Review of OMP, Supply Chain Planning Software Vendor

By Léon Levinas-Ménard

Last updated: December, 2025

OMP is a Belgian software and consulting vendor founded in 1985 that has grown from an academic optimization boutique into a global supplier of supply chain planning suites for asset-intensive industries. Its flagship product, now branded Unison Planning™, aims to cover the entire planning stack—from long-term network design and S&OP down to finite capacity scheduling—within a single data model and application. In recent years OMP has re-positioned the platform as a cloud-hosted, Azure-based solution featuring a “telescopic digital twin”, embedded optimization engines, explainable-AI (XAI) components and AI-branded “value enhancers” layered on top of the core solvers. A 20% growth-capital stake from Belgian investment holding Ackermans & van Haaren (AvH) has reinforced OMP’s status as a mature, privately held vendor with several hundred million euros of annual turnover and a client roster including large manufacturers such as Bayer and Kraft Heinz. The core question for this report is not whether OMP is commercially successful—it clearly is—but how far its technology stack, optimization engines and AI features really go beyond a sophisticated, integrated APS, and how that compares with Lokad’s explicitly probabilistic and code-centric approach to supply chain optimization.

OMP overview

Positioning and scope

OMP is a software and consulting company specialized in supply chain planning rather than execution systems. It traces its origins to an academic spin-off founded in 1985 by professor Georges Schepens, with a strong grounding in mathematical optimization for planning problems.¹² Over time OMP evolved from the earlier OMP Plus suite into today’s Unison Planning™ platform, still targeted primarily at large manufacturers in chemicals, life sciences, consumer goods, metals, packaging and related capital-intensive sectors.³⁴⁵

Externally, OMP is now consistently described as a global vendor headquartered in Belgium with international offices and several hundred to roughly 1,500 employees depending on the source and year.³⁶⁷ An investment agreement in 2020 brought Belgian holding Ackermans & van Haaren in as a ~20% shareholder, explicitly framed as growth capital to support further international expansion and R&D in digital twin, AI and planning automation.⁸⁹ A 2023 OMP fact sheet cites turnover of around €200m, confirming that OMP is a commercially mature rather than early-stage player.¹⁰

On the market-credibility side, OMP has been repeatedly recognized as a Leader in Gartner’s Magic Quadrant for Supply Chain Planning Solutions, including a 9th consecutive recognition as Leader in the 2024 report and a 2025 placement “highest for Ability to Execute” among the vendors evaluated.¹¹¹²¹³ A separate 2025 Gartner Critical Capabilities analysis ranks OMP among the top vendors across multiple planning use cases.¹⁴ While MQ positions are not technical guarantees, they confirm that large enterprises are deploying OMP at scale and that analysts see it as a serious, established APS vendor.

From OMP Plus to Unison Planning™

Historically, OMP Plus provided an integrated suite covering forecasting, inventory optimization, network design, S&OP, production planning and scheduling, all built around a common in-memory data model and solver framework.¹⁵¹⁶¹⁷ Older technical slides describe a central optimization layer able to address campaign planning, network design, inventory optimization, order promising and cutting/allocations using a mix of mathematical programming, graph algorithms, and heuristics, all fed by a central data model and integration layer.¹⁸

In the late 2010s, OMP re-branded its flagship solution as OMP Unison Planning™, now marketed as a unified, end-to-end planning platform that synchronizes all planning stages, horizons, functions and roles within one “telescopic digital twin”.¹⁹²⁰ OMP positions Unison Planning™ as an open, industry-specific platform with generic planning core plus vertical templates, and highlights AI engines, explainable AI, scenario management and a digital twin as differentiating elements.¹⁹²¹

Third-party descriptions (consultancies and partners) are consistent with this positioning: Unison Planning is typically presented as a full-scope APS covering strategic network design, demand planning, S&OP/IBP, supply planning, production scheduling, order promising and control-tower-style visibility, with strong presence in process manufacturing and packaging.⁴²²²³

Industries and reference customers

OMP focuses on capital-intensive manufacturing sectors where long lead times, complex production assets and multi-site networks make advanced planning valuable. Historic materials and case studies show deployments in metals, plastics, floor covering, paper and packaging, chemicals, pharmaceuticals, food & beverage and consumer goods.³¹⁵¹⁶¹⁸

