Review of Lanner, Supply Chain Software Vendor

By Léon Levinas-Ménard

Last updated: November, 2025

Lanner Group Ltd (now part of Royal HaskoningDHV’s Twinn portfolio) is a UK-based simulation software editor whose core product, WITNESS, is a long-standing discrete-event simulation (DES) and “predictive simulation” environment used to build digital twins of factories, warehouses, service operations, and business processes. Lanner’s technology stack revolves around model-centric simulation rather than large-scale data-centric forecasting: users build process models, parameterise them with arrival patterns, cycle times, and resource rules, and then run many stochastic replications to assess performance under different scenarios. Over time, Lanner has extended WITNESS with Experimenter/Optimizer modules, 2D/3D animation, and APIs to external code (C++, .NET, Python), and created the L-Sim simulation engine to embed DES into BPM/BPSim tools such as ARIS and Sparx Enterprise Architect. Since its 2019 acquisition, Lanner’s software has been marketed under the Twinn brand as part of a broader “digital twin and predictive simulation suite,” with supply chain & logistics positioned as one of several verticals alongside manufacturing, healthcare, and energy. In practice, Lanner’s solutions are typically deployed for project-style, scenario-based analysis (designing or re-configuring lines, DCs, and service systems), not as always-on, high-frequency optimisers for day-to-day replenishment or pricing decisions. From a technical standpoint, WITNESS and L-Sim are mature, well-documented simulation engines with deep academic usage and stable desktop-centric deployment; however, their “AI” and optimisation claims mostly reflect classical DES, input modelling, experimentation and search over scenario parameters rather than state-of-the-art machine learning or integrated probabilistic decision optimisation.

Lanner overview

Lanner Group is a specialist in discrete-event and predictive simulation software headquartered in the UK, with WITNESS as its flagship product and L-Sim as a Java-based simulation engine embedded in third-party BPM and BPSim tools.¹² The company traces its origins to British Leyland’s operational research group in the late 1970s, via AT&T Istel, where early visual interactive simulation tools (SEE WHY) were developed and later evolved into WITNESS.¹³ WITNESS is positioned as a general-purpose DES platform for modelling manufacturing, logistics, healthcare, and service systems, with 2D/3D animation and an Experimenter module used to sweep parameter combinations and search for improved system designs.⁴⁵⁶ In 2019 Lanner was acquired by Royal HaskoningDHV; the software is now marketed within the “Twinn” predictive simulation and digital twin suite, alongside other analytics and digital twin offerings.⁷⁸⁹¹⁰ Twinn’s public materials emphasise digital twins that connect physical assets, processes, and resources into a single simulation model to stress-test designs and policies before implementation, with vertical pages for supply chain & logistics and food & beverage highlighting use cases such as warehouse throughput analysis, production scheduling, and inventory strategy testing.⁷¹¹¹² WITNESS remains primarily a model-driven, desktop-centric simulation environment; historical and operational data are used, but typically as parameters and distributions inside user-defined models rather than as the core driver of fully automated optimisation pipelines.

Lanner vs Lokad

Although both Lanner (Twinn) and Lokad situate themselves in the broad domain of “supply chain and operations optimisation,” their approaches, architectures, and typical use cases are fundamentally different.

Model-centric simulation vs. data-centric optimisation. Lanner’s WITNESS is a discrete-event simulation environment: the core artefact is an explicit process model built by analysts, with entities, resources, queues, routing logic, and statistical distributions for arrivals and processing times.⁴¹³¹⁴ Users run many replications and scenario sweeps (via Experimenter/Optimizer) to observe performance metrics such as utilisation, throughput, waiting times, and service levels under different layouts, staffing levels, or operating rules.⁴⁶¹⁵ Lokad, by contrast, is a cloud-native, data-centric platform that ingests large tables of historical transactions, inventory positions, and master data, then generates probabilistic demand forecasts and optimised replenishment/production/pricing decisions via a domain-specific language (Envision) and stochastic optimisation algorithms. Lokad models uncertainty primarily through forecast distributions over demand and lead times, not by building explicit event-level process flows.

