AI agents are arriving in the enterprise. Most companies aren’t seeing the profit yet
McKinsey’s State of AI 2025: 62% of organisations are experimenting with AI agents, but only 39% report any EBIT impact — and for most of them it is under 5%. The gap is data and process, not models.
Corex® Team
McKinsey’s global survey The state of AI in 2025: Agents, innovation, and transformation, released in November 2025, captures a moment of transition. Nearly 2,000 respondents across 105 countries described how far AI has really gone inside their organisations.
The numbers
- 62% say their organisations are at least experimenting with AI agents.
- 23% are scaling an agentic AI system somewhere in the enterprise — most often in IT service management and knowledge work.
- Only 39% attribute any EBIT impact to AI, and most of those say it is less than 5% of EBIT.
- About 6% qualify as “high performers” attributing 5% or more of EBIT to AI.
Why the value isn’t showing up
An AI agent can only act on what your systems know. If sales lives in one spreadsheet, stock in another and invoices in a third, an agent has nothing reliable to work with. The companies capturing value have usually done the unglamorous work first: one source of truth for customers, inventory, finance and operations, with workflows defined clearly enough to automate.
A practical roadmap for mid-sized companies
- Consolidate core data into an integrated ERP/CRM rather than adding another tool.
- Map and simplify processes before automating them — automation multiplies whatever is already there.
- Start agents where rules are clear: service desks, document handling, reconciliation, scheduling.
- Measure EBIT, not activity. Hours saved and errors avoided should show up in the P&L.
CorexOS® brings 22+ enterprise modules — from sales and inventory to accounting, HR and projects — onto one platform, so automation and AI have a solid foundation to build on. Explore CorexOS®.
Sources
Written by
Corex® Team