Publicly named customers include Bayer, which selected OMP for global demand management and multi-regional planning (with OMP emphasizing AI-powered forecasting and support for hundreds of planners),²⁴ and Kraft Heinz, which is showcased in joint presentations on “autonomous supply planning” combining decision intelligence and “advanced mathematical optimization” powered by Unison Planning.²⁵ Earlier case studies under the OM Partners brand describe OMP Plus roll-outs for Axalta, Albéa, UCB and others, often highlighting integrated demand-to-schedule workflows and improved service levels.¹⁶¹⁸

In short, OMP is clearly deployed in large enterprise environments with complex networks and long planning horizons; this is not a small-business planning tool.

OMP vs Lokad

This section focuses on differences in technical approach and product philosophy between OMP and Lokad, not on which vendor is “better”. The contrast is especially relevant for companies evaluating both for supply-chain planning and optimization.

Platform model vs programmable engine

OMP provides a single vendor-controlled platform (Unison Planning™) with a fixed data model, UI, and embedded solvers. Customization happens via configuration, industry templates and scripted rules inside the product, but the core model structure and solver stack are proprietary.
Lokad offers a programmable optimization engine exposed through its domain-specific language Envision, where the entire forecasting and optimization logic is written as code by Lokad’s supply chain scientists or the client.²⁶²⁷ The platform is designed as a multi-tenant SaaS with an event-sourced data store and a custom distributed VM (“Thunks”), and does not impose a fixed planning process; the “application” is effectively the Envision program.

In practical terms: OMP sells a standardized APS suite you adapt to your business; Lokad sells a quantitative engine and DSL used to construct a bespoke decision-optimization app for your business.

Treatment of uncertainty and forecasting

OMP publicly emphasizes demand sensing, forecasting and resilient planning, and implicitly uses statistical and OR methods inside its engines, but does not disclose whether it models full probability distributions of demand and lead times or relies on point forecasts plus safety stock heuristics.¹⁹²¹²⁵
Lokad explicitly builds its entire stack on probabilistic forecasting, computing full demand distributions (quantile grids) and feeding them into stochastic optimization algorithms, an approach it has documented since the early 2010s and validated in competitions like the M5 forecasting challenge.²⁶²⁸

Thus, while both talk about uncertainty, Lokad’s probabilistic machinery is explicitly documented and central to its design, whereas OMP’s treatment is implicit inside solvers and not detailed publicly.

Optimization transparency and control

In OMP, optimization lives inside pre-built engines configured via the UI and project settings. Planners and consultants can adjust constraints, parameters and process flows, but they generally cannot re-write the underlying optimization logic in a first-class programming language. OMP does not publish its objective functions or solver formulations.¹⁸¹⁹
In Lokad, the optimization logic is first-class code: users write Envision scripts that manipulate random variables, build cost functions and call optimization routines. Lokad publishes conceptual descriptions of its algorithms (e.g., stochastic discrete descent, differentiable programming for supply chain) and treats the supply-chain model as a programmable object.²⁶²⁷²⁸

For organizations wanting fine-grained control over the math (and willing to invest in that expertise), Lokad offers more transparency and flexibility. For organizations preferring a productized APS with pre-packaged processes, OMP is more aligned with traditional expectations.

Role of consulting and configuration

Both vendors blend software with expertise, but they do so differently:

OMP projects typically revolve around configuring a large APS: modelling the network in Unison Planning™, integrating ERPs, and tuning planning workflows. OMP consultants and partners lead the implementation, and planners work mostly inside the OMP UI once the system is live.²²²³²⁹
Lokad positions its own team as “supply chain scientists” who co-author Envision code with the client. The engagement is closer to a joint R&D project on top of a reusable platform: the DSL is the central artifact, and Lokad emphasizes rapid iteration and continuous refactoring of that code as the business changes.²⁶²⁷

The net result is that Lokad encourages thinking of planning logic as version-controlled code, while OMP encourages thinking of planning as configurable workflows inside a suite.