Project-style what-if studies vs. daily decision pipelines. WITNESS is typically used in project contexts: designing a new factory, re-configuring a warehouse, validating S&OP capacity plans, or stress-testing a proposed change in operating rules. Twinn case studies show WITNESS models used to evaluate line configurations for Mars Chocolate North America, to test new layouts and control logic for Carrefour distribution centres around Paris, and to design an eco-warehouse for Italian cosmetics brand L’Erbolario.¹⁶¹⁷¹⁸¹⁹ These studies run off-line, often with dedicated simulation specialists, and results are reported as recommended designs or policies. Lokad’s deployments, by design, run as recurring batch pipelines (typically daily): they recompute forecasts and optimisation outputs from up-to-date operational data and generate prioritised decision lists (purchase orders, stock transfers, pricing moves) that can be pushed into ERP/WMS systems. Where Lanner helps decide how a system should be structured and operated, Lokad aims to decide what to buy, move, and price today given that structure.

Granularity of uncertainty handling. In WITNESS, uncertainty is usually represented through classical DES input modelling: fitted probability distributions for arrivals, service times, breakdowns, and other stochastic elements, feeding Monte-Carlo simulations of the process.¹⁴²⁰ The focus is on system-level performance metrics (throughput, queues, utilisation). Lokad’s emphasis is on probability distributions of demand and supply at SKU × location × time granularity, with economic drivers (holding cost, stock-out penalties, obsolescence) used to compute the expected financial outcome of each decision. Lanner’s optimisation is largely oriented toward system design and parameter tuning (e.g., buffer sizes, staffing levels) via Experimenter, whereas Lokad’s optimisation is oriented toward daily stock levels and allocations under uncertainty.

Technology and user roles. WITNESS is a Windows-based simulation environment with drag-and-drop modelling, internal scripting, and the ability to call external code libraries such as C++, C#, VB.NET, or Python when needed.⁴⁷ It is aimed at industrial engineers, process analysts, and simulation specialists who are comfortable modelling flows and logic. Lokad is accessed via a web application; its Envision DSL is used by “supply chain scientists” to express data transformations, forecasting models, and optimisation logic, while planners interact mainly through dashboards and decision lists. Both require specialist skills, but WITNESS skills skew toward DES modelling, whereas Lokad skills skew toward data engineering and quantitative optimisation.

Scope in supply chain. Twinn’s supply chain & logistics marketing emphasises operational design and improvement of warehouses, factories, and logistics flows—capacity planning, bottleneck analysis, labour planning, vehicle flow, etc.¹¹ Case studies illustrate WITNESS models of DC operations, transport hubs, and production lines, but do not describe end-to-end integrated demand forecasting, multi-echelon inventory optimisation, or large-scale SKU-level replenishment as core features.¹⁶¹⁷¹⁸²⁰ Lokad’s stated scope is specifically end-to-end supply chain decision optimisation (demand forecasting, replenishment, production planning, allocation, pricing) across very large assortments, focusing on probabilistic forecasts and decision ranking by expected economic value. WITNESS can certainly be used within supply chain projects (e.g., to design a DC or test an S&OP plan), but it is not, out-of-the-box, a drop-in replacement for a supply chain planning or inventory optimisation system in the Lokad sense.

In short, Lanner/Twinn and Lokad are complementary rather than substitutable: WITNESS is best seen as a general-purpose DES/digital twin environment for designing and stress-testing processes, whereas Lokad is a data-driven optimisation platform geared toward recurring, granular supply chain decisions.

Corporate history and ownership

Lanner Group’s roots lie in the West Midlands automotive industry. The company’s history traces back to BL Systems (British Leyland’s IT department), then ISTEL, then AT&T Istel; in 1978–1980 this group developed SEE WHY, cited as one of the first commercially available visual interactive simulation tools.¹⁶¹³ Following a management buy-out from AT&T Istel, Lanner Group Ltd was formed in 1996 (incorporated in 1995 under earlier names such as PINCO 741 and SEEWHY Solutions), headquartered in Henley-in-Arden and later Birmingham.¹²¹⁰ WITNESS evolved from the SEE WHY lineage, with an IBM PC version launched in 1986 and multiple revisions since.¹²⁸

