The transformational potential of AI is already properly established. Enterprise use circumstances are constructing momentum and organizations are transitioning from pilot tasks to AI in manufacturing. Firms are not simply speaking about AI; they’re redirecting budgets and sources to make it occur. Many are already experimenting with agentic AI, which guarantees new ranges of automation. But, the highway to full operational success continues to be unsure for a lot of. And, whereas AI experimentation is in all places, enterprise-wide adoption stays elusive.
With out built-in knowledge and programs, secure automated workflows, and governance fashions, AI initiatives can get caught in pilots and wrestle to maneuver into manufacturing. The rise of agentic AI and growing mannequin autonomy make a holistic method to integrating knowledge, purposes, and programs extra vital than ever. With out it, enterprise AI initiatives might fail. Gartner predicts over 40% of agentic AI tasks shall be cancelled by 2027 as a result of price, inaccuracy, and governance challenges. The actual challenge just isn’t the AI itself, however the lacking operational basis.
To grasp how organizations are structuring their AI operations and the way they’re deploying profitable AI tasks, MIT Expertise Assessment Insights surveyed 500 senior IT leaders at mid- to large-size firms within the US, all of that are pursuing AI in a roundabout way.
The outcomes of the survey, together with a collection of knowledgeable interviews, all performed in December 2025, present {that a} sturdy integration basis aligns with extra superior AI implementations, conducive to enterprise-wide initiatives. As AI applied sciences and purposes evolve and proliferate, an integration platform might help organizations keep away from duplication and silos, and have clear oversight as they navigate the rising autonomy of workflows.
Key findings from the report embody the next:
Some organizations are making progress with AI. In recent times, research after research has uncovered an absence of tangible AI success. But, our analysis finds three in 4 (76%) surveyed firms have at the very least one division with an AI workflow totally in manufacturing.
AI succeeds most often with well-defined, established processes. Practically half (43%) of organizations are discovering success with AI implementations utilized to well-defined and automatic processes. 1 / 4 are succeeding with new processes. And one-third (32%) are making use of AI to varied processes.
Two-thirds of organizations lack devoted AI groups. Just one in three (34%) organizations have a group particularly for sustaining AI workflows. One in 5 (21%) say central IT is accountable for ongoing AI upkeep, and 25% say the accountability lies with departmental operations. For 19% of organizations, the accountability is unfold out.
Enterprise-wide integration platforms result in extra sturdy implementation of AI. Firms with enterprise-wide integration platforms are 5 instances extra doubtless to make use of extra various knowledge sources in AI workflows. Six in 10 (59%) make use of 5 or extra knowledge sources, in comparison with solely 11% of organizations utilizing integration for particular workflows, or 0% of these not utilizing an integration platform. Organizations utilizing integration platforms even have extra multi-departmental implementation of AI, extra autonomy in AI workflows, and extra confidence in assigning autonomy sooner or later.
This content material was produced by Insights, the customized content material arm of MIT Expertise Assessment. It was not written by MIT Expertise Assessment’s editorial workers. It was researched, designed, and written by human writers, editors, analysts, and illustrators. This contains the writing of surveys and assortment of information for surveys. AI instruments that will have been used have been restricted to secondary manufacturing processes that handed thorough human evaluate.
