Go back to Market Research
Infor supplies enterprise applications for manufacturing, distribution, and other industries. Its supply chain portfolio includes demand forecasting, supply planning, production scheduling, sales and operations planning, and the Nexus network for shipment visibility and collaboration. These products connect with ERP applications through the shared Infor OS platform. Infor has belonged to Koch since 2020. (1, 16, 22)
Overall supply chain score: 5.0/10
| Dimension | Score |
|---|---|
| Supply chain depth | 4.4/10 |
| Decision and optimization substance | 5.0/10 |
| Product and architecture integrity | 5.0/10 |
| Technical transparency | 6.4/10 |
| Vendor seriousness | 4.2/10 |
Infor’s governing process is configurable planning: forecast demand, calculate constrained supply plans, schedule production, and let users compare scenarios or resolve exceptions. Its manuals expose substantial operational machinery, including monetary costs, manufacturing transitions, forecast intervals, and logistics predictions. The central weakness is the connection between those calculations. Published methods do not establish that uncertain demand and uncertain receipts enter a common economic calculation that chooses actions across the network.
The cost of bridging that gap falls on model configuration, integrations, and operational judgment. Supply Planning displays a long inventory of penalties alongside financial costs; Production Scheduling documents cases where a more tractable objective is preferable to directly minimizing changeovers. These choices determine what the software optimizes. A successful solver run cannot establish that the chosen formulation captures the business’s losses and opportunities. (2, 3)
Infor publishes extensive implementation documentation and explicit restrictions on practitioner-authored Python. Its AI communication is less disciplined. The broad business-advisor promise recorded in 2018 now has a successor in the agentic-enterprise proposition, while the latest automotive example still asks a planner to examine a shortage recommendation and request the change. Concrete automation and narrow predictive applications have shipped; evidence of autonomous economic management remains much thinner. (7, 12, 14, 24)
Infor versus Lokad
Infor’s proposition starts with an enterprise operating environment: transaction systems, industry workflows, shared integration services, planning modules, and human workspaces. Lokad’s proposition starts with a recurring economic calculation: which purchases, production orders, allocations, or other actions offer the highest expected return. Its inventory optimization documentation describes probabilistic treatment of demand, lead times, and other uncertainties, with customer-specific decision logic maintained by Supply Chain Scientists.
Consider a manufacturer deciding whether to expedite a component shared by several products. Infor can expose the shipment’s predicted arrival, plan material availability, calculate production alternatives, and display cost differences. The difficult calculation is the value of expediting across possible arrival dates and competing uncertain demand, including the production opportunities displaced elsewhere. An ETA, a forecast interval, and a constrained plan supply different pieces of that calculation. The reviewed interfaces do not show how Infor combines them into one risk-adjusted decision. (2, 4, 5)
Lokad’s stated design places those economic interactions inside executable decision logic. The broader perspective appears in Introduction to Supply Chain and Holimization: operational failures should improve the formulation, rather than generate an expanding collection of manual repairs. The comparison must include the cost of maintaining that logic, including Lokad’s scientific and engineering work.
Infor offers transaction processing, shop-floor workflows, and a trading network beyond Lokad’s scope. For an existing Infor customer, those connections can reduce implementation work. The decision-engine comparison concerns the quality of recurring actions and the total effort needed to keep improving them; the breadth of the ERP portfolio cannot settle it.
What the numerical machinery actually does
Supply plans optimize costs and configured penalties
Supply Planning’s comparison screen lists revenue, profit, holding costs, disposal costs, purchases, regular and overtime production, transport, and backorders. It also exposes penalties for shortages, safety-stock violations, excess stock, capacity violations, lost demand, expedited activity, and shipment limits. Separate feasibility, purchasing, and main solver results include mixed-integer optimality gaps. This is inspectable optimization machinery with a financial vocabulary. (2)
The penalties create an important modeling responsibility. A safety-stock shortfall, an unserved forecast, and a late customer order are different events. Assigning weights to all three can make the same underlying shortage count several times, or can protect a policy target at the expense of a more profitable action. The published cost screen supplies no general method for deriving those weights from economic consequences under uncertainty.