Lanner’s products expanded beyond WITNESS into niche packages such as PRISM (policing) and PX-Sim (healthcare), and into L-Sim, a Java-based simulation engine for embedding DES into BPM tools.¹²⁵¹³ From 1996 to 2010, private equity firm 3i was a main investor; in 2010 NVM Private Equity invested £3m, replacing 3i while the latter kept a minority stake.¹² In January 2019, Lanner Group Ltd was acquired by Royal HaskoningDHV, an international engineering and consultancy firm.⁹¹⁰ M&A databases describe Lanner at that point as a predictive simulation specialist whose technology connects physical assets, processes, and resources into a single digital model for resilient operations and supply chains.³⁹ Post-acquisition, Lanner’s products have been integrated into Royal HaskoningDHV’s Twinn digital twin brand, with Lanner Group Ltd remaining as a legal entity based in the UK (Companies House lists it as active, with SIC codes under business software development and IT services).⁷¹⁰

Commercial databases (D&B, Tracxn, Mergr) show Lanner as a small to mid-size software firm—dozens rather than hundreds of employees—with a long track record and global distribution network but not a hyperscale SaaS player.²⁴¹⁹ Overall, Lanner is best characterised as a mature, niche simulation vendor now embedded in a larger engineering consultancy.

Product portfolio and deployment model

WITNESS predictive simulation & digital twin platform

WITNESS is Lanner’s core discrete-event simulation environment. Product descriptions emphasise its ability to build 2D/3D animated models of factories, warehouses, transport systems, and service processes, with built-in elements such as machines, buffers, conveyors, vehicles, labour, and paths.⁴¹³¹⁴ The tool supports:

Discrete-event, continuous, and hybrid simulation world-views, with time-advance, random sampling, and statistics collection mechanisms typical of DES engines.¹³¹⁴
Graphical model building via drag-and-drop elements combined with WITNESS code for logic, routing, and control rules.⁴¹⁴
Input modelling and statistics, including random number generators, distribution fitting, and data import from databases or spreadsheets.¹⁴
Output analysis with charts, cost tracking, scenario management, and documentation tools.¹⁴

The Experimenter (and formerly Optimizer) module allows users to define sets of scenarios—combinations of parameter values such as buffer sizes, staffing levels, or scheduling rules—and run parallel replications to compare performance metrics.⁶¹⁵²¹ Third-party literature shows WITNESS used with design of experiments (DOE), Taguchi methods, and meta-heuristic approaches such as genetic algorithms: some studies link external genetic algorithm engines to WITNESS models to optimise manufacturing processes or assembly line balancing.⁷⁸²²²¹ This confirms that WITNESS provides mechanisms for experimentation and optimisation, but via classical simulation-plus-search setups rather than native large-scale mathematical programming.

Recent versions of WITNESS marketed under WITNESS Horizon highlight modernised UI, improved Experimenter, support for parallel runs, and connectivity to external code libraries (C++, C#, VB.NET, Python), giving users flexibility to embed custom logic or integrate with other systems.⁴⁶¹⁵ Release notes and product news emphasise usability and incremental performance improvements; there is no evidence of a radical shift toward cloud-native, multi-tenant SaaS or integrated machine-learning pipelines within WITNESS itself.⁶¹⁵

L-Sim: embedded simulation engine for BPM/BPSim

L-Sim is a Java-based simulation engine derived from WITNESS technology, designed to embed process simulation into BPM and BPSim-compliant tools.¹⁵²³ The 2006 Winter Simulation Conference paper on L-SIM describes it as a purpose-built DES engine for BPMN models, focusing on executing process models defined in BPMN and BPSim specifications, with features such as:

Reading BPMN/BPSim models and mapping them to simulation semantics.
Handling events, queues, and resources according to BPSim parameters.
Producing performance metrics like utilisation, cycle times, and bottleneck identification.⁵

Commercial integrations include:

IDS Scheer’s ARIS Business Simulator, where L-Sim acts as the simulation engine for BPMN models in ARIS.⁵¹³¹¹
Sparx Systems’ Enterprise Architect MDG BPSim Execution Engine, which uses L-Sim under the hood to execute BPSim scenarios.²³²²
Other BPM tools that leverage L-Sim via the BPSim standard for process simulation.⁷⁸²²

This architecture reinforces Lanner’s positioning as an engine provider: WITNESS is the modelling environment; L-Sim is the embedded engine that third-party BPM tools can use to simulate processes in a standardised way.

