Enterprises Face $18T in Untapped AI Value

The report, published jointly by Genpact and HFS Research, estimates the largest 2,000 public companies hold nearly $18 trillion of untapped value tied to artificial intelligence, attributing the unrealised potential to what the report terms "enterprise debt." The study defines enterprise debt as a mix of outdated technology, poor-quality data, inefficient processes and workforce readiness gaps, and states these factors are interconnected.

HFS and Genpact report they calculated the aggregate figure by applying respondent-reported revenue uplift and cost-reduction estimates across the combined revenue base of the Global 2000. The study states organisations that successfully address these debts could see about 8% faster annual revenue growth and 16% lower annual costs. The report includes the phrase "agentic AI trapped in pilot purgatory" to describe stalled deployments. Phil Fersht, founder and CEO of HFS Research, is quoted: "AI is exposing every weakness enterprises have spent decades learning to live with."

Companies confronting comparable stacks of legacy systems, fragmented data, and process inefficiencies typically face long lead times to operationalise machine learning models. Data quality and process standardisation are common blockers for moving from proofs of concept to production ML pipelines. Observers in the enterprise AI space often note that resolving cross-functional dependencies between data engineering, platform teams and line-of-business owners is a prerequisite for reliable model deployment at scale.

The study quantifies a large business opportunity and reframes familiar adoption barriers as a single, interlocking concept: enterprise debt. For practitioners, that framing highlights why isolated investments in models or tooling frequently underdeliver when upstream data hygiene, orchestration and staffing readiness are weak. Industry reporting positions this work alongside other analyses that emphasise operational maturity, not just model performance, as the gating factor for AI ROI.

Signals that observers and practitioners can follow include changes in enterprise budgeting toward data platform remediation, uptake of production ML observability tooling, cross-functional hiring patterns that show stronger product and data-engineering coordination, and case studies that quantify realised revenue uplift or cost savings after debt-reduction initiatives. The report itself does not prescribe a single sequencing; it notes proven organisations operate at "dual velocities" rather than following a strict linear path.