Skip to main content

The Biggest Bottleneck Is No Longer Writing Code

Core idea: Decisions, data quality, and business understanding limit progress.

The Bottleneck Has Shifted​

For decades, the bottleneck in software development was clear: can we build it? The limiting factor was engineering capacity — how many developers you had, how fast they could write code, and how quickly they could ship features.

AI has changed that equation. The question is no longer "can we build it?" The question is now:

"Do we have the data, and can we decide fast enough to keep up with what we can build?"

This is what Steven Brovich called the bottleneck shift — one of the four forces reshaping software teams in the agentic era.

The New Bottlenecks​

When implementation speed is no longer the constraint, three new bottlenecks emerge:

1. Decision Velocity​

AI can generate code faster than humans can decide what to build. The limiting factor becomes the quality and speed of your decision-making:

  • Do we have clear requirements?
  • Have we prioritized correctly?
  • Are stakeholders aligned?
  • Can we make decisions fast enough to keep up with generation speed?

2. Data Quality​

AI models are only as good as the data they work with. If your data is incomplete, inconsistent, or inaccessible, no amount of AI acceleration will help:

  • Is your data clean and well-structured?
  • Do you have the right data to train or prompt effectively?
  • Can your data pipeline keep up with AI-driven development?

3. Business Understanding​

AI can generate code, but it cannot understand your business context. The people who can translate business needs into clear specifications become the most valuable members of the team:

  • Do you understand the problem well enough to specify it?
  • Can you evaluate whether the AI's solution actually solves the business problem?
  • Do you have the domain expertise to know when the AI is wrong?

The Expert Multiplier​

Brovich also highlighted the expert multiplier effect (referencing Project Mantle): a group of senior, knowledgeable people with AI at their disposal can achieve an order of magnitude increase in speed.

But this only works if the bottlenecks above are addressed. AI amplifies senior expertise — but it also amplifies poor requirements, bad data, and unclear decisions.

What This Means for You​

  • Invest in decision-making processes. The faster you can decide, the faster you can build.
  • Clean your data. Data quality is now a competitive advantage.
  • Develop business understanding. The people who can bridge business and technology are more valuable than ever.
  • Leadership and planning are competitive advantages. When implementation is cheap, the ability to decide what to build — and why — is what separates successful teams from chaotic ones.
  • AI reduces implementation time. But it does not reduce the time needed to think, plan, and decide. If anything, it makes those activities more important.

Portions of this article are based on insights from Steven Brovich's talk "A leader's guide to advanced team structures in an agentic world" at AWS Events. You can watch the full talk here: A leader's guide to advanced team structures in an agentic world.

Last updated on