Deployment patterns

Public information suggests that WITNESS is primarily deployed as a desktop or client-server Windows application, with licensing and local installation, possibly complemented by server-side farms for running large Experimenter campaigns.⁴⁶¹⁵¹⁴ Twinn’s marketing emphasises cloud and digital twin narratives at the portfolio level, but does not provide detailed, independent evidence of WITNESS operating as a fully multi-tenant SaaS platform akin to modern web-native analytics tools. The prevalent deployment model remains:

Simulation specialists build and run models locally or on internal servers.
Data is imported from ERP/MES/WMS exports or databases.
Results are consumed via reports, dashboards, and presentations rather than via direct transactional automation.

This is consistent with the broader DES ecosystem, where tools like WITNESS, AnyLogic, and others are typically used in engineering projects rather than as 24/7 operational systems.

Use of Lanner in supply chain and logistics

Typical supply chain problems addressed

Twinn’s supply chain & logistics product page positions WITNESS as a tool to address:

Warehouse and distribution centre design, including conveyor systems, picking strategies, and automation investments.
Production and inventory planning across factories, focusing on capacity, buffer sizing, and scheduling decisions.
Transport and logistics flows, such as cross-dock operations and carrier/resource allocation.¹¹¹²

Academic and practitioner literature confirms WITNESS usage in:

Warehouse modelling, where researchers build WITNESS models of storage, picking, and replenishment to evaluate layouts and control policies.¹⁷²⁴
Airport check-in, petrol station queues, and other service logistics, to assess resource allocation and queue dynamics.¹³¹²
Manufacturing line balancing and layout optimisation in automotive and other industries.²²¹⁵

These use cases are classic DES applications: WITNESS models the process; parameters (arrival rates, service times, staffing levels) are varied; performance metrics are observed.

Evidence from named case studies

Lanner/Twinn provides several named client case studies relevant to supply chain and operations:

Mars Chocolate North America (MCNA) – WITNESS was used as a core engine in an S&OP analysis tool to evaluate production capacity, buffer sizing, and investment decisions. Twinn materials and third-party articles describe WITNESS helping Mars maximise capacity, reduce risk, and support network-wide planning decisions.¹⁶⁹ The focus is on strategic and tactical S&OP scenario analysis, not on daily replenishment optimisation.
Carrefour – A WITNESS-based “analysis and operations tool” was developed for Carrefour’s French distribution centres, integrating an Excel front-end with a WITNESS model to test operational scenarios such as conveyor speeds, staffing, and work-organisation rules.¹⁷¹⁹ This is a concrete example of WITNESS applied to warehouse operations in a major retailer.
L’Erbolario – The Italian cosmetics company used WITNESS to design and validate an “eco-warehouse,” assessing alternative layouts, automation investments, and operating rules to reduce environmental impact while maintaining service levels.¹⁸
Safran (aerospace) – Twinn case material highlights WITNESS usage in aerospace manufacturing and MRO contexts, for example to evaluate shop-floor flows and resource allocation, contributing to more robust production and maintenance processes.²⁰

These case studies confirm that Lanner has verifiable named clients across retail, FMCG, and aerospace, and that WITNESS is used for significant, high-stakes operational design projects. However, they also show that the tool’s role is to simulate and compare scenarios, not to automatically generate operational replenishment decisions or run as an embedded optimiser inside ERP/WMS.

From a supply chain-specific perspective, WITNESS is well-suited to:

Designing new facilities or re-engineering existing ones.
Testing S&OP and capacity-planning scenarios.
Exploring “what-if” questions about throughput, buffers, and service levels.

It is not, based on public evidence, marketed or deployed as:

A demand forecasting engine at SKU × location granularity.
A multi-echelon inventory optimiser.
An end-to-end supply chain planning suite with embedded transactional integration.

