When developing internal AI tools, adopt a 'fail fast' mantra. Many use cases fail not because the idea is bad, but because the underlying models aren't yet capable. It's critical to regularly revisit these failed projects, as rapid advancements in AI can quickly make a previously unfeasible idea viable.

When AI tools are not adopted, leadership often blames resistance and prescribes more training. The real issue is typically a structural failure, such as not involving local teams in the model's design or misaligned incentives between insight generators and decision-makers.

AI tools accelerate development but don't improve judgment, creating a risk of building solutions for the wrong problems more quickly. Premortems become more critical to combat this 'false confidence of faster output' and force the shift from 'can we build it?' to 'should we build it?'.

Product leaders must personally engage with AI development. Direct experience reveals unique, non-human failure modes. Unlike a human developer who learns from mistakes, an AI can cheerfully and repeatedly make the same error—a critical insight for managing AI projects and team workflow.

While AI dramatically increases development speed, it's a double-edged sword. Without a solid product foundation, user understanding, and clear principles, teams will simply accelerate the shipment of low-value features. AI amplifies both good and bad practices.

Without a strong foundation in customer problem definition, AI tools simply accelerate bad practices. Teams that habitually jump to solutions without a clear "why" will find themselves building rudderless products at an even faster pace. AI makes foundational product discipline more critical, not less.

The 85% AI project failure rate isn't a technology problem. It stems from four business and process issues: failing to identify a narrow use case, using data that isn't clean or ready, not defining success and risk, and applying deterministic Agile methods to probabilistic AI development.

AI tools, likened to "1,000 interns," require explicit instructions to be effective. This new reality of one-day sprints quickly reveals which product managers have a clear vision and which do not, as ambiguity leads directly to poor development results and exposes a core skill gap.

When an AI-coded feature is flawed, the instinct is to patch the specific output. A more effective, long-term approach is to analyze *why* your agent system produced a bad result and improve the underlying agent, skill, or process that failed.