Review of GEP, Supply Chain and Procurement Software Vendor

By Léon Levinas-Ménard

Last updated: November, 2025

GEP is a US-headquartered procurement and supply chain software vendor that grew out of a consulting and outsourcing business and now sells a broad cloud suite spanning source-to-pay (GEP SMART), supply chain execution and planning (GEP NEXXE), and cross-suite AI layers (QUANTUM low-code platform and MINERVA AI/ML engine) running on Microsoft Azure. Over roughly 25 years, GEP has accumulated a sizeable enterprise customer base, particularly in procurement, and more recently positioned itself as an “AI-powered” platform using Azure SQL Database, Azure Marketplace deployment and Azure OpenAI integrations for generative features. The suite is wide in functional scope – procurement workflows, supplier collaboration, logistics visibility, control-tower-style monitoring and some planning – but public technical materials describe mainly workflow automation, analytics and generative assistance, with limited detail on the underlying forecasting and optimization algorithms that drive actual supply chain decisions. Taken at face value, GEP’s technology looks commercially mature and cloud-native, but technically closer to an enterprise process and visibility layer with AI-enhanced analytics than to a deeply quantitative optimization engine for inventory, capacity and pricing.

GEP overview

GEP traces its origins to 1999 as Global eProcure, founded by a group of Indian-born entrepreneurs and headquartered in Clark, New Jersey, combining strategic sourcing consulting with outsourced procurement operations and early e-sourcing tools.¹ Over the 2000s the company expanded through global delivery centers in India and elsewhere, growing primarily as a services business that also licensed proprietary technology to support sourcing and spend management.¹² Around the early 2010s Global eProcure rebranded as GEP and progressively repositioned itself as a unified software + services provider, culminating in the launch of SMART by GEP, later shortened to GEP SMART, as a single, cloud-based source-to-pay (S2P) platform.¹³

To expand beyond procurement into broader supply chain execution and planning, GEP made several acquisitions. In 2012 it acquired Enporion, a US-based B2B marketplace and supply chain management platform serving energy and utilities, in a deal reported by FreightWaves and confirmed by GEP’s own press release.⁴⁵ In 2024 GEP acquired OpusCapita’s procurement, e-invoicing and AP automation business, adding Nordic/European customers and technology focused on purchase-to-pay and invoice processing.⁶⁷ Secondary sources such as Owler list additional acquisitions (for example COSTDRIVERS and DATAMARK) that broaden analytics and managed services capability, though these are poorly documented in public technical materials.⁸

Today, GEP presents itself as an integrated provider of consulting, managed services (BPO) and GEP SOFTWARE, the umbrella brand covering GEP SMART (S2P), GEP NEXXE (supply chain) and AI/low-code layers GEP MINERVA and GEP QUANTUM.¹⁹ Analyst coverage (e.g., Spend Matters) describes GEP as a “hybrid” vendor: part strategy consultancy, part outsourced procurement operator, part software company.³ Commercially, GEP claims hundreds of large enterprise clients across CPG, pharma, manufacturing, financial services and energy, managing tens or hundreds of billions of spend – figures repeated across company profiles on Everipedia and Umbrex, but not independently audited.¹²

Gartner’s 2025 Magic Quadrant for Source-to-Pay Suites placed GEP in the Leaders quadrant, citing its unified S2P vision and strong customer traction in complex global organizations.⁹ However, this recognition is confined to the S2P domain; there is no equivalent Magic Quadrant entry for GEP in core supply chain planning. Overall, public evidence indicates GEP is a commercially established vendor in procurement technology with growing ambitions in supply chain, but much less externally validated in the latter domain.

GEP vs Lokad

GEP and Lokad both address “supply chain” problems but from fundamentally different angles. Comparing them directly is only meaningful if these structural differences are made explicit.