Mixed-integer optimization is particularly exposed to this problem. Supply chains combine uncertain receipts, discrete purchases, substitutions, expiration, and interactions across products and resources. Making the formulation tractable can remove precisely the interactions that determine value. Expanding a fixed planning model into a large collection of uncertain scenarios adds computational burden; it does not establish that the resulting formulation remains usable or economically faithful. Infor’s disclosed solver results show how well a supplied model was solved, not how well that model represents the business.
Production scheduling exposes real physical detail
Production Scheduling documents a paint-filling example in which changing colors normally requires a one-minute flush, but an idle period exceeding thirty minutes lets paint dry and increases cleaning to fifteen minutes. A transition matrix can represent both the sequence and elapsed-time dependency. This is meaningful manufacturing detail: the cost of a sequence depends on when it runs. (3)
The same documentation says time-dependent changeovers are harder for the solver and suggests span or distance-to-work-area objectives as generally better ways to obtain the desired arrangement. That recommendation reveals a practical tension between physical fidelity and computational convenience. The economic test is whether the alternate objective preserves profitable production choices when due dates, scarce materials, and downstream demand change. Minimizing a convenient scheduling measure provides no such assurance.
The maintainability test follows directly. When an operator discovers that cleaning behavior, yield, or available substitutions differ from the model, somebody must revise the formulation, replay affected decisions, and deploy the correction. A new penalty or manual rescheduling convention can hide the defect without fixing it.
Forecast uncertainty exists, but its decision use is unclear
Infor’s forecasting documentation names established methods, including Holt-Winters, state-space exponential smoothing, BATS, and model selection using information criteria. Its newer AI engine documents configurable training losses and prediction intervals formed by clustering forecast errors. The default interval coverage is 90%; that is a coverage setting, not a claim of 90% forecast accuracy. (4, 18)
These disclosures support more than point forecasting. They do not describe a joint distribution of demand and replenishment outcomes feeding Supply Planning’s objective. An interval around each demand series leaves dependencies unresolved: simultaneous promotion effects, common supplier disruptions, or several products competing for the same component. Those dependencies matter when evaluating a purchase or production commitment.
Infor also markets demand sensing through frequent updates, incremental data, and pattern recognition. The label adds no separate technical capability. Its contribution would need to be isolated against a forecast using the same information and refresh schedule, with inputs available before the relevant procurement or production deadline. The published overview does not provide that comparison. (20)
Nexus predicts logistics outcomes, including hypothetical changes
Nexus’s Predict API accepts changed shipment attributes and hypothetical shipments, then returns a predictive arrival through an asynchronous request-and-poll interface. Examples include changing the carrier or transport mode. The documentation says hypothetical predictions are excluded from ordinary ETA error calculations and reports. (5)
A hypothetical shipment has no observed arrival against which to calculate an ordinary forecast error. The material issue is validation of the intervention: whether changing the carrier or mode produces the predicted benefit under comparable conditions. The published interface provides neither that validation nor a probability distribution for valuing the change against its freight cost and inventory consequences. A point arrival prediction helps visibility but leaves the economic decision unfinished.
The 2025 engineering account co-written by Nexus’s technology leader describes predictions before departure and during transit, data-quality checks, and Databricks-based distribution of network data. This supports a functioning predictive and data-sharing operation. It does not document the transfer of calibrated supply uncertainty into SCP’s purchasing and production optimization. (6)
Technology history: broad intelligence, narrower delivery
Infor’s history contains substantial platform investment. Its corporate chronology dates ION to 2009, the AWS partnership and CloudSuite launch to 2013, GT Nexus to 2015, and Birst and Coleman to 2017. The 2019 filing records the 2016 acquisition of Predictix, promoted through cloud-native predictive methods, machine learning, and big-data analytics. These are identifiable products and acquisitions behind the successive labels. (13, 25)
The 2018 filing nevertheless made an expansive claim for Coleman: an industry-aware assistant that anticipates and advises, acting as a “true business advisor.” Voice access and repetitive-task automation were offered as support for that proposition. Such interfaces and task-specific models do not substantiate broad enterprise intelligence. The filing did not delimit the assistant’s reasoning competence, failure cases, or economic responsibility. (12)
A more concrete result appeared in August 2022. Zeelandia used Coleman to rank five products for a sales visit and reported reducing preparation from thirty minutes to five. Pricing recommendations were another application. The announcement identifies deployment in the Czech Republic and distinguishes later geographic rollout. This is credible evidence of a useful, bounded recommendation workflow. Its measured labor saving concerns sales preparation, not autonomous supply chain control. (24)
Infor’s 2022 internship book provides a revealing engineering counterpart. Nexus Science sought faster ETA code for more than 100 million rows and heuristics for mapping recommended resolutions back to existing orders and shipments. Another team proposed automating forecast-model implementation and tuning, explicitly acknowledging the specialist effort ML required. These projects identify substantive work and persistent operational difficulties beneath the earlier advisor language. (17)
An added note on the Zeelandia announcement states that Coleman was renamed Infor Artificial Intelligence in April 2024 while its functionality continued. The 2025 Nexus account adds a concrete Databricks data-sharing mechanism. The November 2026 LN automotive release promotes MCP access to more than 6,000 APIs and consolidation of over 100 single-purpose agents into twelve domain agents. These developments concern different layers: model services, data distribution, and orchestration. (6, 14, 24)
The historical weakness is the recurring distance between broad intelligence claims and bounded demonstrated tasks. Infor has delivered useful applications, but has not supplied a comparable record of unattended, economically measured purchasing and production decisions. Rebranding and adding a conversational integration layer leave that earlier promise unresolved.
