Infor says organizations are deploying AI faster than they are collecting value from it. The second edition of its Infor Enterprise AI Adoption Impact Index, released alongside the company’s Industry AI announcement, sets out the evidence: a comparison of April and August results in four markets, plus a seven-market view of where leaders see obstacles.
YouGov surveyed 2,111 business decision-makers for Infor between August 5 and 19. Infor calls the distance between adoption and results the post-adoption gap and links it to industry fit, cost predictability, and AI governance.
Deployment Moved Faster Than Reported Gains Between April and August
Of the four markets surveyed in both waves, only the US saw deployment and reported efficiency gains rise together. Deployment there climbed about 13 percentage points, while reported gains rose about 7.
France shows the widest split, with deployment up about 12 points and reported gains down about 14. The United Kingdom saw gains fall as deployment edged up, and Germany held steady on both measures. August samples ran from 254 to 550 respondents per market, so the smaller movements are best read as direction.
Spending is set to keep rising, with 59% of leaders expecting their business to increase AI investment over the next 12 months. Cost clarity is thinner: 88% say predictable costs matter when evaluating AI, yet most do not have a comprehensive view of long-term costs.
Respondents Report Capability and Data but Still Cite Poor Industry Fit
About two in three respondents (68%) say off-the-shelf AI does not address their industry’s needs, and the answer does not come from leaders who see themselves as unprepared. In industrial manufacturing, distribution, and retail, roughly two in three to three in four report enough in-house capability, and a similar share rate their data as mature.
Infor reads the result as a gap in industry fit, arguing that general-purpose AI treats industries as interchangeable. Its report contrasts high-volume automotive lines with low-volume, highly configured aerospace and defense builds.
Distribution respondents report the highest misfit and the lowest efficiency gains of the three industries, and retail respondents report the lowest misfit and the highest gains.
Shashi Bellamkonda, principal research director at Info-Tech Research Group, said in Infor’s announcement that the next phase of enterprise AI depends on how well organizations apply it within their industry and business processes.
Leaders Are Comfortable With Autonomous Agents, but Ownership Is Split
More than half of leaders (54%) are mostly or very comfortable with autonomous agents executing critical business processes without human input at every step. Asked who primarily owns AI governance, respondents split: 23% name the CEO or executive leadership, 22% the CIO or CTO, and 15% say no one has clear responsibility.
Confidence in data does not remove the concern. About three in four respondents (74%) say their data is mature and well governed enough to support reliable AI, yet one in three cite security, sovereignty, privacy, or compliance among their top barriers. Japan is the low point on data confidence by a wide margin. The report offers one possible explanation: clean data without governance around it still produces outputs that need checking before use.
Infor argues that ownership is splitting further rather than consolidating, and that governance has to be built into the technology, with agent actions that can be traced and audited and a person able to step in, override, or approve. Its Industry AI architecture places human-approval workflows, agent permissions, and audit trails in its orchestration layer.
What This Means for ERP Insiders
Track reported gains next to deployment. The index shows the two measures moving apart in most markets. ERP teams that report both can see whether a rollout is paying back before they extend it.
Test data readiness and governance separately. Respondents rate their data highly yet still cite security and compliance as barriers, so each needs its own check. Reviewing both gives ERP teams a clearer view of what limits scale.
Settle ownership before widening agent authority. With responsibility split across executives, IT leaders, and committees, a challenged agent decision may have no clear person to answer for it. Naming an owner gives ERP teams a direct answer when that happens.