Decision outputs

OMP primarily outputs plans within its workbenches: constrained demand plans, production schedules, supply plans and scenarios, visible inside Unison Planning™ and sometimes feeding back into ERP as planned orders or schedules.¹⁹²²²³
Lokad primarily outputs ranked lists of decisions (purchase orders, transfers, pricing moves) with associated economic metrics (expected ROI, stock-out risk, etc.), exported to ERPs/WMSs and often used as the basis for automated or semi-automated execution.²⁷²⁸

Both can support S&OP and tactical planning; however, Lokad’s philosophy is more decision-centric and financially explicit, whereas OMP’s is more process-centric and plan-centric.

AI narrative

OMP leads with AI-flavoured platform features: telescopic digital twin, XAI engines, generative AI assistants and value enhancers around collaboration and analytics.¹⁹²¹³⁰³¹ The focus is on empowering planners with intelligent tools embedded in a unified suite.
Lokad leads with AI/ML as internal machinery: deep learning and differentiable programming used to train forecasting and decision models, but with relatively little emphasis on chatbots or assistants. The front-end is spreadsheets and dashboards built on Envision, not an all-in-one planning cockpit.²⁶²⁸

A buyer looking for a single, role-based planning cockpit with embedded AI helpers is closer to OMP’s sweet spot; a buyer wanting a programmable quantitative engine that can be wired into existing tools is closer to Lokad’s.

Capabilities and deliverables of the OMP solution

Practical planning capabilities

From a functional perspective, Unison Planning™ aims to behave as a single APS platform covering:

Demand planning and forecasting (statistical forecasts, promotion planning, demand sensing)
Inventory and supply planning (multi-echelon planning, safety stock policies, material and capacity-constrained supply plans)
Sales & Operations Planning / Integrated Business Planning (S&OP/IBP)
Finite capacity scheduling and sequencing at plant level
Network design and strategic planning
Order promising / ATP–CTP and available-to-promise allocation
Control tower / analytics dashboards for cross-site visibility¹⁸¹⁹²²²³

Older technical decks on “The OM Partners supply chain suite” explicitly show a matrix of supported problem types (network design, campaign planning, inventory optimization, order allocation, cutting/blending, etc.) linked to factories, distribution, and demand across strategic/tactical/operational horizons, all implemented in one solver framework (OMP Plus).¹⁸ This is consistent with the current Unison Planning™ messaging, which emphasizes “one logic, one model” and end-to-end alignment from corporate to plant level.²⁰³²

From a user’s point of view, the deliverable of an OMP project is typically:

A unified planning model of the client’s network and constraints, instantiated in OMP’s data structures.
A configured set of planning workflows (demand, supply, scheduling, S&OP, etc.) with planning cycles and governance.
A set of optimization engines producing executable plans (order proposals, campaigns, schedules) and scenarios.
Role-specific workbenches and analytics dashboards providing KPIs, alerts and scenario comparison.¹⁹²⁰²²

Because OMP also sells consulting services, a substantial part of the value is in configuring this generic platform to a specific industry and client, often in collaboration with system integrators or partners.

Telescopic digital twin and planning “unison”

The central conceptual device in OMP’s current marketing is the “telescopic digital twin”. In OMP materials this is described as a single model of the supply chain that supports zooming seamlessly from strategic to detailed operational views, feeding both 360° Analytics dashboards and planning engines.¹⁹²⁰³²³⁰

OMP describes Unison Planning™ as a unified platform that “synchronizes all planning stages, horizons, functions and roles” and explicitly links this to a telescopic digital twin and explainable-AI engines for scenario exploration and decision support.¹⁹²⁰
Industry-specific pages (e.g., metals) emphasize that this twin provides end-to-end visibility across multiple plants and layers and informs centralized decisions using “advanced intelligence” and “mathematical methodology”.³²

Technically, this appears to be an evolution of the earlier common data model + central solver approach in OMP Plus, now enriched with:

more detailed representation of capacities, constraints and flows,
real-time or near-real-time data refreshes in some contexts,
and analytics dashboards layered on top.

The digital-twin branding is conceptually reasonable (the platform does maintain an internal representation of the supply network and its constraints), but public documents do not provide algorithmic details on how the twin is built, maintained or validated beyond conventional master-data management and integration practices.

