ERP is the foundation. That was the through-line at Infor Velocity Week, and it is the reason the Velocity story landed.

Infor has spent years building industry-specific systems of record. Velocity Suite uses that foundation to make intelligence actionable. The agents, the automations, and the recommendations run on Infor’s process data, configuration, and live telemetry. The roadmap is proactive: the system looks at how the customer actually runs, ranks the next opportunities, and gives the team a sequence to follow.

That is a strong place to build from. The process is already there. The data is already there. Velocity is how that base starts producing steps a customer can take.

The catalog is the product

Infor’s thesis from the start was depth. Industry nuances are different, so the cloud generation is a set of multi-tenant platforms: discrete manufacturing, process manufacturing and distribution, and US healthcare and state and local government. Kevin Samuelson has put that rewrite at about $5 billion. By the company’s count, the acquisition era had left it with on the order of 150 code bases. The rewrite kept the industry focus and gave every micro-vertical a shared, modern base.

Infor Velocity Week showed how fine that base goes. A food-and-beverage company and a discrete manufacturer can share a payables outline. Receiving, putaway, and quality change once you reach dairy, protein, or beverage. Infor calls that object the Industry Process Catalog. Agents read it. Finished goods, yield, byproduct, and semi-finished are catalog terms with a specific meaning in that industry.

Benchmarks sit on the catalog. Infor Velocity Week cited APQC-sourced vertical benchmarks, used to highlight activities that can move. Process mining then shows where this tenant is spending time. Automation and prebuilt agents pick up from there. Because the processes are already in the product, Infor can start from configuration and let the agent work with a full picture of the flow.

Three ways the same loop shows up

The team walked three experience types. Each one draws on the catalog and the customer’s own data.

Task. A buyer workspace for quantities, approvals, and the work already on the desk. The agent brings that job forward with the context of the process behind it.

Event. A supplier calls and moves a delivery by about fourteen days. The buyer can see which production orders are affected, whether an alternate item has enough quantity, whether an alternate supplier can expedite, what to release now, and what to watch so later orders stay on track. Planning systems have long been able to run this what-if. Velocity collapses it into a live impact note, with the next action available on the workspace. The gain is speed, on top of the ERP the customer already trusts.

Process. Mining opens the third path. The public Xpress Boats story is a clear example: process mining across procure-to-pay, order-to-cash, and demand-to-build, a published 50 percent reduction in expedited shipping costs for materials, and a 98 percent faster diagnosis of process issues. In the room, the same customer showed the next layer. Normal aluminum-boat lead time sat around 95 to 97 days. Late orders stretched to roughly 125 or 126, with the delay concentrated in procure-to-receive and a handful of items driving it. The agent proposed new stocking parameters from history, receipts, and demand. A person applies the configuration. It also returned the operating list: vendor SLAs, cycle counts, supplier on-time metrics. The recommendation comes with the process change that makes it stick.

A food-and-beverage demo, described as live at a customer, used the same pattern in another dialect. Finished goods at risk of expiring inside three weeks, open customer orders in the same window, a cross-reference, then a match. The rule is first-expire-first-out, and the match follows the customer contract. A retailer glad to take short-dated organic product is a different delivery from a grocer that wants remaining shelf life. When no order is open, the next branch is donation while the goods are still inside date. The agent can walk that graph because the catalog already defines it.

Telemetry is how the roadmap stays ahead

Rick Rider’s session at Infor Velocity Week showed how this becomes a standing practice.

Velocity Suite is a suite on top of the ERP. In the room, Samuelson put multi-tenant SaaS at about $2 billion of roughly $3.5 billion in revenue, and Velocity at about 250 percent growth, attached to about two-thirds of new-customer transactions and expected at 15 to 20 percent of bookings. Those figures are from Infor Velocity Week. The packaging fits the goal: a flat fee and open consumption, so customers can take on the next use case as the value shows up.

The path to that model ran through proof. Innovation showcases around 2020, then Advanced Services in 2024, showed that each customer’s version of a use case is personal. Rider’s forward deployed engineering is a subscribed pod that stays with the account. Process mining surfaces the next bottleneck. The AI Adoption Hub ranks it and answers the three questions he hears from customers: where to start, how to start, and what the return looks like. The customer owns the assumptions in the calculator. Starforge is the internal engine that turns telemetry, industry patterns, and analyst-injected ideas into requirements, a design, and a package. He cited 435-plus use cases, and put time-to-production at under four weeks, down from about twelve. Infor’s August note says most of its AI solutions went to production in under four weeks in 2026, with direct Adoption Hub access for customers arriving in the coming months.

