Infor announced an expanded version of its Infor Industry AI architecture, alongside the next release of Infor Velocity Suite, the package through which customers buy its AI capabilities. The update adds more industry-specific AI agents, an adaptive user experience that assembles role-based views, and governance controls that extend across a customer’s enterprise.
Infor tied the launch to the second edition of its Enterprise AI Adoption Impact Index, a survey of 2,111 business decision-makers conducted by YouGov on Infor’s behalf. Two-thirds (68%) of respondents said off-the-shelf AI does not adequately address their industry’s needs. Infor CEO Kevin Samuelson put the company’s argument plainly: “What matters is what those models know about your business.”
The announcement puts two questions in front of ERP teams: where an AI agent gets its understanding of the business, and who answers for what it does.
How Infor Gives Agents Industry Context
Infor says its agents work from a shared semantic layer called Infor IQ. The layer gives every application in a CloudSuite consistent definitions of core concepts, such as an item, a location, or a customer, along with the industry detail underneath them. Infor’s example: an agent ordering raw materials for a food manufacturer must understand a sugar shipment’s Brix measurement, while an automotive parts agent must understand a vehicle identification number.
Agents also draw on Infor’s Industry CloudSuites, Industry Process Catalogs, and industry-specific domain language models. Infor exposes business-level process APIs, so an agent can make a single call to create a purchase order instead of chaining dozens of technical calls. The company says fewer steps reduce the chance of error and lower token costs.
The architecture connects to non-Infor applications and to third-party orchestration and analytics tools, and users can work through the Infor GenAI Assistant or AI assistants they already use. Infor lists over 350 prebuilt use cases and reports customer results of up to 60% faster shipment processing and up to 90% time savings in some procurement, supply chain, manufacturing, and sales workflows.
What the Survey Says About Fit and Authority
YouGov polled respondents in the US, UK, Germany, France, Australia, Singapore, and Japan in August 2026, according to Infor’s blog summary of the results. The industry-fit concern was sharpest in hands-on operating sectors: 76% of distribution respondents, 73% of manufacturing respondents, and 69% of retail respondents said generic AI does not fit their needs.
More than half (54%) of leaders said they are comfortable with autonomous agents fully executing critical processes without human input at every step, and 59% expect AI investment to rise over the next 12 months. The same respondents estimate that 44% of AI-generated insights and workflows still need expert human review or adjustment. Infor’s report calls this a post-adoption gap, in which companies move AI into daily operations faster than they build the structures to support it.
Governance in the Platform, Ownership With the Customer
Infor’s governance changes run through its orchestration layer. Human-approval workflows, agent permission structures, explainable AI logic, and immutable logs of every agent action sit in a governance, risk, and compliance layer at the core of the Infor Industry Cloud Platform. Infor’s blog argues that agents should be treated like users under a zero standing privilege model, receiving access only for the task at hand and losing it when the task ends.
Inside customer organizations, ownership is still unsettled: asked who is responsible for governing AI, 23% of respondents named the CEO or executive leadership, 22% the CIO or CTO, 15% an AI committee, and 10% department heads, while 15% said no one has clear responsibility. That figure reached 21% in Japan and 4% in Singapore. A third of respondents (33%) cited data security, sovereignty, or compliance as their biggest barrier to advancing AI.
Infor sells Velocity Suite, which bundles these controls with its agents, process mining, automation, and implementation experts, for a flat annual fee instead of per-user, per-bot, or per-transaction pricing. Platform controls can limit and record what an agent does, but deciding who owns its outcomes remains a management decision.
What This Means for ERP Insiders
Test how agents learn your business. Ask vendors where their agents get definitions for items, locations, and customers, and pilot on a process that depends on industry-specific data. ERP teams that check this early avoid paying later for workarounds and manual correction.
Assign AI accountability before agents act. Name an owner for agent-driven decisions and define which actions need human approval, then confirm the platform records them. A named owner gives auditors and process leads a clear answer when an agent’s decision is challenged.
Model AI costs for full-scale use. Compare flat-fee packages such as Velocity Suite with usage-based pricing against expected agent volume and the review work your teams will still carry. A full-scale cost model gives finance a realistic budget before adoption expands.
