The Constraint Didn't Go Away. It Moved.

You cannot scale delivery without understanding your constraint. AI agents create the feeling of unlimited capacity. There is always another task to start. It feels like progress, even when the system is not actually moving faster.

In The Goal, Eliyahu M. Goldratt’s point is simple: every system has a constraint, and that constraint determines output. Improving anything else does not increase throughput.

If you are only using AI for coding, you may be adding capacity where the constraint is not. If coding was the bottleneck, speeding it up can shift the limit somewhere else — so more work enters the system, more starts, and more piles up in front of whatever is actually limiting progress. Delivery may not speed up. It can slow as work backs up at the constraint.

Most teams started by applying AI to code and are now (or still) feeling the bottleneck elsewhere. It makes sense that they are looking to apply AI across the whole SDLC. Rather than taking it all on at once, start with the biggest constraint and move from there.

Originally published on Substack.