ERP has an AI adoption problem. Vendors have moved to put copilots, agents, and generative AI into enterprise applications. What has moved more slowly is connecting those capabilities to real business processes and proving they make the business run better.
The real question is not whether Infor is adopting forward-deployed engineering (FDE). It is whether Infor can make it work less like a services engagement and more like a repeatable engine for AI-enabled ERP execution. I have seen this pattern across ERP for years. Capabilities arrive faster than organizations can absorb them. Placing engineers closer to customer operations is not new. What matters is whether it shortens the feedback loop with product engineering, turns customer problems into product learning, and makes the next deployment easier. Infor now appears to be pushing beyond conventional FDE by investing in new technology to reduce repeat engineering work, increase reuse, and build a more scalable model across enterprise applications.
Infor is pairing FDE with its AI Adoption Hub, using AI to help build, refine, deploy and manage AI solutions. Rick Rider, Infor’s SVP of AI Innovation, argues that FDE is becoming a standard approach to enterprise AI delivery and that the next issue is how the model evolves.
Infor’s four-week deployment provides a useful proof point. More important is whether each engagement improves the ERP platform and guides future deployments.
AI Availability Is Not AI Adoption
The hard work starts when AI operates inside a business process. It needs trusted data, access to the right workflow, business rules, approvals, and exceptions. Someone also has to decide what it can recommend, what it can execute, and who owns the outcome when it gets something wrong.
Manufacturing and distribution make that challenge clear. An inventory or production exception can touch demand, purchasing, schedules, warehouse capacity, suppliers, and financial constraints before the right action becomes clear.
An agent on top of ERP does not remove that complexity. The issue is getting intelligence into the workflow, not adding another AI feature.
FDE Closes The Distance Between Infor’s Product And The Customer’s Problem
Rider describes FDE as a structured way to scope, build, review, and validate an AI solution with a customer before production. In ERP, the more important effect is reducing the distance between product engineering and the operational problem.
The vendor develops the software. An integrator implements it. Consultants may redesign processes, and customer teams make it work day to day. AI cuts across those roles, and a single agent may depend on vendor application logic, integrator connections, customer data, and operating rules known only to a few business users.
FDE puts engineering closer to the operational reality. That can shorten the path from identifying an issue to building something useful while showing product teams where workflows break and which exceptions matter.
The point of FDE in ERP is turning customer problems into product learning. If a lesson from one manufacturer makes a capability easier to deploy for the next, the model scales. If every engagement stays unique, it remains a services model.
FDE Itself Will Not Differentiate Infor
Infor is not alone. Microsoft is investing $2.5 billion in its Frontier Company, AWS committed $1 billion to a global FDE organization, ServiceNow and Accenture launched a joint program, and SAP is hiring forward-deployed AI engineers and architects.
As FDE expands across the market, vendors will scale it through different combinations of capital, partners, and software. For ERP buyers, the better comparison is whether each approach reduces engineering dependence, creates reusable product capability, and makes the next deployment easier.
Infor’s answer is the AI Adoption Hub. Rather than scale only by adding engineers, Infor is trying to automate more of the prototyping, refinement, deployment, and ongoing management work. The test is whether the Hub increases reuse while reducing engineering effort.
Industry Context Gives Infor A Different Starting Point
Infor is not introducing FDE in isolation. Its industry-specific CloudSuites connect with Velocity Suite capabilities including process mining, Value+ automation, generative AI, and agents.
Process mining can show where workflows slow or break. Automation can remove defined manual work. Agents can recommend or take actions. FDE can connect those capabilities to a specific operating problem.
Infor’s industry orientation means teams start with more context around orders, inventory, production, scheduling, and quality than a general-purpose platform. That can shorten discovery, but even similar manufacturers can have very different data, approvals, plant configurations, supply networks, and operating practices. Infor still has to show that industry standardization shortens deployment without treating similar customers as identical.
As Rider explained in a KramerTALK conversation, “Because even if we’ve got the exact same industrial manufacturing customer on the exact same footprint, it’s always going to be uniquely different based on how they handle their business process and the data underneath because the data definition can be slightly different at the transactional level from customer to customer.” A shared industry foundation can accelerate deployment without assuming every customer operates the same way.
AI Adoption Hub Extends The FDE Model
Rider makes an important distinction: FDE is the delivery methodology while the AI Adoption Hub is the architecture intended to accelerate and manage it. Infor is applying AI to the process of building and managing AI.
