For decades, product management has been defined by the management of scarcity. There were always more ideas than engineering hours, more “desirability” questions than research budgets, and more market hypotheses than we had the time or money to test. We built complex bureaucratic structures—sprints, backlogs, steering committees—specifically to gatekeep that expensive, finite resource: delivery. But what happens to your operating model when the cost of execution drops toward zero?
We are currently witnessing the collapse of the “Feasibility Barrier.” In the past, a functional prototype might take a squad two weeks to pull together; today, as noted by leaders at Atlassian, complex interfaces and functional CRMs are being built in an afternoon. When building becomes the easiest part of the cycle, the traditional “Product Trio” of PM, Design, and Engineering faces an existential vibe check. If we aren’t careful, we’ll find ourselves managing a factory that produces high-quality software faster than anyone actually needs it, leading us straight into the “Rot Economy.”
The Death of the Backlog as a Power Center
In the “old” world—the one I inhabited at Nokia and during the early days of easyJet—the backlog was a source of power. It represented the queue of reality. But as AI-assisted development and Small Language Models (SLMs) become embedded in the workflow, the bottleneck is shifting. We are moving from implementation scarcity to curation scarcity.
If an engineer can now produce the output of three people using agentic orchestration, the old ratios (one designer to eight engineers) break down. We are seeing a movement toward what Joca Torres describes as the reorganization of the system around cheaper execution. The risk is no longer “Can we build it?” but “Should we have built it in the first place?” Our processes are often lagging indicators, still worshipping the “sprint demo” when the real value has shifted to the rapid, empathetic discovery of user problems.
Protecting the User from “Agentic Enshittification”
There is a darker side to this efficiency. As the cost of generating features and “agents” falls, corporate bureaucracy tends to fill the vacuum with “Enshittification”—a term popularized by Cory Doctorow to describe how platforms eventually cannibalize user value for shareholder rent. We see this when Salesforce or other giants start imposing “data tolls” or when AI agents are deployed not to help the user, but to exhaust them into a purchase.
As innovation professionals, our job is to act as the “empathy firewall.” We must protect the user experience from this automated rot. Desirability is the only risk that AI cannot solve for us. You can simulate a user interview with a synthetic persona, but it will never tell you the uncomfortable, messy truth that a real human will. When the “how” becomes instant, the “why” becomes the only competitive advantage. Are you using your newly found speed to iterate on value, or just to ship more “shit” faster?
Moving from Copilots to the Silicon Workforce
We are transitioning from “chatting with boxes” to orchestrating a silicon workforce. This isn’t just a tech upgrade; it’s an architectural shift. Companies like NVIDIA and Lenovo are doubling down on the hardware to support this, but the organizational “software”—our teams—is still running on v1.0.
- Agent-Compatible Architecture: Stop trying to drop autonomous agents into workflows designed for human hand-offs and manual sign-offs. It produces friction and “brittleness.”
- The End of “Task” Management: Move your team’s focus from “completing tickets” to “orchestrating outcomes.” If the AI can write the code, the engineer’s role becomes one of Systems Design and Ethical Governance.
- Small Model Sovereignty: Avoid the “Rot Economy” of rising API fees and degrading public models by investing in fine-tuned SLMs. Owning your “intelligence infrastructure” is the new site reliability engineering.
The “honeymoon phase” of AI experimentation is over. We are entering the era of the reckoning: the Great AI Sobriety Test. The winners won’t be the companies with the largest LLMs, but those who redesigned their product teams to be thinner on execution and thicker on empathy. We have been handed a tool that can build almost anything. Now, for the first time in the history of the digital age, we have no excuse for building the wrong thing. It is time to stop hiding behind the backlog and start looking the user in the eye.
Leave a Reply