Enterprise AI is moving well beyond the experimentation phase. Investment is rising, businesses are giving AI a greater role in consequential decisions and confidence in the technology is growing.
But the harder challenge begins after adoption. As organisations move AI into everyday operations, a post-adoption gap is emerging between deploying the technology and making it work effectively at scale. Industry fit, long-term cost visibility, and governance are becoming the key issues businesses need to solve next.
That raises a broader set of questions: Does your AI understand the business well enough to be usable? Can its costs remain predictable as usage grows? And who is ultimately responsible for the outcomes it produces?
New global research from Infor™, conducted by YouGov among 2,111 business decision-makers across the US, UK, Germany, France, Australia, Singapore and Japan, shows where the post-adoption gap is appearing.
Confidence is rising, but readiness looks different depending on where you sit
Businesses broadly believe they can handle AI. 79% say they can manage an AI implementation project without disrupting daily core operations. Experience is changing those perceptions, and not always in the same direction.
C-suite respondents are also more positive about AI’s impact on frontline workers than department heads are: 70% of C-suite respondents report a positive impact, compared with 59% of department heads.
The takeaway is simple: the view from the top can look different from the reality closer to day-to-day operations.
The first post-adoption challenge is making sure AI actually fits the business
Fit becomes more important as AI moves from experimentation into operational workflows.
68% of businesses say off-the-shelf AI does not adequately address challenges specific to their business or industry.
That becomes harder to ignore as AI moves into everyday workflows. A system can be technically capable while still lacking the context to understand how a particular manufacturer, distributor, retailer or other enterprise actually operates.
When that context is missing, organisations can end up compensating with more customisation, more workarounds and more human intervention just to make the technology useful.
Industry context has to be part of the foundation, not something layered on after the fact.
Giving AI more responsibility also changes the cost equation
Investment is rising, but long-term cost visibility has not materialised.
Going live is only the beginning. As AI expands across more users, workflows and agents, costs can compound through fragmented systems, unnecessary steps, and repeated human review.
The question is no longer simply what AI costs on day one. It is what it costs to run once usage scales and the effects ripple throughout an organisation.
As AI gains authority, governance has to keep pace
Businesses are increasingly comfortable giving AI a role in consequential work.
In the US, for example, the share of respondents who are very comfortable with autonomous AI agents handling many processes rose from 10% in April to 15% in the latest wave.
Globally, 80% want AI involved in high-stakes decisions such as financial forecasting and workforce planning, while only 12% prefer humans to make those decisions without AI input.
But greater AI involvement does not mean humans can step away. AI may be gaining authority, but humans remain the backstop.
Businesses estimate that 44% of AI-generated insights and workflows still require expert human review or adjustment.
Human review alone will not catch every problem. As AI agents work in teams and begin improving themselves, the potential for harm grows with every system and set of data they are allowed to access.
With RSI (Recursive Self Improvement) and teams of agents becoming more pervasive in agentic architectures, a well-intended agent can still create material risk if it is given more authorisation than it needs. This is why agentic governance needs to be embedded within data access and security layers. Agents need to be treated like a physical user and work with a 'zero standing privilege' approach.
In practice, that means an agent should receive access only for the task in front of it and lose that access when the task is done.
AI governance has an owner in most businesses, but there is no single model for where it belongs
Responsibility for governing AI use, managing risk, and ensuring compliance is spread across the enterprise:
- 23% name the CEO or executive leadership.
- 22% name the CIO or CTO.
- 15% point to an AI committee or governance group.
- 10% name a dedicated Chief AI Officer.
- 10% put responsibility with individual department heads.
- 15% say no one person has primary responsibility, or it is unclear who does.
The differences between markets reinforce how unsettled that model remains. In Japan, 21% say no one has primary responsibility. In Singapore, that figure is just 4%.
That does not mean every organisation needs the same governance structure. What it does mean, however, is that every organisation needs to know where accountability sits before an AI-driven decision is tested.
Closing the post-adoption gap is the next phase of enterprise AI
Adopting AI and operating it successfully at scale are two different undertakings.
Businesses are making real progress: investment is climbing, deployment is advancing and comfort with AI authority is growing.
Those gains do not close the post-adoption gap on their own, though. Industry fit, cost efficiency and governance still determine whether adoption turns into lasting value.
For leaders looking to close the post-adoption gap, three questions stand out:
Answering those questions is what moves an organisation toward a truly agentic enterprise, where AI agents and people work as one coordinated system under human oversight.
Infor Industry AI is built to help businesses close the post-adoption gap
Infor Industry AI starts with a simple principle: AI should be built around how an industry actually works, with the connectivity and governance needed to scale responsibly.
Industry context is built into the foundation rather than layered on afterward, helping businesses start closer to value without relying on excessive customisation or workarounds. Infor Industry AI connects applications, data, workflows, and agents into one operating environment, while bringing AI directly into the businesses processes and role-aware experiences where work already happens.
Governance is built in from the start. As agents take on more responsibility, their actions remain explainable, traceable and auditable, helping organisations maintain human oversight without slowing adoption.
The goal is not simply to deploy more AI. It is to create the foundation for an agentic enterprise, where people and AI agents can work as one coordinated system with industry context, connectivity and governance built in.
Download the full report here
Learn how Infor's industry-specific AI solutions can help your organisation move from experimentation to execution.
Learn more:
About the research
This analysis is based on survey data collected from business decision-makers across the United States, United Kingdom, Germany, France, Australia, Singapore and Japan between August 5 and 19, 2026. The research was conducted by YouGov on behalf of Infor.
The survey included:
- 550 respondents in the United States
- 254 respondents in the United Kingdom
- 258 respondents in Germany
- 260 respondents in France
- 263 respondents in Australia
- 260 respondents in Singapore
- 266 respondents in Japan
All figures reflect the global aggregate unless otherwise stated. Respondents represent a cross-section of industries and organisation sizes, providing a global view of how businesses are navigating the shift from AI adoption to AI governance.