Claude Code Flipped Quarterly Planning on Its Head

Adam Aldred makes a sharp observation about what happens to engineering planning when AI coding tools actually work: the bottleneck moves. Engineering capacity, the thing teams have obsessed over for 15 years, is no longer the constraint. QA and stakeholder alignment are.

The Old Question Is Now the Wrong Question

For the past 15 years, engineering organizations have revolved around capacity math. How many sprints in a quarter. How many points per sprint. How many engineers on staff. Can we get more engineers. Every quarterly planning cycle started and ended with that arithmetic.

Aldred describes an older executive coming to him with exactly these questions -- how many points for the quarter, do we have enough engineering resourcing -- and his reaction was blunt: that is the wrong question to be asking right now.

Where the Bottleneck Actually Lives

With Claude Code and similar AI-assisted development tools, the coding itself is no longer the slow part. Two things are:

End-to-end QA testing. The features get built faster than they can be validated. QA becomes the rate limiter.

Stakeholder alignment. Scheduling meetings with humans, getting product managers and stakeholders on the same page about what should get built, and reaching consensus on priorities -- all of that now takes longer than writing the code.

The implication is significant. If you are still planning quarters around engineering capacity, you are optimizing for a constraint that has already loosened. The new constraint is human coordination.

Key Takeaways

• AI coding tools have shifted the bottleneck from engineering capacity to QA and stakeholder alignment
• Quarterly planning questions about sprint points and engineer headcount are increasingly outdated
• The hardest part of shipping software is now deciding what to build and validating it, not building it
• Teams that adapt their planning process to this new reality will move faster than those still doing capacity math