So many firms are not getting meaningful results from AI investments. Genpact’s June 2026 report, “The $18 Trillion Opportunity: How Four Enterprise Debts Will Make or Break Your AI Future,” argues that the constraint is not the model layer alone. It is the state of the business beneath it: data, process, technology, and talent. The study, done with HFS Research, surveyed more than 2,000 enterprise executives globally and says nearly $18 trillion in value is trapped inside Global 2000 companies.
The report shows many enterprises are trying to scale AI on top of systems that were never designed for it. Genpact says 92% of senior executives believe agentic AI will change how work gets done, but only 13% say it is already integrated into operations. At the same time, 85% of leaders say enterprise debt is actively limiting their AI value, and more than half have no funded plan to resolve it.
Genpact and HFS frame $18 Trillion AI Gap as value already within large enterprises that is not being captured because the operating foundation is weak. The report says fixing these foundations could translate into about 8% faster annual revenue growth and 16% annual cost reduction across the Global 2000.
[ALSO READ: IBM Study Reveals Growing Disconnect Between AI Ambition and Governance ]
Before the rise of AI, outdated processes and legacy systems were inefficient, but we managed to work with them. Now, in the age of AI, those same flaws have turned into significant obstacles. When a model is trained on bad data, it quickly produces bad results. If you insert an AI agent into a broken process, it can end up amplifying the existing problems instead of fixing them.
What enterprises do
For enterprises, do not just “buy more AI.” It is “fix the environment AI has to run in.” Genpact’s report says the companies that make progress treat debt resolution and AI transformation as one program.
If data is fragmented, and workflows are manual, systems are old and hard to integrate, and employees are not trained to work with AI, then even good software will stall. In the report’s terms, the bottleneck is execution quality, not ambition.
[ALSO READ: AI Data Debt: The Risk Lurking Beneath Enterprise Intelligence ]
What are enterprise debts, and why do they matter?
Genpact defines four enterprise debts: process debt, data debt, technology debt, and talent debt. These are not accounting liabilities. They are accumulated operating weaknesses. Process debt refers to inefficient, manual, or ungoverned workflows. Data debt refers to poor-quality, fragmented, or non-AI-ready data. Technology debt refers to legacy systems and integration complexity. Talent debt refers to skills gaps and workforce misalignment.
The report argues that these debts matter more now because agentic AI depends on the entire workflow, not just the model. A company can deploy automation tools and still fail if approvals are messy, master data is unreliable, systems do not connect cleanly, or employees do not know how to use the tools. That is why the report says AI and debt resolution are the same program, not separate ones.
Why enterprise debts are blocking AI value, and how to fix it
According to the report, data debt is the biggest AI blocker, cited by 33% of respondents. Technology debt follows at 28%, process debt at 23%, and talent debt at 16%. In the report’s view, data debt keeps AI stuck in pilot mode, technology debt makes AI expensive to integrate, process debt makes AI unreliable in production, and talent debt slows adoption and limits human oversight.
Genpact says enterprises need cleaner and better-governed data, simpler and more consistent processes, more modern systems, and a workforce prepared to work in a human-agent operating model. The report also notes that more than 40% of enterprise capacity is tied up maintaining, correcting, or working around these debts, which is capacity that cannot be used for transformation.
How four enterprise debts will make or break your AI future
Genpact quantifies each debt separately; It says data debt and process debt each represent about $7.7 trillion in potential value. Technology debt represents $1.5 trillion, and talent debt about $1 trillion. The report also says data debt alone consumes up to 40% of employee time in data-intensive functions because of reconciliation, rework, and quality correction.
Technology debt is often the most visible problem because it shows up in system age, integration headaches, and developer time spent on maintenance. But the report says that focusing on technology alone misses the larger issue. If the process is broken and the data is poor, then new tools only automate dysfunction. That is the difference between digitizing work and improving it.
[ALSO READ: AI Is Growing Faster Than the World’s Data Centers Can Handle ]