Productivity: useful automation and a persistent planner loop
The latest automotive shortage example is explicit: an agent checks expected stock, recommends moving material or changing an order, and the planner reviews the recommendation, requests the change, and reruns the check. That can reduce navigation and coordination. It retains recurring human intervention in the decision loop. MCP access simplifies calling existing software; it supplies neither the economic objective nor the uncertainty model required to judge competing actions. (14)
A more substantial planning reference is Grupo Peñafiel. The 2025 customer story covers three plants, more than 400 SKUs, and forty distribution centers. Its supply chain director estimates a 30% improvement in production planning time after consolidating the national plan. The same account describes automated material planning and integration of further demand data as subsequent steps. The measured scope is a planning improvement during a continuing implementation, not complete automation of material decisions. (23)
Infor’s February 2025 Pfaudler account reports seventeen hours per day saved on sanctions checks and sharply reduced supplier-selection work. Its separate item-master project covers 800,000 items; completion within six months and inventory savings of €1–2 million are presented as expectations. Those projected savings cannot be counted as realized returns. The distinction matters because data cleanup, compliance checks, supplier scoring, and economic replenishment eliminate different workloads. (15)
A useful productivity assessment would follow both recurring decisions and the work needed to maintain their calculation: planners, integration administrators, model specialists, and vendor or partner engineers. The reviewed cases do not provide that complete labor account. Replacing spreadsheet consolidation or repetitive data lookup can produce genuine savings while leaving the cost of revising the decision system largely intact.
Software materials and ownership cost
Infor OS provides a common integration layer, but the supply chain portfolio still combines several numerical and operational systems. The following materials have identifiable scopes.
| Component or family | Documented role and scope |
|---|---|
| AWS API Gateway, S3, Lambda, CloudFront, Route 53 | Named in the May 2023 Infor OS reference architecture. |
| Infor Data Fabric, ION, Birst, and Coleman/Infor AI | Data storage and preparation, event/workflow integration, reporting, and model services in documented OS workflows. |
| Databricks and Delta Sharing | Nexus data-quality, prediction, and network-data distribution described in June 2025. |
| R and AWS | Nexus Science’s 2022 ETA engineering project; historical project scope. |
| Python, Spark/Pandas, and bundled numerical libraries | Current Infor AI preprocessing and user-script interfaces. |
| Supply Planning and Production Scheduling engines | Separate documented planning and scheduling calculations with their own parameters and objectives. |
The architecture has a visible operational cost. AWS and Infor’s 2023 scrap-monitoring example moves historian data through periodic ingestion, model processing, Data Fabric, workflow alerts, and Birst reporting. Each stage has data semantics, failure handling, and maintenance responsibilities. Adding Nexus predictions to a production decision also requires consistent identifiers, units, timestamps, and interpretation across the planning and logistics systems. Shared authentication does not perform that reconciliation. (16)
Infor publishes service entitlements for integration traffic, API execution, script time, storage, cloud egress, and model training. Its scripting meter charges execution that times out or fails after starting. This makes the infrastructure burden concrete: failed or repeatedly revised computations consume the same contracted resources used for productive work. Public limits help a buyer estimate the workload; they do not establish the final subscription or implementation cost. (9, 10)
Configuration creates a separate maintenance bill. Forecast-engine documentation includes model settings, mapped regressors, masks, and troubleshooting actions. For a failed forecast-magnitude test, one suggested response is raising the threshold or disabling the check. That gives an administrator a recovery path, but also makes it possible to silence a symptom without correcting the data or model. Such changes require regression tests tied to operational consequences. (21)