AI, XAI and value enhancers

OMP’s current technology pages introduce “value enhancers” built “on AI, data science, deep learning and cognitive features”, intended to solve specialist challenges and “make smart plans even smarter”.²¹ Categories listed include:

Collaboration and simulation (workflow, scenario management, cross-functional collaboration)
360° Control (360° Analytics, 360° Response)
Various helper components around risk, what-if and monitoring.²¹

The same ecosystem also references explainable AI (XAI) engines, generative-AI assistants and optimization engines under the UnisonIQ label, presented as tools to explore scenarios and explain recommendations.¹⁹³⁰ OMP’s own news release on AvH’s investment lists “demand sensing, resilient planning and optimized planning automation” among the latest technologies embedded in the suite and again claims uniqueness of the telescopic digital twin in synchronizing strategy and execution.⁸¹⁶

Concrete optimization content is clearer in older, more technical material. A 2008 slide deck on OMP’s solver system lists a mix of mathematical programming, statistics and graph-based algorithms (including genetic algorithms and simulated annealing) as the backbone of the optimization layer across network design, inventory optimization, campaign planning and allocation problems.¹⁸ More recent marketing references “advanced mathematical optimization” in joint presentations with Kraft Heinz on autonomous supply planning, but without disclosing specific formulations or solvers.²⁵

Separately, OMP’s Data Genie tool is presented in podcasts and articles as a data-science component that improves the accuracy of digital twins by statistically inferring or validating master data (capacities, run rates, etc.) against observed operational data.³¹³³³⁴ This is arguably one of the more concrete “AI” components: it leverages data to adjust model parameters, but OMP does not publish methodological details beyond high-level descriptions.

Overall, OMP clearly uses a combination of OR algorithms + data science under AI branding. However, public sources stop short of exposing full model structures, objective functions or training regimes; the AI/XAI claims are credible but not deeply substantiated in open technical documentation.

Data management and integration

A recurring theme in OMP materials is that data quality and integration are primary bottlenecks for advanced planning. The Data Management & Integration component is positioned as the bridge between ERPs and the Unison Planning model, responsible for exchanging “accurate high-quality data” and maintaining consistent data with clear ownership and stable interfaces.²⁹ OMP emphasizes that smooth interfaces and consistency between ERP and planning system are key to reliable plans, which aligns with industry experience.

Historically, OMP Plus already offered OMP Data Manager and OMP Integrator for central data management and ERP integration, reinforcing that OMP has long focused on being the analytical system of record for planning, not replacing ERP but sitting on top of it.¹⁸ That pattern appears unchanged: Unison Planning™ still assumes ERPs and MES/WMS as execution systems, and focuses on becoming the planning system of record with its own data model and workbenches.

Technical implementation of the OMP platform

Architecture and stack

OMP describes Unison Planning™ as a cloud-based, Azure-hosted solution, with OMP Cloud offering the planning environment as SaaS.³⁵ The cloud proposition stresses standard expectations—elastic resources, global access, managed upgrades—rather than exposing deep architectural details.

Publicly available information on the tech stack mostly comes from job postings and third-party tech-stack trackers. These consistently indicate a Microsoft-centric backend:

Multiple OMP vacancies for Senior Software Engineer C# / .NET mention C#, ASP.NET Core, Azure Functions, Azure DevOps, Azure Service Bus/Event Hub and microservices-style architectures.³⁶³⁷
A “Senior .NET Software and DevOps Engineer” posting mentions the need to build cloud-native applications and APIs on Azure using event-driven architecture.³⁷
Tech-stack listings for OMP include C#, .NET, JavaScript/TypeScript, Angular/React, and Python, suggesting that the UI is web-based (SPA) and that some data-science components may leverage Python or similar tooling.³⁸

Older technical materials on OMP Plus describe an in-memory common data model with integration modules and a central database, and emphasize modular solvers and scenario management on top of this shared model.¹⁸ Modern Unison Planning™ messaging retains the notion of “one model, one logic” and real-time data usage, which is compatible with an in-memory, microservices-oriented architecture built around a central planning model.

However, OMP does not publish the kind of detailed architectural diagrams that, for example, Lokad exposes for its own platform. We therefore cannot confirm whether OMP uses event sourcing, particular columnar storage formats, or specific commercial solvers; only that it is a cloud-native, .NET/Azure-heavy stack with rich web front-ends.