That is the forward deployed enterprise he described. Business creators shape the extension, IT promotes it, and Infor’s pod stays close as a partner. The proactive roadmap is what keeps the work moving between those moments.

Governance keeps the action inside the business

The governance session at Infor Velocity Week, led by Vignesh Subramaniam, gave the operating frame for all of this.

Two recent behaviors set the agenda: agents improving themselves on a task, and agents calling other agents when they need help. Infor’s answer is already in the October release, with a roadmap behind it.

The control plane sits on Infor OS, in two products.

GRC, the practice that traces back to Approva and still serves companies beyond the Infor base, owns business integrity. The new object is an AI use-case registry. Registered functions are the functions AI can reach, which gives an auditor a clear list of where AI is in use. The team tied that list to EU AI Act transparency duties, including clarity when a person is working with a system and when content is generated. Segregation of duties uses the same product. In the demo, a buyer asking why a supplier was late was guided to an AP colleague for the payable detail. The agent works within the human’s remit, on that person’s behalf.

Infor OS security owns identity. Every agent has a human owner. Rights arrive when a supervisor or a defined trigger needs the work done, scoped to the task and the window, through an access broker, and they retire when the task is done. The team cited on the order of 70 to 80 agents per CloudSuite today, with new ones coming through the FDE practice. AI Studio is how customers will create more of their own. Registration is how that growth stays accountable.

Data completes the picture. Albanero, acquired in 2024 and folded into the data fabric, is the quality and migration layer. Clean, governed data is what lets agents do reliable work. Customers can replicate data for analytics in their own fabric. Transactions returning to the ERP travel through the API gateway and an MCP gateway Infor is building, under OAuth and the same on-behalf-of or system identity. Use cases live in GRC. Agents live in OS security. Together they are the system of record for how intelligence acts.

The Industry AI architecture

Infor Velocity Week put this on one slide, titled Infor Industry AI Architecture. Precise outcomes, connected and open, and easy to use sit across the top. Governed by design sits underneath.

Industry Foundation is the ERP. CloudSuite applications, application logic, and Infor Data Fabric, with AI, machine learning, RPA, and document management already in the base. Agent Intelligence sits on that foundation: product supervisors, process and domain agents, Infor IQ at the micro-ontology level, and an MCP layer of process APIs and tools. Enterprise Orchestration connects the estate through an ecosystem supervisor agent and Infor IQ at the macro ontology, so Infor and the customer’s other systems can work from the same business context.

The Interaction Layer is where people meet it. Personalized and adaptive experiences, Claude, Copilot, and other assistants, autonomous process orchestration, and a third-party orchestrator. Value Intelligence carries targeted optimization, adoption roadmaps, and innovation readiness. AI Studio is the build path: use-case discovery, rapid build, and deployment. Governance, risk, and compliance run under the whole stack, on the Infor Industry Cloud Platform interoperability layer, on Amazon Bedrock.

That picture matches what Infor Velocity Week laid out. The data and the process live in the foundation. Velocity makes them actionable. The roadmap in Value Intelligence points to the next step. AI Studio is how that step gets built, inside a governance boundary the architecture already names.

What Velocity adds

Infor’s open platform means an interoperability layer, with Amazon Bedrock underneath, so Infor applications and the customer’s other systems can participate. Headless means a new experience can sit above core business logic while the ontology and the process definition stay intact. MCP is the assembly layer on Infor OS that lets those experiences come together quickly.

The foundation is the industry ERP. Velocity uses that foundation to make intelligence actionable, grounded in the customer’s data and processes, with a proactive roadmap that keeps pointing to the next best step. Value Intelligence carries the same indicators from the pre-sales value map into live process mining, so the goal the team set is the goal the team can see.

That is the loop from Infor Velocity Week. Catalog and mining show the work. The pod, and soon the customer’s own builders, ship a governed extension. The roadmap updates from how the business is actually running. It is a practical picture of enterprise AI, built on the system companies already use to operate.

About Shashi Bellamkonda

Loading bio…

LinkedIn Profile

Get new posts by email:

Subscribe

Powered by