What is the BCG 10 20 70 rule?

The BCG 10 20 70 Rule: Why People Still Beat Algorithms in AI Success

When it comes to AI, most companies obsess over the tech—the latest models, the flashiest algorithms. But according to a recent insight from BCG, that’s exactly where they’re going wrong. Enter the “10/20/70” rule: a no-nonsense formula for making AI actually work in real-world business.

Here’s how it breaks down: 10% of your resources should go toward algorithms—the AI models themselves. That’s right, the part most executives lose sleep over? It’s the smallest piece. Then, 20% should fund technology and data infrastructure—clean data pipelines, scalable platforms, the backbone that lets AI function reliably.

But the real kicker? The final 70% must go to people and processes. Not servers. Not PhDs in machine learning. But change management, training, leadership buy-in, and redesigning workflows so that humans and machines actually collaborate. Because no matter how brilliant an AI is, it fails if the team doesn’t trust it, doesn’t know how to use it, or if the organization resists shifting how work gets done.

This rule flips the script on traditional tech rollouts. It’s a reminder that AI isn’t some plug-and-play upgrade—it’s a cultural shift. The most successful implementations aren’t driven by the smartest code, but by the smartest people strategies. Companies that pour energy into communication, upskilling, and clear ownership of AI-driven processes are the ones seeing real ROI.

In a world racing to adopt AI, BCG’s 10/20/70 rule is a sobering counterpoint: technology is the easy part. The human side? That’s where the real work lies—and wins.