Functional lines are not inherently wrong. Functional teams that organize specialized capability well are not rare: they unify standards, accumulate deep expertise, and push one link of the chain to its limit, so the system runs efficiently like well-oiled gears. It has its historical rationale — when the division of labor is deep enough, functional lines do make professional strength run efficiently.

But organizational structure should not be permanently locked in by historical justification. Watching AI and automation tools gradually lower workflow barriers, I have become more and more convinced of this.

Tools Lower Barriers, and Capability Begins to Shift

When AI and automation tools fill in many of the links that once required specialized skill to complete, the shortcomings of any single function become easier to cover. As a result, the capability that matters starts to shift from “pushing one link of the chain to its limit” toward understanding and running an end-to-end business: spotting the problem, organizing resources, completing delivery, and taking responsibility for the outcome.

This is not a declaration that specialization is useless. Expertise remains the foundation of quality and judgment; the change is that expertise must ultimately serve business results rather than become an identity walled off from the business. The most awkward state is when a functional team treats its expertise as a moat — not making judgments for the business, but declaring “you must go through us.”

Signals to Watch

An organization should watch several signals rather than decide by gut feeling whether to change:

  • Whether cross-functional collaboration has become the main cost — with most time spent on handoffs and coordination rather than on actual work;
  • Whether people who understand the full business picture deliver better than pure specialists;
  • Whether tools have already significantly lowered the original skill barriers — capabilities that once took three years to develop can now be picked up in weeks;
  • Whether cost and output pressure demand shorter decision and execution chains.

When the signals hold, functional lines should shift more toward shared capabilities that serve the business, rather than remaining independent endpoints; when they do not hold, forcing a merger will only destroy the professional division of labor that was already working.

This Is an Organizational Judgment, Not a Slogan

To be clear: the judgment in this essay does not apply to every team. “Whether to change the functional lines” is an organizational judgment that depends on the stage of your business, the maturity of your tools, and the current state of your team’s collaboration. There is no universal answer — and precisely because of that, the signals and evidence should be laid out on the table for discussion, rather than dodging the choice with “merge for the sake of merging” or “stay put for the sake of staying put.”

The goal is always one thing: stronger end-to-end execution capability, not what the structure itself looks like.

A Fuller Argument

This short essay is a condensed version of the position paper “When AI Lowers Workflow Barriers, How to Redraw Functional and Business Lines”. The paper covers: a four-quadrant spectrum of organizational forms and the signals to watch, comparisons with Team Topologies and Netflix, the differences between Chinese and American organizational starting points, and a falsifiable pilot protocol. If you are considering whether to change your functional lines, read the paper directly.