The Hub can help create a solution blueprint from industry and customer requirements, refine it against workflow data, move it into production, and manage changes as requirements evolve.
Rider describes four stages: prototype, refine, deploy and evolve. The evolve stage may matter most. AI systems will need to change as processes, data, regulations and business priorities change, making ongoing improvement part of the lifecycle rather than another project.
As Rider put it, “This is a partnership-driven, continuous innovation cycle and that’s how we treat our FDE practice.” That makes ongoing improvement part of the customer relationship, but it also raises the importance of ownership, change control, and measurable progress over time.
If the Hub can manage that lifecycle without repeated scoping and deployment cycles, it addresses more than implementation speed. It supports ERP as a system that increasingly participates in business execution.
That is not proven at scale. Infor still needs to show how much human engineering the evolve stage requires and whether subsequent deployments become easier.
Four Weeks Is A Deployment Metric, Not The Outcome
Infor says most of its AI solutions have gone into production in under four weeks during 2026.
Rider cites Team Air Distributing, where three agents were built and deployed in under two weeks, along with other customer examples reaching production in two to three weeks.
Those examples matter because enterprise technology often takes too long to produce value. But ‘in production’ needs a clear definition. ERP buyers should ask whether the timeframe includes integration, security review, testing, training, change management, and customer effort before and after go-live.
Then ask what changed operationally: production exceptions, fulfillment, manual intervention, inventory, cycle time, or warehouse productivity. Those are the measures that survive budget review, not how quickly something reached technical go-live.
Rider connects that operational improvement to a broader business goal: “The real power is when a focus on efficiency turns into a business growth mindset.” Efficiency matters, but the larger opportunity is using the capacity and insight created by AI to serve customers better, respond faster, and support growth.
Time to value should mean time to measurable operating improvement.
FDE Must Not Rebuild ERP’s Old Customization Trap
ERP has spent decades dealing with the cost of customization. Customer-specific modifications solved immediate problems but often made upgrades harder and systems more expensive to maintain.
FDE could recreate that problem if every engagement produces another unique AI solution. The better model turns customer-specific learning into reusable product capability.
The AI Adoption Hub could help manage reuse and updates without repeating the full development cycle. Buyers should ask what Infor learned from an engagement and whether that learning improves the standard product.
A scalable FDE model should make the tenth deployment easier than the first.
AI Building AI Raises The Governance Bar
Using AI to create, refine, or modify other AI systems raises the governance bar. Customers need visibility into what changed, why it changed, and who approved it.
Permissions have to define what an agent can recommend, approve, or execute. Changes need testing and auditability, with clear rollback and exception handling. Accountability still belongs to people. Identifying an issue is different from changing an order, adjusting inventory or taking action inside a financial process.
In manufacturing and distribution, that difference has real consequences. An agent that reschedules production, releases a purchase order, or adjusts a financial posting needs controls that let the right person see, question, and reverse the action before it propagates.
The same principle applies upstream when the AI Adoption Hub generates or refines an agent. Someone still has to review what the AI created, not just what it later does in production.
Faster delivery only works if control keeps pace.
Infor’s Partner Ecosystem Is Part Of Whether FDE Scales
FDE also has to fit with Infor’s partner ecosystem. Systems integrators and consultants still matter for integration, data readiness, and change management. Infor can bring product and engineering knowledge while partners provide implementation scale and customer context.
The tension is role clarity. FDE moves vendor engineers into territory systems integrators and consultants have traditionally owned. Infor will need to ensure partners can build repeatable services around the model, rather than being left only with integration and data cleanup.
If the AI Adoption Hub reduces the engineering load required for each deployment, partners could help extend FDE without recreating custom services at scale. If it does not, the model simply moves the same work into a different workflow.
What Infor Still Has To Prove
As FDE becomes more common, the methodology itself will matter less than what a vendor learns from each deployment and how quickly that learning becomes reusable product capability.
I would like Infor to publish a reuse rate: how often a capability built for one customer moves to others without re-engineering. Pair it with customer-confirmed results such as fewer exceptions, shorter cycle times, or lower inventory.
If each engagement increases reuse, reduces engineering effort, and improves a customer process, Infor has a more scalable model. Four weeks gets attention. Repeatable operating improvement is the measure.
