NTT DATA, Inc. President & CEO Abhijit Dubey From Semafor World Economy

Why It Matters

Understanding the full cost structure prevents under‑budgeting and ensures AI initiatives deliver measurable business value, a critical insight for CEOs and CFOs planning digital transformation.

Key Takeaways

• AI deployment costs span four distinct investment layers.
• $1 covers agent development; $2 funds employee change management.
• $3 finances ecosystem integration, security, and orchestration efforts.
• $4 is needed for data preparation and readiness.
• Over‑focus on technology ignores critical change, ecosystem, and data costs.

Summary

In a recent Semafor World Economy interview, NTT DATA President and CEO Abhijit Dubey outlined a four‑step framework—dubbed the “rule of 1‑2‑3‑4”—that companies must consider when extracting economic value from enterprise AI.

Dubey broke down the total cost of ownership into four buckets. The first dollar funds the creation of AI agents and core models. The second dollar is spent on change‑management initiatives that bring employees up to speed. The third dollar builds the surrounding ecosystem—inter‑agent communication, orchestration, security guardrails, and token optimization. The fourth dollar prepares the data pipeline, cleaning and structuring information so agents can operate effectively.

“While we get too fixated on the one, sometimes a lot of times we forget the two, three, and the four,” Dubey warned, emphasizing that neglecting any of these layers can erode ROI. He cited early‑stage pilots that stalled because firms under‑invested in data readiness or ignored the need for robust governance.

The framework signals that AI budgets must expand beyond pure model development. Executives who allocate resources across all four pillars are more likely to achieve scalable, secure, and financially sustainable AI deployments, reshaping competitive dynamics across industries.

Original Description

NTT DATA, Inc. President & CEO Abhijit Dubey from Semafor World Economy answering the question "What is the most surprising impact of the AI revolution, and how will that impact redefine how companies compete?": “The real cost of AI isn’t just building it. I call it the 1-2-3-4 rule: $1 to create the tech, $2 to drive adoption, $3 to build the ecosystem, and $4 to prepare the data. Too often, we fixate on the $1 — and forget the rest is where the real value (and effort) lies.”