Every engineering leader investing in AI is asking a version of the same question: if AI writes code this fast, why isn't software reaching customers faster too? James Jarvis, CEO of ClearRoute, addressed this gap directly in a recent interview with TalkDev Bureau, drawing on findings from the State of the Route to Live report.
James mentions in his interview where leaders should focus next.
James argues that writing code was never the hardest part of enterprise software delivery. The harder work starts once the code exists: testing, security, compliance, governance, and release. His team's research uncovered approval and review cycles stretching from days into multiple weeks, propped up by manual coordination rather than automation. AI has accelerated the smaller, more visible slice of the lifecycle, while the larger portion of the work still moves at its old pace.
For leadership, this isn't a technical footnote. Each hold-up in that larger portion shows up downstream as a delayed launch, a slower response to the market, and a business that adapts more slowly than it could.
James positions platform engineering as the structural layer that lets AI operate safely inside complex enterprise systems. Consistent environments, built-in automated checks, and reusable building blocks give teams room to move quickly while staying within guardrails. He frames this as a mindset shift: leaders should stop evaluating individual AI tools and start building an operating model designed for continuous, well-governed change.
One of James' sharpest points was: AI doesn't level the playing field between organisations. It amplifies whatever is already there.
Companies with disciplined engineering practices and mature automation see genuine gains because AI slots into systems already built for speed. Companies working with disconnected tools or manual sign-offs end up creating more volume that still can't move through the delivery pipeline efficiently. Strong foundations make AI worth more. Weak foundations make its limitations more visible.
• Code speed isn't delivery speed. The 80% Problem points to where the real bottleneck lives, and it isn't in engineering output.
• Platform engineering is now a business capability, not just a technical one. It's what gives AI room to scale without losing control.
• AI exposes organisational maturity. The distance between strong and weak performers grows, rather than shrinks, as AI adoption increases.
• Readiness is about governance, not tooling. James is clear that AI-readiness depends on platform-level controls for identity, access, and oversight, not another assistant added to the stack.
• Operationalisation is the next differentiator. The organisations that shorten the distance between deciding on a change and delivering it will out-adapt and out-respond their competitors, no matter how much AI they've adopted.
James' message to leadership is direct: the companies that win this next phase won't be the ones producing the most code. They'll be the ones that have rebuilt their delivery systems enough to turn AI into steady, measurable business results.
Explore the full State of the Route to Live report for the complete data, benchmarks, and analysis behind these insights.
Here is the full interview with James Jarvis for more context on where enterprise AI adoption is headed.
Every engineering leader investing in AI is asking a version of the same question: if AI writes code this fast, why isn't software reaching customers faster too? James Jarvis, CEO of ClearRoute, addressed this gap directly in a recent interview with TalkDev Bureau, drawing on findings from the State of the Route to Live report.
James mentions in his interview where leaders should focus next.
James argues that writing code was never the hardest part of enterprise software delivery. The harder work starts once the code exists: testing, security, compliance, governance, and release. His team's research uncovered approval and review cycles stretching from days into multiple weeks, propped up by manual coordination rather than automation. AI has accelerated the smaller, more visible slice of the lifecycle, while the larger portion of the work still moves at its old pace.
For leadership, this isn't a technical footnote. Each hold-up in that larger portion shows up downstream as a delayed launch, a slower response to the market, and a business that adapts more slowly than it could.
James positions platform engineering as the structural layer that lets AI operate safely inside complex enterprise systems. Consistent environments, built-in automated checks, and reusable building blocks give teams room to move quickly while staying within guardrails. He frames this as a mindset shift: leaders should stop evaluating individual AI tools and start building an operating model designed for continuous, well-governed change.
One of James' sharpest points was: AI doesn't level the playing field between organisations. It amplifies whatever is already there.
Companies with disciplined engineering practices and mature automation see genuine gains because AI slots into systems already built for speed. Companies working with disconnected tools or manual sign-offs end up creating more volume that still can't move through the delivery pipeline efficiently. Strong foundations make AI worth more. Weak foundations make its limitations more visible.
• Code speed isn't delivery speed. The 80% Problem points to where the real bottleneck lives, and it isn't in engineering output.
• Platform engineering is now a business capability, not just a technical one. It's what gives AI room to scale without losing control.
• AI exposes organisational maturity. The distance between strong and weak performers grows, rather than shrinks, as AI adoption increases.
• Readiness is about governance, not tooling. James is clear that AI-readiness depends on platform-level controls for identity, access, and oversight, not another assistant added to the stack.
• Operationalisation is the next differentiator. The organisations that shorten the distance between deciding on a change and delivering it will out-adapt and out-respond their competitors, no matter how much AI they've adopted.
James' message to leadership is direct: the companies that win this next phase won't be the ones producing the most code. They'll be the ones that have rebuilt their delivery systems enough to turn AI into steady, measurable business results.
Explore the full State of the Route to Live report for the complete data, benchmarks, and analysis behind these insights.
Here is the full interview with James Jarvis for more context on where enterprise AI adoption is headed.