Technology stack, modelling paradigm and extensibility

Discrete-event simulation core

Technically, WITNESS implements the standard components of a modern DES engine: simulation clocks, event lists, random sampling, statistics accumulation, and animation.¹³¹⁴¹⁶ The “Process Simulation Using WITNESS” textbook provides a comprehensive description of WITNESS’ internal world-views (entities, resources, queues, activities) and modelling constructs (machines, buffers, conveyors, vehicles, labour, etc.).¹⁴ Key characteristics:

Time-driven by events: The engine advances the simulation clock to the next event (arrival, completion, breakdown, etc.), updating system state and statistics.
Randomness via RNG and fitted distributions: Input modelling tools fit empirical data to theoretical distributions (normal, exponential, Weibull, etc.) and feed them into the simulation.¹⁴
Hybrid discrete/continuous capabilities, allowing combined flow and DES models for certain processes (e.g., fluid tanks and conveyors).¹³¹⁴

This architecture is mature and standard in the DES field; the main differentiator for WITNESS is its long industrial track record and broad library of modelling components, not a fundamentally novel simulation algorithm.

Experimentation and optimisation features

WITNESS’ Experimenter and related optimisation tools provide an environment to:

Define design variables (e.g., number of operators, buffer capacities, shift patterns).
Specify performance metrics (e.g., throughput, WIP, waiting times, operating cost).
Run scenario sweeps and DOE (full or fractional factorial designs, Taguchi methods, etc.).
Use built-in heuristics and ranking to identify high-performing configurations.⁶¹⁵²¹

Third-party studies show how WITNESS is coupled with genetic algorithms or meta-heuristics to optimise manufacturing processes and assembly line layouts, with the simulation acting as the evaluation function for candidate solutions.⁷⁸²² This is a powerful pattern, but also a standard one in simulation-based optimisation: the key technical challenge is usually compute time and search strategy, rather than novel AI algorithms.

From a state-of-the-art AI perspective, these capabilities are robust but classical:

There is no public evidence of integrated deep learning, gradient-based differentiable programming, or end-to-end learning of policies inside WITNESS.
Optimisation appears to be driven by scenario sweeps, DOE, and external or internal heuristics, not by large-scale mathematical programming or reinforcement learning.

This does not diminish WITNESS’ practical value for design problems, but it means that when Lanner/Twinn uses terms like “predictive simulation” and “digital twin”, the underlying engine is still a DES/DOE stack, not modern ML-driven control.

Embedded engine and standards support (L-Sim, BPMN/BPSim)

L-Sim demonstrates a different aspect of Lanner’s tech stack: standard-based, embeddable engines. The WSC 2006 paper and vendor materials describe L-Sim as:

A Java-based simulation engine for BPMN models annotated with BPSim parameters.
Integrated into IDS Scheer’s ARIS Business Simulator and Sparx Enterprise Architect’s BPSim Execution Engine.⁵²³¹³²²
Focused on mapping BPMN constructs (activities, gateways, events) to DES semantics and delivering performance metrics for business process improvement.

This indicates solid engineering depth and a focus on interoperability with BPM standards—again, technically credible but not bleeding-edge AI/ML in itself.

AI, machine learning and optimisation claims

Twinn’s high-level marketing uses contemporary terms such as “predictive simulation,” “digital twin,” and “data-driven decision making,” with some references to AI within the broader Royal HaskoningDHV portfolio.⁷⁸¹⁸ However, when we inspect product-level evidence for WITNESS and L-Sim, the picture is more conservative:

Product pages emphasise discrete-event simulation, experimentation, and integration with external code libraries (C++, C#, VB.NET, Python), but do not detail built-in machine-learning models or end-to-end AI workflows inside WITNESS.⁴⁷¹⁵
Academic and textbook publications describe WITNESS as a DES tool with robust input modelling and output analysis, not as an ML platform.¹³¹⁴¹⁶
Optimisation is positioned around Experimenter, DOE, and scenario-based search; where genetic algorithms are used, they are typically external tools coupled to WITNESS as a simulator.⁷⁸²²²¹

In other words:

“Predictive” largely refers to the predictive nature of simulation—running forward in time under stochastic assumptions—rather than to predictive modelling in the ML sense.
“AI” references in the context of WITNESS are sparse and mostly marketing-level; we do not find detailed technical documentation of native deep learning, reinforcement learning, or large-scale optimisation algorithms inside the core product.