Optimization engines and algorithmic transparency

The technical heart of OMP is its set of optimization engines. Historical documents give a relatively candid view:

The “OM Partners supply chain suite” deck explicitly cites mathematical programming, statistics and graph-based algorithms (e.g., simulated annealing, tabu search, constraint logic programming) as the main solver classes across network design, inventory optimization, campaign planning and allocation problems.¹⁸
The same slides highlight use cases covering capacity-constrained scheduling, inventory and portfolio optimization, and ATP/CTP allocation, all driven by this solver layer.¹⁸

In modern marketing, these become “advanced mathematical optimization” and “optimized planning automation”, with references to autonomous or automated planning in press releases—e.g., the 7_01 release was described as “next-generation digital supply chain planning software” offering better performance, scalability and supporting autonomous planning, real-time data integration and cloud deployment.³⁹ Supply Chain Digital’s profiles of OMP likewise refer to key capabilities such as advanced forecasting, production and distribution optimization, and the telescopic digital twin, and state that UnisonIQ incorporates generative AI assistants, explainable AI and optimization engines.³⁰

The gap, from a skeptical standpoint, is that OMP does not provide:

Open mathematical formulations of its core optimization problems.
Benchmarks or competitions showcasing solver performance.
Public APIs to the solver layer for arbitrary optimization tasks.

The vendor’s optimization footprint is credible and long-standing—the historical slides and academic-style talks show real OR heritage¹⁸⁴⁰—but external observers must treat the current “AI-powered optimization” claims as partially opaque, since detailed algorithmic descriptions are absent from public documentation.

AI, XAI and digital-twin tooling

Beyond the core solvers, OMP offers several AI-branded components:

Data Genie, framed as a data-science tool to calibrate and improve the digital twin’s master data using real operational data.³¹³³³⁴ This is a plausible application of statistics and ML to identify mismatches between assumed and actual capacities, yields, etc.
A family of “value enhancers” built on AI, deep learning and cognitive features, adding specialized functionality around collaboration, simulation, analytics and response management.²¹
The UnisonIQ framework, as summarized in Supply Chain Digital, which wraps generative AI assistants, explainable AI and optimization engines around the telescopic digital twin.³⁰

OMP’s own blog also acknowledges the proliferation of new planning buzzwords—telescopic digital twin, resilient planning, XAI—and attempts to give high-level definitions, but again remains conceptual rather than algorithmically precise.⁴¹

Taken together, these components suggest that OMP does employ:

Some ML models (e.g., for demand sensing, anomaly detection, master data calibration).
UI-level “assistants” and dashboards leveraging AI outputs.
Explanatory layers to help planners understand why a plan was generated (KPIs, constraint explanations, etc.).

However, the absence of detailed technical documentation means that OMP’s AI/XAI must be treated as marketing-backed but partially black-box from a third-party perspective. This contrasts with vendors that openly document their probabilistic models or differentiable-programming frameworks.

Deployment and roll-out methodology

Public case studies and partner materials indicate a fairly standard enterprise APS deployment pattern:

A multi-month project to model the supply chain in Unison Planning™, including network structures, BOMs, routings, resources, calendars and business rules.
Integration work to feed ERP, MES and other systems into the OMP data model via data-management tooling and interfaces.²⁹¹⁸
Iterative configuration of planning workflows and optimization engines, often with industry-specific templates.
Progressive go-live by region, plant or scope, with planners transitioning from legacy tools to OMP workbenches.

OMP and partners like EyeOn, Bluecrux and others frame Unison Planning™ projects as transformations that unify siloed planning processes, moving from local spreadsheets or legacy tools to a harmonized planning application.²²²³ Concrete deployment timelines vary per case study, but the pattern is clearly project-based, with combined software + consulting effort.

From a skeptical standpoint, this is typical of large APS: the solution is only as good as the model built. There is no public evidence that OMP offers a radically different deployment model (e.g., fully code-as-model, continuous integration of planning logic) beyond what is standard in large APS programmes.

Commercial maturity and client base

OMP’s customer list is not fully public, but the combination of:

Long-running case studies with industrial manufacturers (Axalta, Albéa, UCB, VDM Metals, etc.) under the OM Partners brand,¹⁶¹⁸⁴²
Sector-specific solution pages for metals, chemicals, life sciences, consumer goods and packaging,³³²
Named references like Bayer and Kraft Heinz,²⁴²⁵
Repeated Gartner Leader positions and strong Critical Capabilities showings,¹¹¹²¹⁴¹⁹
Turnover in the ~€200m range with substantial reinvestment in R&D,¹⁰¹⁵

all point to a vendor that is well beyond early adopters and entrenched in large-scale planning landscapes.

