Strategy used to be mostly handed down from the top, set in advance and passed down for others to carry out. AI is now handling more of the execution layer, which means the decisions underneath every plan — at every level, in every function — matter more than they used to. Strategy isn't only a specialist function anymore. It's everyone's job now, whether the title says so or not.

The idea that frugality should be built into architecture from the start — not bolted on when the cloud bill arrives — is gaining ground. Vogels, CTO of Amazon, made a similar point from a cost angle. Frugal Architect by Design takes it further: efficiency as a design instinct, not a budget response. Systems built this way tend to be leaner, more resilient, and more sustainable — not by intention, but by consequence.

As AI agents move from concept to deployment, it is worth stepping back and thinking carefully about what you are actually trying to solve. Automation and agents have different strengths, different requirements and different organisational implications — and today the choice between them is a real and consequential one. In time, as agent technology matures and organisations build the foundations to support it, the balance may shift significantly. But right now the organisations that navigate this well will be the ones that started with the problem and worked toward the right technology — not the other way around.

The exception is where the real work begins.

From data to direction.