Given the available evidence, it is fair to characterise Lanner’s technology as state-of-the-practice for discrete-event simulation and digital twins, with strong academic grounding and industrial robustness, but not state-of-the-art in machine learning or AI-driven decision optimisation. Organisations looking for ML-heavy demand forecasting or algorithmic replenishment should see WITNESS as a complementary design and analysis tool, not as a substitute for specialised ML/optimisation platforms.

Commercial maturity and client base

Lanner has been in operation in its current form since the mid-1990s and has built a sizeable client base across multiple industries. Public materials and databases indicate:

Global distribution with subsidiaries or presence in the US, China, France, and Germany, plus distributors in many countries.¹²
Usage across automotive, aerospace, consumer goods, logistics, healthcare, and other sectors.¹²⁷
Named clients such as Mars Chocolate North America, Carrefour, L’Erbolario, Safran, and various manufacturers and service organisations.¹⁶¹⁷¹⁸²⁰

As part of Royal HaskoningDHV, Lanner’s software is now often delivered within consulting engagements, leveraging the parent company’s engineering and advisory teams.¹⁸⁹¹⁰ Commercial intelligence platforms classify Lanner as an acquired, mature niche vendor with many competitors in the simulation and digital twin space.³²³¹⁹ It is not an early-stage startup: its technology is well-established, widely taught in universities, and embedded in third-party tools. At the same time, it does not have the scale or breadth of an ERP or APS mega-vendor.

From a buyer’s perspective, Lanner/Twinn should be evaluated as:

A technically solid, specialised simulation vendor with decades of track record.
Primarily suited to organisations that value model-driven analysis (industrial engineering, process improvement, digital twin initiatives).
Less suited as a standalone solution for companies whose main need is ongoing, large-scale supply chain planning and optimisation.

Conclusion

Lanner (now Twinn within Royal HaskoningDHV) delivers a mature, technically credible suite of discrete-event simulation and digital twin tools—primarily WITNESS and L-Sim—used to design, analyse, and improve complex operational systems. The core strengths of the offering are clear and well-evidenced:

A long-standing DES platform (WITNESS) with rich modelling constructs, input/output analysis, Experimenter/Optimizer features, and extensive academic and industrial usage.⁴¹³¹⁴²¹¹⁶
An embeddable engine (L-Sim) integrated into BPM/BPSim tools such as ARIS and Sparx Enterprise Architect, showing strong interoperability with process-modelling standards.⁵²³²²
Demonstrated applications in high-profile supply chain-adjacent contexts—Mars, Carrefour, L’Erbolario, Safran—where simulation models supported major design and policy decisions.¹⁶¹⁷¹⁸²⁰¹⁹

At the same time, a rigorous, sceptical reading of the evidence highlights important limitations and clarifications:

Lanner’s technology is model-centric simulation, not data-centric probabilistic optimisation. It excels at scenario analysis and design, not at automated, recurrent decision-making at SKU × day granularity.
The “predictive” and “digital twin” labels are justified in the DES sense but should not be conflated with modern ML-driven forecasting or control; AI claims at the WITNESS level are sparse and not backed by detailed technical documentation of native ML algorithms.
Optimisation capabilities are strong in the classical DES/DOE sense—Experimenter, heuristic search, external meta-heuristics—but do not constitute state-of-the-art integrated stochastic optimisation engines for large-scale supply chain decisions.

Commercially, Lanner is a mature, stable specialist vendor embedded within a larger engineering firm. For organisations planning to use WITNESS or L-Sim in supply chain contexts, a realistic framing is:

Use WITNESS to design and stress-test warehouses, factories, S&OP plans, and logistics processes, exploring how structural changes affect performance under uncertainty.
Combine it with separate data-driven forecasting and optimisation tools if the goal is to automate day-to-day replenishment or pricing decisions.

When compared with platforms like Lokad, Lanner sits in a different technical niche: highly capable for digital twin and process design via simulation, but not a direct replacement for probabilistic, ML-driven supply chain decision optimisation. Buyers should resist marketing buzzwords and evaluate Lanner on what its technology demonstrably does best: robust discrete-event simulation and scenario experimentation for complex operations.