Infor documents useful change controls in its shared platform. Security-package upgrades compare new templates with live customizations and leave the customer to select which changes to adopt. ION scripts can be tested with supplied payloads and inspected for runtime and size limits. These facilities make maintenance more manageable; they also identify work that remains with administrators and application owners. (11, 19)
Security: limits on practitioner code
Infor’s ION documentation sets a fifteen-second runtime limit, five-megabyte input and output limits, and a memory ceiling of 1,000 MB. Scripts cannot access external or online resources and are limited to the submitted code, attached libraries, and supplied input. Infor AI’s preprocessing scripts are separately documented as network-isolated. (7, 8)
The boundaries have specific value. Blocking external communication constrains data exfiltration; bounded runtime and memory limit resource abuse; passing explicit datasets narrows the data available to a script. The remaining questions concern how those datasets are authorized, how dependencies are controlled, how tenant isolation is enforced, and how compromised or erroneous transformations are promoted and rolled back. The reviewed pages do not fully document those mechanisms.
ION’s test environment allows up to thirty seconds while production permits fifteen. Its guide instructs users to inspect that difference and test their largest payloads. This is useful operational transparency. It also means a successful interactive test alone is insufficient evidence that an extension will run safely and reliably in production. (11)
Economic accountability
Infor publishes identifiable customer activities, named customer speakers, and some bounded before-and-after measurements. Those accounts are more useful than an unqualified portfolio-wide ROI percentage. They still leave the full economic proposition open: implementation and recurring costs, a comparison period, decision-quality losses, and responsibilities when projected gains fail to arrive.
Peñafiel’s planning-time estimate, Zeelandia’s sales-preparation saving, and Pfaudler’s projected inventory reduction should remain separate outcomes. None establishes the net return of the entire supply chain architecture. A deployment needs an agreed baseline, a record connecting recommendations to execution, and responsibility for repairing a model that performs poorly. The public material reviewed here does not define those commitments. (15, 23, 24)
Detailed scorecard
Scores express an assessment of the documented product and evidence. Each dimension is the mean of five criteria; the overall score is the mean of all twenty-five.
Supply chain depth
| Criterion | Score |
|---|---|
| Economic framing | 5/10 |
| Decision end-state | 5/10 |
| Conceptual sharpness on supply chain | 5/10 |
| Freedom from obsolete doctrinal centerpieces | 3/10 |
| Robustness against KPI theater | 4/10 |
Average: 4.4/10
Economic framing — 5/10. Supply Planning exposes revenue and operational costs, but also policy penalties whose economic derivation is unspecified. The mapping from business consequences to the objective remains a consequential implementation responsibility. (2)
Decision end-state — 5/10. Supply plans, production sequences, and logistics recommendations address executable activity. Scenario comparison and the current automotive shortage workflow still depend on users to settle and authorize recurring choices. (2, 14)
Conceptual sharpness on supply chain — 5/10. Time-dependent cleaning and hypothetical shipment changes address real manufacturing and logistics interactions. Their connection to joint economic decisions under uncertain demand and supply remains unresolved. (3, 5)
Freedom from obsolete doctrinal centerpieces — 3/10. S&OP alignment, forecast improvement, exceptions, and policy targets still organize the offering. The published process gives planners extensive machinery to maintain plans without demonstrating that this recurring work is necessary to produce superior actions. (1, 2)
Robustness against KPI theater — 4/10. Detailed manuals distinguish prediction intervals from accuracy, and customer stories identify specific tasks. Forecast quality at operational decision granularity and returns net of total cost remain largely unreported; some prominently stated inventory benefits are projections. (4, 15)
Decision and optimization substance
| Criterion | Score |
|---|---|
| Probabilistic modeling depth | 4/10 |
| Distinctive optimization or ML substance | 5/10 |
| Real-world constraint handling | 7/10 |
| Decision automation and labor productivity | 5/10 |
| Resilience under real operational complexity | 4/10 |
Average: 5.0/10
