategic direction.
He’s written a tome on the product model – the right way of ‘consistently creating technology-powered solutions that deliver value to users and that also drive results for the organisation’. This model is where product managers are responsible for outcomes, not just features. They define the problem, build solutions, and deliver value to customers. “The product manager is really responsible for outcomes.” (And product leaders are responsible for strategy.)
A good point he made was that aspects of product delivery can be automated, which means that there is more time for strategic thinking and discovery. He didn’t overstate how much could be automated though, he was pretty clear that the question isn’t whether people can operate automated tools, the question is whether skilled people can still deliver outcomes with automated tools in the mix. “If you’re trying to build in the product model, it’s not enough to just follow a process. You have to actually think.”
Another good seam running throughout was about the dangers of following a process over truly thinking things through. He said that the principles of Agile are great, but many are simply following a process rather than thinking about how to deliver value to users. “In a lot of big companies, the real problem is process is used as a substitute for thinking.” Similarly he warned against abdicating product thinking or decision-making to AI.
“It’s super easy to write a prompt to create a written narrative [or PRD], which kind of misses the whole point. Because this is literally the definition of thinking. And can you use Gen AI to create a better written narrative? Absolutely. But you don’t want to abdicate your thinking. You want to amplify your thinking.”
What this means is that developing product sense is of the utmost importance. The most important skill for product managers isn’t technical proficiency, it’s product sense – the ability to understand customer needs, build a strategy, and make informed decisions about what to build. “The product manager is bringing deep understanding of the customers, deep understanding of the data, deep understanding of the business.” Organisations that operate under a product model – where product managers are empowered to make decisions and drive outcomes – tend to be more successful. “The best the best product model organisations they believe in pushing decisions down to the teams because they’re closest to the technology and closest to the customers.”
There was a good note on product in the private sector…or in truly transformative parts of the public sector. AI is lowering the barrier to entry, making it harder to differentiate products “because you can essentially copy anybody”. Differentiation now hinges on delivering superior outcomes and possessing strong product sense – the ability to understand what truly matters to users and the organisation. So if you want to beat your competition, develop product sense. If you want to make public services truly better (and that’s what the public want), develop product sense. Just shipping something isn’t good enough.
On the whole, I liked this talk. It didn’t really focus on the tools much, and he wasn’t breathless about being able to automate all and sundry. He was against shoving intelligence layers into everything, backing one of my principles that AI should augment human intelligence, not replace it.
Loads of it wasn’t new. Nothing has really changed, except that it’s easier to do less process, more thinking and more achieving.