A product team saved $150 million in margin improvement not by building new features, but by decommissioning a long tail of customized, on-prem legacy products. This "unsexy" work eliminated significant operational drain from support and maintenance, directly impacting the bottom line in a way new features rarely can.

Instead of a full rewrite, identify the specific pain points of a legacy system (e.g., a command-line UX) and solve them with minimal development. This delivers immediate value, reduces risk, and validates the market need for a larger investment later, preventing a costly failure.

The initial step in modernizing is not to rebuild, but to understand. AI can ingest source code, user manuals, and even screen recordings to map existing processes and identify optimization opportunities, ensuring the new system improves upon the old rather than just replicating it.

In a complex legacy environment, internal motivations like improving developer experience or modernizing technology often fail to gain traction. The initiatives that successfully navigate the process are those that can clearly articulate and deliver tangible value to the end customer.

Government procurement processes are rooted in a pre-digital, paper-based mental model. They treat software like a physical commodity that must be procured anew for each jurisdiction, preventing them from leveraging software's inherent scalability and leading to massive, redundant development costs.

In government, digital services are often viewed as IT projects delivered by contractors. A CPO's primary challenge is instilling a culture of product thinking: focusing on customer value, business outcomes, user research, and KPIs, often starting from a point of zero.

Even if legacy code is stable and functional, it should be replaced when the user experience it provides becomes obsolete. When user expectations (e.g., mobile access, modern UI) have fundamentally shifted, the old system becomes a liability regardless of its technical stability.

Don't just plug AI into your current processes, as this often creates more complexity and inefficiency. The correct approach is to discard existing workflows and redesign them from the ground up, based on the new paradigms AI introduces, like skipping a product requirements document entirely.

To transition to AI, leaders must ruthlessly dismantle parts of their existing, money-making codebase that are not competitively differentiating or slow down AI development. This requires overcoming the team's justifiable pride and emotional attachment to legacy systems they built.