At the same time, the heavy emphasis on consulting and implementation services—within OMP and via partners—means that the boundary between software product and services is fluid. Many clients likely experience OMP as an APS programme rather than a pure off-the-shelf product.

Critical and skeptical observations

Summarizing the skeptical takeaways:

Integrated APS, not a radically new paradigm Functionally, Unison Planning™ is a powerful and integrated APS, but it remains in the tradition of monolithic planning suites: a single vendor-controlled data model, solver set and UX covering all planning levels. The “telescopic digital twin” concept appears as the natural evolution of OMP’s long-standing common data model and solver framework, now re-branded with digital-twin terminology and enriched dashboards, rather than a fundamentally new technology class.¹⁸¹⁹³²
AI/XAI branding ahead of published detail OMP’s marketing uses a dense cluster of AI labels—demand sensing, deep learning, XAI, generative AI assistants, autonomous planning—but technical transparency is limited. Aside from historical slides and high-level podcasts about Data Genie, there is little public information about model classes, objective functions, or training procedures.¹⁸²¹³⁰³¹³³ Claims should therefore be treated as plausible but not independently verifiable.
Optimization heritage is real but closed Historical material clearly shows deep OR expertise and a serious solver framework, and joint presentations with clients such as Kraft Heinz emphasize advanced mathematical optimization.¹⁸²⁵⁴⁰ However, modern OMP does not expose these solvers as open components or publish benchmarks. From an external perspective, Unison Planning™ behaves like a black-box optimizer wrapped in a rich UI.
Heavy implementation footprint Unison Planning™ projects require substantial modelling and integration, often supported by OMP consultants or partners.²²²³²⁹ This is expected in the APS segment, but means that the effective capability of the solution at a given client is dominated by how well the model was built and maintained, not solely by the vendor’s generic functionality.
Decision automation vs decision support OMP materials speak about “autonomous planning” and “optimized planning automation”, yet the public emphasis is still on planners using workbenches, exploring scenarios and making decisions supported by XAI dashboards.¹⁹²⁵³⁹ There is limited public evidence that large OMP clients are running fully automated replenishment or scheduling at scale without human approval; the realistic interpretation is advanced decision support with some automation, not full autonomy.

These points do not undermine OMP’s usefulness—its platform is evidently valuable to many manufacturers—but they highlight that OMP should be understood as a modern, integrated APS with strong OR roots and evolving AI components, rather than as a fully transparent, code-centric optimization platform.

Conclusion

From a purely technical and architectural standpoint, OMP is a mature, integrated APS platform with strong OR heritage, substantial commercial adoption and a broad functional footprint across demand, supply, S&OP and scheduling. Its Unison Planning™ suite reflects decades of evolution from OMP Plus: a common data model, embedded solvers for a wide range of planning problems, and increasingly sophisticated analytics and dashboards built around a “telescopic digital twin”. The partnership with Ackermans & van Haaren and repeated Gartner Leader recognitions confirm that OMP is a mainstream choice for large manufacturers seeking to modernize complex, multi-site planning landscapes.

At the same time, a skeptical, evidence-driven reading of public information suggests that OMP’s technology should be understood as high-end APS with strong optimization capabilities, rather than as a fully transparent, AI-native optimization platform. Its AI and XAI claims are credible but lightly documented externally; its optimization engines are proven but closed; and its deployment model is project-heavy, with success hinging on the quality of the model built by OMP and its partners. None of this is unusual in the APS world—but it is important context for buyers comparing OMP with more code-centric, probabilistic platforms such as Lokad.

For organizations that want a unified planning suite with one vendor owning data model, UI and solvers—and are comfortable with a traditional APS implementation approach—OMP is a strong, commercially validated option. For organizations that instead prioritize full mathematical transparency, probabilistic optimization exposed as code, and decision-centric outputs, Lokad offers a materially different paradigm. The right choice depends less on high-level AI/APS marketing terms and more on the organization’s appetite for modelling as a configurable suite (OMP) versus modelling as maintained code (Lokad).