Scope and product shape. GEP sells a broad enterprise suite: GEP SMART for source-to-pay, GEP NEXXE as a supply chain platform, and cross-suite AI/low-code layers (MINERVA, QUANTUM), plus substantial consulting and managed services. Its sweet spot is end-to-end procurement workflows (sourcing, contracts, supplier management, invoicing) with adjacent supply chain visibility and collaboration. Lokad, by contrast, is a narrow but deep platform focused almost exclusively on quantitative supply chain planning and optimization – demand forecasting, inventory and capacity decisions, some pricing – and does not attempt to replace S2P, ERP or WMS systems. Lokad must integrate with whatever procurement/ERP stack a client already has (which may include GEP); GEP aims to be the transactional backbone itself.
Architecture and programmability. GEP’s suite is a cloud-native, Azure-hosted application stack. SMART and NEXXE are offered via the Azure Marketplace, built on Azure SQL Database and other Azure services.¹⁰¹¹¹² Public material and engineer profiles indicate a microservices architecture with a low-code layer and plugin-style front-end, using common web technologies; NEXXE’s back-end is described as microservices with saga orchestration, and the platform is explicitly positioned as low-code/no-code for workflow customization.¹³¹⁴ Lokad, on the other hand, runs an in-house tech stack centered on its domain-specific language Envision, backed by an event-sourced store and a distributed VM. The platform is programmable first: every forecast and optimization is code, rather than configured via low-code UI. This makes Lokad closer to a supply-chain-specific analytics engine; GEP is closer to a generalized enterprise application suite extended by low-code.
Decision model and “AI”. GEP’s AI narrative emphasizes MINERVA – an AI/ML engine that now integrates Microsoft’s generative AI via Azure OpenAI – and QUANTUM, a low-code platform marketed as “AI-first”.¹⁵⁹ In supply chain, NEXXE promises “advanced AI and ML to solve real-world supply chain problems”, predictive insights and closed-loop planning.¹¹¹³ However, public sources mainly describe use-cases like conversational querying, document summarization, classification, anomaly detection and generic prediction; there is little technical detail on probabilistic demand modeling, multi-echelon inventory optimization or combinatorial scheduling algorithms. Lokad, conversely, is explicitly built around probabilistic forecasting and stochastic optimization (Monte Carlo, gradient-based and heuristic methods) with published evidence such as performance in the M5 competition and detailed technical write-ups. In other words, GEP’s AI appears primarily as assistive analytics and automation layered onto workflows, while Lokad’s “AI” is mostly under-the-hood mathematical machinery that directly generates optimized decisions.
Outputs: workflows vs. prioritized decisions. GEP SMART and NEXXE are heavily workflow-centric: user journeys revolve around sourcing events, contract approval, supplier onboarding, supply chain exception handling, control-tower views and scenario simulations.¹⁰¹¹ The system may generate recommendations (e.g., suppliers to approach, inventory policies, logistics plans), but these sit inside broader process flows and governance structures. Lokad’s primary deliverable is a ranked list of decisions – purchase orders, stock transfers, production batches, pricing moves – each annotated with expected financial impact under uncertainty. Lokad leaves procurement workflows and approvals to other systems; GEP embeds decisions inside its own processes.
Delivery model. GEP often sells large transformation programs mixing software, consulting and BPO; Umbrex notes GEP’s positioning as a “full-service procurement transformation partner” with long-term outsourced operations for some clients.¹ Lokad typically operates with small “supply chain scientist” teams building and maintaining Envision scripts on top of a client’s data, without taking over operational procurement. For a company wanting to outsource parts of procurement and standardize processes globally, GEP is structurally aligned; for a company wanting a specialized optimization brain that plugs into an existing ERP/S2P landscape, Lokad is closer to that role.

In short, GEP is a suite vendor with AI-enhanced enterprise workflows, while Lokad is an optimization platform that assumes someone else handles transactional processes. For evaluating supply chain decision technology, the relevant comparison is not UI sophistication or S2P breadth, but the depth and transparency of decision modeling; on that axis, GEP’s public materials are still thin compared with Lokad’s technical disclosures.

Corporate history and acquisitions

GEP’s early history is relatively well documented by third-party profiles. Umbrex reports that Global eProcure was founded in 1999, initially providing procurement consulting and outsourcing services, before gradually building its own technology to support strategic sourcing and spend analytics.¹ Everipedia similarly describes GEP as evolving from a boutique consultancy into a global provider of procurement solutions with offices across North America, Europe and Asia, and highlights its growth in managed services alongside software.²

Acquisitions appear to have been used primarily to expand domain coverage and geographic reach:

