Most organisations think AI will transform how they build software.
I’m starting to think it will expose how they think.
We used to scale by increasing execution capacity. More capability = more output. That model is changing.
I’ve been looking closely at specification-driven design (SDD). And there’s a shift hiding inside it:
👉 The quality of the specification now determines the quality of the outcome.
Not the developer.
Not the tool.
Not even the AI.
Specification quality = output quality.
Most organisations struggle here:
jump straight to tools (“let’s use AI”)
skip clarity
write vague specifications
get average outputs
blame the technology
But the constraint isn’t the tooling.
👉 It’s the thinking.
If you break it down, there are only three things that matter:
Build the machine
Run the machine
Feed the machine properly
This doesn’t reduce the value of people.
It shifts where value sits.
👉 From execution --> definition.
Engineers won’t just build. They’ll need to:
frame problems
define specifications
shape outcomes
👉 The leverage moves upstream.
This is the part I find interesting.
Because it’s not new.
It’s the same muscle:
founders clarifying strategy
sales teams articulating value
leaders defining priorities
Just applied in a different context:
👉 “How do we define work so clearly that a machine can execute it?”
In the past, we scaled execution.
Now, we scale the quality of our thinking
Because that’s what drives the machine.
💬 Where are you seeing this show up. Strong outputs vs disappointing ones?
Written by Jonathan Stern
ICF Certified Coach | Gallup CliftonStrengths Certified | Former MuleSoft ANZ Leader
I coach high-potential leaders and high-potential scale-ups.
Schedule a conversation