Most teams are still stuck in old metaphors when it comes to product delivery. In the age of AI, we need an update.
The construction era
Modern product delivery still relies on a deeply ingrained metaphor: construction.
Product managers & designers draft plans. Engineers build to spec. Delivery managers ensure things run smoothly & deadlines are met.
This works with traditional code because it's deterministic. We architect solutions & then use code to bring them to life brick by brick.
The gardening era?
Generative AI doesn't work like this. It doesn't allow us to lay bricks faster - it does away with brick-laying altogether.
A more apt metaphor may be gardening. We plant seeds and provide conditions for growth, but ultimately do not control what emerges from the soil.
Our role, like a gardener's, is to constantly observe & evaluate. We prune, redirect, replant (prompt, tune, add tools) - getting closer as we do to whatever it is that we set out to achieve.
So what?
The implications for delivery are significant:
1. Defining & communicating the "why" becomes more important than ever
Teams cannot mindlessly execute on preordained plans, they must critically evaluate progress against a clear, shared understanding of their objective.
2. Gardening isn't done in sprints or cycles, it's constant
The build phase with generative AI is constant & (near) instant. PMs, designers, engineers & data scientists must work together to react to model outputs in real-time.
3. Success is defined at a systems-level, not a feature-one
Our goal isn't to create a good output, it's to teach a system to reliably create good outputs. We must understand how a system behaves, how our inputs influence outputs, and when & where we need to intervene.
Generative AI is a shift from constructing features to cultivating systems. The teams that are willing to revisit their approaches from first principles are the ones that will boast the most impressive gardens.