Enporion (2012). In January 2012, GEP (then still often referred to as Global eProcure) announced the acquisition of Enporion, a US-based provider of supply chain management software and electronic marketplaces for the energy and utilities sector.⁴⁵ FreightWaves reported the deal as a strategic move to deepen GEP’s capabilities in utilities and to gain Enporion’s hosted procurement platform.⁴ GEP’s own press statement (PDF) frames Enporion as bringing “advanced supply chain management solutions” and an established marketplace network.⁵ There is little public technical detail on how Enporion’s technology was integrated or retired; GEP’s later product branding (SMART, then NEXXE) suggests a gradual consolidation into a single cloud stack rather than maintaining multiple code bases.
Other acquisitions (DATAMARK, COSTDRIVERS). Owler’s company profile lists further acquisitions including DATAMARK and COSTDRIVERS, but without transaction dates or technical detail; these appear to be primarily aimed at expanding analytics, data and BPO capabilities, not at creating separate product lines.⁸ Given the lack of corroborating sources, these should be treated as indicative rather than exhaustively documented.
OpusCapita procurement business (2024). In July 2024, GEP announced the acquisition of the procurement and AP automation software business of OpusCapita, a Finland-based provider of procurement, e-invoicing and accounts payable solutions.⁶⁷ Press releases state that OpusCapita’s products and customers will be integrated into GEP SMART, with GEP emphasizing expansion of its European footprint and strengthened e-invoicing/compliance capabilities.⁶⁷ No technical migration roadmap is publicly available; based on typical S2P consolidation patterns, one should expect OpusCapita’s features to be progressively folded into SMART while standalone branding is phased out.

Overall, the acquisition pattern is consistent with a platform-build strategy: buy domain-specific assets (utilities marketplaces, Nordic e-invoicing), then subsume them into a unified, Azure-based suite.

Product portfolio and positioning

GEP SMART: unified S2P suite

GEP SMART (often stylized “SMART by GEP” in older materials) is the company’s flagship source-to-pay platform. Microsoft’s Azure Marketplace describes it as a “unified procurement platform” covering spend analysis, sourcing, contract management, supplier management, savings project tracking, procurement, invoicing and category workbench functions, delivered as a multi-tenant cloud service running on Azure.¹⁰ GEP positions SMART as a single, integrated platform rather than a collection of loosely-coupled modules; Spend Matters’ vendor snapshot supports this, noting that SMART was architected as a unified suite built natively for the cloud, in contrast to legacy on-premise tools that were later refactored for SaaS.³

Feature-wise, SMART’s capabilities largely align with modern enterprise S2P expectations: advanced sourcing events, guided buying, contract repository with obligation tracking, supplier onboarding and performance scorecards, catalog management, P2P workflows and invoice matching. Case studies from GEP’s marketing (e.g., a SMART implementation for a global manufacturer) describe deployments with hundreds of users, multi-year historical data migration, and a dedicated implementation team combining GEP subject-matter experts with client stakeholders.¹¹¹⁴ One published example mentions migrating three years of historical data and enabling simultaneous access for ~300 users on a unified source-to-contract platform, which is in line with a typical enterprise SaaS rollout rather than an experimental or immature system.¹⁴

From a technology standpoint, SMART is tightly coupled to the Microsoft stack. A Microsoft Azure blog on SQL Database elastic pools cites GEP as a reference customer, noting that GEP migrated more than 800 databases into Azure SQL Database elastic pools, shut down its own data centers, and now operates as a “datacenter-free” company, with SMART by GEP described as a cloud-based procurement and supply chain solution built on Azure SQL.¹² This is strong evidence that SMART runs on Azure SQL Database in a multi-database, elastic-pool configuration, which fits well with a multi-tenant SaaS platform serving many customers.

In more recent materials, SMART is often described as being “powered by GEP MINERVA™ AI” – implying that the AI/ML engine sits underneath or alongside SMART to drive classification, predictions and generative features – but technical details on the exact algorithms used (e.g., for spend classification, supplier risk scoring or fraud detection) are not publicly disclosed. We therefore treat claims of AI-driven optimization in SMART as partially substantiated (Azure OpenAI integration is real; the internal ML models remain opaque).