Probabilistic modeling depth — 4/10. Demand intervals and logistics prediction are documented. Their coupling into a common distribution of demand and receipts, used to value purchasing and production actions, is not. (4, 5)
Distinctive optimization or ML substance — 5/10. Inspectable forecasting methods, solver outputs, and engineering projects establish numerical substance. Public evidence does not demonstrate a distinctive solution to the stochastic, nonlinear economic problem that conventional planning formulations simplify. (2, 17, 18)
Real-world constraint handling — 7/10. The scheduling manual represents elapsed-time-dependent transitions, while supply plans include production, transport, stock, and capacity effects. These concrete mechanisms justify substantial credit despite the limits of their combined economic formulation. (2, 3)
Decision automation and labor productivity — 5/10. Customer accounts document shorter planning and sales-preparation tasks; current agents can execute requested changes. Total labor across planning, model maintenance, and integration is unmeasured, and recurring planner review remains prominent. (14, 23, 24)
Resilience under real operational complexity — 4/10. Parameters, diagnostics, and scenario comparison support adaptation. The documented preference for easier scheduling objectives and the option to disable forecast checks expose the risk of accommodating model defects rather than correcting their economic consequences. (3, 21)
Product and architecture integrity
| Criterion | Score |
|---|---|
| Architectural coherence | 4/10 |
| System-boundary clarity | 6/10 |
| Security seriousness | 6/10 |
| Software parsimony and ownership cost | 3/10 |
| Compatibility with programmatic and agent-assisted operations | 6/10 |
Average: 5.0/10
Architectural coherence — 4/10. Infor OS supplies common integration and data services. SCP, Nexus, analytics, and model services still expose distinct processing paths, with no documented common stochastic objective linking their outputs. (6, 16)
System-boundary clarity — 6/10. Public APIs, workflow descriptions, and service entitlements make many integration responsibilities visible. The precise handoff from supply predictions to economic planning remains less clear than the surrounding data transport. (5, 9, 16)
Security seriousness — 6/10. Explicit script input boundaries, network restrictions, and execution limits directly reduce attack surface. Published material is incomplete on dependency admission, production-data authorization, tenant-boundary validation, and recovery from a damaging extension. (7, 8)
Software parsimony and ownership cost — 3/10. Integration, data preparation, model execution, reporting, and customer configuration impose several operating responsibilities. Metered execution and manual customization reconciliation make parts of that burden visible, without an end-to-end maintenance-cost account. (9, 10, 19)
Compatibility with programmatic and agent-assisted operations — 6/10. Versioned Nexus APIs, scripts, and LN’s agent interfaces offer useful callable operations. Access to existing functions does not give an agent a tested economic policy or unrestricted control over the numerical formulation. (5, 7, 14)
Technical transparency
| Criterion | Score |
|---|---|
| Public technical documentation | 8/10 |
| Inspectability without vendor mediation | 7/10 |
| Portability and lock-in visibility | 5/10 |
| Implementation-method transparency | 6/10 |
| Evidence density behind technical claims | 6/10 |
Average: 6.4/10
Public technical documentation — 8/10. Detailed current manuals expose forecast settings, solver measures, physical scheduling behavior, scripting limits, and troubleshooting. Numerical implementation and complete economic validation remain proprietary or unpublished. (2, 3, 4, 7)
Inspectability without vendor mediation — 7/10. A prospective customer can inspect API requests, model parameters, and operational restrictions before a sales process. Reproducing the complete decision engine and its customer-specific results still requires licensed access and implementation work. (5, 18)
Portability and lock-in visibility — 5/10. APIs and Delta Sharing improve data access. Workflow configurations, planning semantics, and numerical engines remain tied to their Infor environments; data extraction alone does not transfer a working decision system. (5, 6, 19)
Implementation-method transparency — 6/10. Reference architectures, payload tests, and package-upgrade instructions reveal practical responsibilities. The work required to validate economic objectives and interactions across products is less explicitly documented. (11, 16, 19)
