Cycle time, throughput and aging WIP in an age of agentic coding

[#1]

Hey folks, hope you are well.

I'd love to get some perspective and experience shares around how you are approaching metrics for agile teams with the rise of agentic coding. My sense is that as individuals in teams leverage these new technologies we'll see a shift in the metrics and interpreting flow for a team will evolve.

For example:

• Cycle time: expect this to decrease as team members code and review code faster for a user story
• Throughput: expect this to increase as team members will pick up more tasks throughout the week, perhaps even running multiple streams of work in parallel (so much for minimising context switching!)
• aging WIP: expect this to decrease as work doesn't hang around in the system for a long time and planning horizons shrink

What are you seeing?

Thanks,
Nick Muldoon
Insights by Easy Agile

[Nigel Budd]

Hey Nick,  

Terrific question, I'm seeing a lot of my customers actively tracking cycle time metrics, even without the introduction of agentic AI, it seems to be a hot topic.

The expectation is, that those metrics you mention will improve, and from the teams I've seen introducing AgAI have seen reductions in cycle time, reduction in WIP might not neccessarily improve as teams try to do more at the same time.

Of course cycle time is a really useful Agile metric, how quickly can we deliver stuff to our customers, but I always like to balance that with a qualitative metric as well.  There's no point going faster if our quality drops, or we then spend all the team's time fixing bugs, or fixing security holes, for one thing there's not an actual cost saving, the other it could negatively effect the company'r reputation or customer trust.  

[Claudia Love]

Hi Nick, what you ask here resonates a lot as IT engineers and even managers leverage AI to answer questions and solve problems the amount of work increases everywhere!

AI is improving ideation and execution speed, but it’s exposing flow constraints. The real signal isn’t just faster metrics, it’s whether work is finishing cleanly end-to-end and the handshakes between development, deployments and reviews are keeping up (places where its harder, and more technical to insert automation).

Cycle time may flatten until the whole system is optimized, even while throughput and work in progress increases even without measurable value delivery.

Instead we might want to focus on flow efficiency; PRs from open to close, merge and deploy time (what if there is tech debt?!?!), and how efficiently we get through rounds of testing and defect management on the way.