Inside the solution: 5 questions customers ask about agentic AI in Infor

As more Infor customers explore what agentic AI actually means for their business, the same questions keep coming up, whether we are talking to a CIO thinking about where to start or a delivery lead trying to understand if agents can be used for the repetitive work they do. Some are technical like: How do agents actually talk to Infor systems? Others are more foundational: What does this mean for governance, and can Infor Velocity Suite help build this intelligence into systems? 

Here are five questions we hear most often, answered by our teams working on agentic AI across Fortude’s Infor practice. 

This is usually the first question, and the honest answer is that it’s harder than it sounds without the right layer in between. AI agents don’t naturally know how to call Infor APIs, handle authentication, or interpret the structure of ERP data. Build that from scratch for every use case, and you end up with a different custom integration for each one. 

That’s the gap our Model Context Protocol (MCP) server for Infor M3 is built to close. Think of it as a standardized interface layer: A user asks a question in plain language, the MCP server routes it securely to the right Infor endpoint, and the response comes back structured and ready for the agent to act on or present. 

The bigger benefit is reusability. Once the MCP layer is in place, it supports multiple agents and use cases rather than a new integration project every time, and it’s compatible with leading AI platforms that support the MCP standard, not locked to one vendor. 

Infor Velocity Suite is Infor’s own AI package for running agentic workflows inside CloudSuite, covering things like industry-specific agents, generative AI, and process orchestration. Our role as an implementation partner is helping customers actually get value from it – mapping which business processes are ready for agentic execution, configuring agents against a customer’s specific data model and workflows, and connecting Velocity Suite into the wider system landscape so agents aren’t working in isolation. 

In practice, that means starting narrow. We help identify one workflow with clear friction, whether that’s order management, supply planning, or a finance process, and configure Velocity Suite’s agents around that before expanding. The technology can coordinate a lot, but it only delivers value when it’s grounded in how a specific business actually runs, not a generic template.

The best way to answer that is with what this actually looks like on the floor, not in a slide deck. A few examples we point to include: 

• A purchasing agent that adjusts sourcing decisions based on supplier performance and delivery schedules, rather than just flagging a stock shortfall for someone to act on manually. 

• Finance teams see only the handful of close-cycle anomalies that genuinely need a human look, instead of reviewing every transaction that has already been processed cleanly. 

• A disruption in the supply chain, say a delayed freight shipment, triggering a coordinated response across systems: rerouting, updating schedules, and notifying the right people, instead of one team member manually chasing five others. 

• An accounts payable agent that matches invoices against purchase orders and receipts, flags only genuine mismatches for review, and routes clean invoices straight through for approval and payment. 

What ties these together is that the agent isn’t just surfacing information. It’s carrying a decision through to action, inside the workflow, with a person still able to step in and review.

This is consistently one of the first concerns customers raise, and rightfully so. Letting AI agents act on live ERP data, not just query it, is a meaningful step, and it only works if governance is built in from the start rather than bolted on afterward. 

On the technical side, our MCP server enforces authentication and permission scopes for every request, so an agent only ever has access to what it’s explicitly allowed to use. Velocity Suite works on a similar principle at the platform level: a central governance layer sits over agent actions, whether initiated by a person or an agent, so teams can see what an agent did, what data it used, and whether it stayed within defined limits. 

For most customers, the practical starting point is deciding where agents can act autonomously and where a human needs to approve first. Low-risk, routine actions are a reasonable place to start; higher-stakes decisions stay human-reviewed until there’s a track record to build confidence on. That boundary should be a deliberate design choice, not an afterthought.

Yes, though it’s worth being precise about which part of the process this affects. There’s a difference between agents that run a client’s business processes once CloudSuite is live (that’s the Velocity Suite side) and agents that help build the implementation itself. 

We tested the second one directly. Our Americas consulting practice ran a controlled internal pilot, led by Rakhita Samarakone, putting AI agents to work on the repeatable, documentation-heavy parts of CloudSuite delivery: first drafts of specs, test scripts, and delivery documentation that consultants would otherwise write from scratch. The findings were consistent: meaningful time saved on that repeatable work, output good enough to use with review rather than a rewrite, and every consultant in the pilot said they’d keep using it. 

The principle behind it is simple. A senior consultant editing a strong first draft is faster and sharper than the same consultant starting from a blank page, and every participant still validates the agent’s output before it goes anywhere near a client. Agents don’t replace the judgment that makes an implementation succeed. They give experienced people more time to spend on it. 

…….. 

If you are thinking about where agentic AI fits into your own Infor environment, whether that’s connecting agents securely, building on Velocity Suite, or speeding up the implementation itself, our team is happy to talk through what a sensible starting point could look like for your business.