Evidence density behind technical claims — 6/10. Primary manuals and product-scoped engineering material substantiate many mechanisms. Broad autonomy and end-to-end economic claims lack comparably detailed operational evaluations. (14, 17, 23)
Vendor seriousness
| Criterion | Score |
|---|---|
| Technical seriousness of public communication | 4/10 |
| Technology roadmap integrity | 4/10 |
| Conceptual sharpness | 5/10 |
| Economic accountability | 3/10 |
| Defensibility in an agentic-software world | 5/10 |
Average: 4.2/10
Technical seriousness of public communication — 4/10. Engineering documentation is specific, but the marketing narrative repeatedly generalizes from bounded applications to broad intelligence and orchestration. Demand-sensing language adds further claims without an isolated decision-value evaluation. (12, 14, 20)
Technology roadmap integrity — 4/10. The history includes delivered platforms and measurable narrow applications. It does not resolve the distance between Coleman’s earlier business-advisor promise and today’s planner-mediated agent workflows. The 2024 renaming preserves continuity while leaving that capability gap open. (12, 14, 24)
Conceptual sharpness — 5/10. Infor’s strongest materials identify concrete manufacturing and integration problems. Its overall proposition gives less attention to the distinction between making information available, solving a configured plan, and selecting profitable actions under uncertainty. (2, 3, 16)
Economic accountability — 3/10. Named customers and bounded activity savings provide useful reference points. Public accounts do not define complete costs, an economic failure threshold, or responsibility for correcting underperforming decision logic. (15, 23, 24)
Defensibility in an agentic-software world — 5/10. Industry transaction semantics, manufacturing constraints, and an established logistics network are substantive assets. Their value exceeds that of a conversational interface, but a productivity advantage over agents improving executable decision logic remains unmeasured. (3, 6, 14)
Sources
Supply Chain Planning. accessed November 4, 2026. [2] Plan comparison: financial and solver measures
Plan comparison: financial and solver measures. 2026.11; accessed November 4, 2026. [3] Idle-dependent changeover times
Idle-dependent changeover times. 2026.11; accessed November 4, 2026. [4] AI forecast engine parameters
AI forecast engine parameters. 2026.11; accessed November 4, 2026. [5] Predictive ETA API
Predictive ETA API. accessed November 4, 2026. [6] Infor Nexus and Databricks: The Data Intelligent Future of Supply Chains
Infor Nexus and Databricks: The Data Intelligent Future of Supply Chains. 2025-06-02; accessed November 4, 2026. [7] Scripts in data flows
Scripts in data flows. 2026; accessed November 4, 2026. [8] Prepare the Data activity catalog
Prepare the Data activity catalog. 2026.09; accessed November 4, 2026. [9] Infor OS service limits
Infor OS service limits. accessed November 4, 2026. [10] Scripting Time
Scripting Time. accessed November 4, 2026. [11] Testing a script
Testing a script. 2026.10; accessed November 4, 2026. [12] Infor fiscal 2018 Form 10-K
Infor fiscal 2018 Form 10-K. 2018-06-28; accessed November 4, 2026. [13] Infor fiscal 2019 Form 10-K
Infor fiscal 2019 Form 10-K. 2019; accessed November 4, 2026. [14] CloudSuite Automotive 2026.10: automation to agentic enterprise
CloudSuite Automotive 2026.10: automation to agentic enterprise. 2026-11-04; accessed November 4, 2026. [15] Manufacturer transforms efficiency with GenAI
Manufacturer transforms efficiency with GenAI. 2025-02-04; accessed November 4, 2026. [16] Infor OS on AWS
Infor OS on AWS. 2023-05-03; accessed November 4, 2026. [17] Infor Tunisia internship book 2022
Infor Tunisia internship book 2022. 2021; accessed November 4, 2026. [18] Forecast engine parameters
Forecast engine parameters. 2026.10; accessed November 4, 2026. [19] Security package post-deployment tasks
Security package post-deployment tasks. 2026.10; accessed November 4, 2026. [20] What is demand sensing?
What is demand sensing?. accessed November 4, 2026. [21] Forecast engine error codes
Forecast engine error codes. 2026.11; accessed November 4, 2026. [22] Koch completes Infor acquisition
Koch completes Infor acquisition. 2020-04-06; accessed November 4, 2026. [23] Grupo Penafiel customer story
Grupo Penafiel customer story. 2025-05; accessed November 4, 2026. [24] Zeelandia Coleman AI customer results
Zeelandia Coleman AI customer results. accessed November 4, 2026. [25] Infor corporate milestones
Infor corporate milestones. accessed November 4, 2026.