Only 12% of Companies Generate Value From AI. Here's What They're Doing | Sanjeev Vohra, Genpact

What It Covers

Sanjeev Vohra, Chief Digital Officer at Genpact, shares findings from a 500-executive study revealing only 12% of companies generate measurable AI value. He outlines the three core barriers blocking enterprise AI adoption and explains what separates leaders from the 47% stuck in assisted, non-production AI use.

Key Questions Answered

• •AI Maturity Distribution: Only 12% of companies qualify as AI leaders — generating tangible business value from AI in production workflows at scale. Another 15% are classified as advanced, 47% as assisted (using copilots but not embedding AI into operations), and 26% as emerging, meaning they have yet to establish a clear adoption path.
• •The Frozen Middle Problem: Senior leadership and junior employees are engaging with AI, but middle management — operationally overloaded and managing direct reports — represents the critical bottleneck. Companies must design specific rotation and training mechanisms targeting this layer, as ideas and execution both flow through mid-level managers who control day-to-day process workflows.
• •Governance Before Scale: Enterprises deploying multiple agents need a centralized platform that inventories all agents, enforces responsible AI and data privacy policies, monitors compute costs, and flags rogue behavior. Without this infrastructure, companies risk uncontrolled agent sprawl — analogous to past shadow IT problems — with no visibility into which agents are underperforming or operating outside policy boundaries.
• •Progress Over Perfection Framework: AI adoption cannot follow a fixed implementation playbook like an ERP rollout. Because the technology evolves faster than any organization can fully map it, waiting for a complete roadmap guarantees falling behind. Companies should set outcome-driven targets — Genpact's own goal is reducing G&A costs by 50% — then build programs downward from that specific, measurable commitment.
• •10x Engineers, 3x Practitioners: AI tools are enabling engineers to become roughly 10 times more productive by coding in natural language rather than traditional syntax. Business-side professionals — accountants, lawyers, procurement staff — can achieve approximately three times their current output. Scaling this requires systematic company-wide enablement programs, not isolated pockets of high performers, to bring entire workforces to a consistent capability level.