What Is the 10-20-70 Rule for AI?

The 10-20-70 rule for AI is a framework for successful AI implementation in organizations. It states that 10% of effort should go toward building AI algorithms, 20% toward the data and technology infrastructure, and 70% toward business processes and organizational change.

This rule was popularized by McKinsey and other consulting firms to counter the common mistake of focusing too heavily on the technology itself while neglecting the human and process elements that determine whether AI projects succeed or fail.

The 10% for algorithms covers the actual AI models and machine learning code. In 2026, with pre-trained models and APIs widely available, this is the easiest part. You do not need to build models from scratch in most cases. The 20% for data and tech infrastructure includes data pipelines, cloud computing, storage, security, and integration with existing systems. Clean, organized data is critical because AI outputs are only as good as the data they learn from.

The 70% for processes and people is where most AI projects fail or succeed. This includes training employees to use AI tools, redesigning workflows, managing change resistance, setting quality standards for AI outputs, and measuring business impact.

A 2025 BCG study confirmed this ratio, finding that companies allocating at least 60% of their AI budget to people and processes were 3x more likely to report positive ROI from AI investments.

Practical takeaway: If you are implementing AI in your business or workflow, spend most of your time on training, process design, and change management, not on picking the perfect tool.