Catch Your Process up to AI

AI has changed your team’s work, and the process needs to catch up. But how?

As with earlier evolutions in work methods, the first response is often to reach for suggested workflows, practices, rules, ceremonies, standards, or templates.

Instead, I suggest taking a page from the story of agility. Even before AI, leaders realized that minimizing time to value, adapting to change, and creating healthy environments take more than practices. Rather, these effects result from basing behaviours, choices, and actions on certain principles.

This became clear because agile “best practices” morphed over time, while the principles endured:

Early Scrum had 30-day sprints. Most teams don’t do that anymore, yet they still aim to work in short, meaningful cycles.
Daily standups are no longer exclusively about answering the classic three questions, but frequent micro-planning for flow still makes sense.
The technical structure of unit tests, or the testing framework, was never as important as having the tests probe micro-behaviours.
Some organizations don’t want Scrum Masters anymore, but most teams do benefit from someone supporting their teamwork and facilitating their progress.

AI accelerates some parts of the work, but it also creates imbalances. As a result, many practices and coordination mechanisms don’t work well anymore. For example, two-week sprints can be really long, story point estimation is even more fallible, and waiting days for a code review is problematic when code is generated instantly.

Rather than force-fit practices to cope with AI, redraw the process based on principles. Here are some that support both system health and productivity:

Outcome-thinking: your users still want the product to create value for them, not just to do stuff.
Feedback: you still need to quickly learn how your ideas land with others.
Evolutionary design: you still can’t have all the answers up front, even if your AI seems to think it does.
Continuous improvement: today’s best ideas for the product and process may not be tomorrow’s best ideas.
Psychological safety: as long as humans remain in the loop, they should feel safe to engage.

Revisit the process across the system of value delivery, not just in parts of it, to gain real benefit.

The fundamental challenge here for leaders is not in adopting AI tools. It is in redesigning the process so that faster work actually turns into faster value.

Are your executives encouraging (or pressuring) you to increase productivity with AI? Is your team ramping up their adoption of AI, but the process hasn’t changed? If you’re looking for practical and comprehensive guidance, consider my course Reshaping Product Development for AI’s Impacts (for directors/VPs) or Leading AI-Enabled Product Teams (for engineering/product/program managers).