Need AI to deliver faster? Maybe there’s a better way.
by Tony Fifoot, 3 mins reading time
When organisations want to “Go Faster” (with the actual goal of creating more value), the usual playbook (for software products) looks something like this:
Write code faster (AI tools will save us)
Plan better, keep people busy
More efficient tasks
It feels logical on the surface, coding is how Devs generate value - so we want them to be more efficient, faster, and keep them busy. As a manager, it’s natural to assume this is where we should focus our energy.
But here’s the trap: coding is rarely the slowest part of the system.
The hidden reality: Flow Efficiency
Across most product development environments (and most knowledge work), flow efficiency is surprisingly low - often less than 10%. That means for every 100 days a piece of work spends “in the system”:
Only about 10 days are spent actively adding value (designing, coding, testing, etc.).
The other 90 days are waiting - for alignment, for approvals, for priorities to be clarified, for another team to finish something, for dependencies to line up.
Yet organisations pour most of their energy trying to shave small increments from that 10%.
Why improving coding speed barely moves the needle
Let’s do a back-of-the-envelope calculation.
Say your team operates at 10% flow efficiency. After a lot of great management on your part, you improve developer productivity by 20% i.e.: they get all their coding done in four days instead of five - so you can give them Friday off ;)
So, what happens?
The 10 days of work in the example above, now only takes 8 - great work!
The 100 day delivery now only takes 98 days.
The total system improves by just 2%.
People are so much more efficient, but you’ve made a marginal dent in the overall timeline.
Alternatively, reducing waiting time, even modestly, can yield outsized benefits. A 20% improvement in that 90% of waiting translates into an 18% improvement overall.
Where you choose to focus makes an enormous difference.
Why we keep focusing on the wrong thing
If the math is so clear, why do so many organisations keep zeroing in on coding speed? Three reasons stand out:
It’s visible: Developers coding is the easiest part to observe. The waiting? That’s often hidden in emails, in queues, meetings, or invisible blockers.
We need to take action: Buying new tools, introducing AI copilots, or pushing people to be more efficient creates a quick sense of “we’re doing something.” Tackling systemic causes of waiting feels harder and less tangible.
Traditional management thinking: For decades, industries have improved efficiency by focusing on “the workers” - getting value for money. But in knowledge work, the system, not the individuals, is usually the constraint.
Where the real leverage is
If you want to make a dent in your overall delivery speed, shift your attention from coding (build speed) to the overall system of work. That is:
Clarify goals and priorities so teams aren’t stuck waiting for direction.
Reduce handoffs by reorganising teams around end-to-end value.
Make work visible so people can see where it’s stuck and unblock it.
Reduce batch size so work isn’t delayed by the slowest part.
Move decisions to where the work happens so work doesn’t sit waiting.
Improve feedback loops so issues are spotted and fixed quickly.
and lots more…
These systemic changes shorten the 90% of time that gets wasted. That’s where you get the most juice for the squeeze - fixing major bottlenecks can have a huge impact on the overall system - and isn’t about pushing people to work harder.
A shift in mindset
It’s time to challenge trad management practices, to look at developers and think, “If only they could code faster.” The real constraint is almost never “writing code”, it’s the waiting built into the system.
So the question to ask is: How do we make the system flow?
👉 Focus on the 90%, not the 10%. That’s how you accelerate speed to market.
We help lots of companies with this? Let’s chat.