GEP NEXXE: supply chain platform

GEP NEXXE is the company’s supply chain platform, marketed as a “cloud-native unified supply chain platform” for end-to-end visibility, collaboration and planning.¹¹¹⁶ The Azure Marketplace listing describes NEXXE as supporting demand planning, supply planning, inventory optimization, logistics, and “control-tower” style capabilities, and highlights “advanced AI and ML to solve real-world supply chain problems” as a selling point.¹¹¹³

Independent software comparison sites (e.g., eBool) characterize NEXXE as combining real-time visibility across tiers, predictive analytics for risk and disruption, and a low-code/no-code design that allows users to tailor workflows and dashboards, emphasizing flexibility in building supply chain applications without deep coding skills.¹³ Technical profiles of GEP engineers mention microservices and saga orchestration in the NEXXE back-end and a plugin-based front-end, affirming a modern, distributed architecture consistent with that low-code positioning.¹⁴

However, public technical detail on NEXXE’s core optimization logic is limited. Marketing and analyst-style descriptions reference:

real-time visibility and alerts (typical of control-tower systems),
predictive risk analytics (e.g., anticipating supply disruptions),
multi-echelon inventory planning and scenario analysis,
AI/ML-based demand sensing and predictions.

What is not described in any concrete way is:

how demand forecasts are modeled (classical time-series vs. ML vs. probabilistic distributions),
what objective functions underlie “inventory optimization” (service level, cost, profit),
whether uncertainty is handled via full probability distributions or simplified safety-stock heuristics,
what algorithms are used for complex combinatorial decisions (e.g., network flows, scheduling).

Case studies on GEP’s site mention results like “40% reduction in inventory” or “30% improvement in order fulfillment” attributable to NEXXE, but these narratives remain qualitatively focused and do not expose the mathematical mechanism or provide enough detail for independent replication.³ From a skeptical technical perspective, NEXXE today looks like a modern, Azure-native control-tower and collaboration layer with embedded analytics, rather than a transparently specified optimization engine.

GEP QUANTUM and MINERVA

GEP QUANTUM is marketed as an “AI-first, low-code platform for procurement, supply chain and sustainability applications”. An AIThority article on its launch describes QUANTUM as providing composable building blocks, integrated AI engines and a visual low-code environment to rapidly assemble new applications atop GEP SMART and NEXXE, targeting both professional developers and “citizen developers”.¹⁵ QUANTUM is therefore best understood as platform glue and extensibility, not a standalone optimization product.

GEP MINERVA is the AI and machine learning engine that underpins both SMART and NEXXE. A 2023 MarketScreener/Microsoft press release states that GEP launched a suite of solutions within GEP SOFTWARE that use OpenAI’s ChatGPT via Microsoft Azure OpenAI Service, offering an intuitive interface to query data, automate processes and improve decision-making; it further notes that Microsoft’s generative AI capabilities are incorporated into the GEP MINERVA AI and ML engine to deliver cross-organizational data analytics and decision support.⁹

Taken together, this indicates that GEP’s AI stack is built around:

a proprietary AI/ML engine (MINERVA) handling classic ML tasks (classification, clustering, predictive models),
a low-code platform (QUANTUM) to expose these capabilities in apps,
Azure OpenAI for generative functions (conversational interfaces, summarization, document understanding).

What remains unclear is how far this AI stack is used for prescriptive optimization in supply chain, beyond descriptive and predictive analytics. There is no public discussion of, for example, training objective functions around end-to-end cost or profit, joint learning of forecasts and decisions, or advanced constrained optimization akin to operations research solvers.

Architecture and technology

Public information points to a Microsoft-centric, Azure-native architecture for GEP’s suite:

Azure SQL and “datacenter-free” operations. In Microsoft’s blog announcing the general availability of Azure SQL Database elastic pools, GEP is cited as a SaaS customer that migrated over 800 databases into elastic pools and closed its own data centers, with a GEP technology VP noting that this move made GEP “datacenter-free” and delivered significant cost savings.¹² SMART by GEP is explicitly referenced as a cloud-based procurement and supply chain solution built on Azure SQL Database, establishing that relational data is stored in Azure SQL and that multi-tenant database provisioning is handled via elastic pools rather than self-managed servers.
Azure Marketplace deployment. Both GEP SMART and GEP NEXXE are listed in the Microsoft Azure Marketplace as “Unified Procurement Platform – GEP SMART” and “Unified Supply Chain Platform – GEP NEXXE”, respectively, emphasizing cloud-native deployment, scalability and global availability in Azure regions.¹⁰¹¹ This suggests that at least some customers procure the software as a SaaS subscription through Microsoft’s marketplace, though direct contracts with GEP remain the norm for large transformation deals.
Microservices and low-code. Engineer profiles and third-party descriptions give hints about the internal architecture. A senior software engineer describes working on GEP NEXXE using microservices and saga patterns on the back-end and a plugin-style front-end, and contributing to making NEXXE a low-code platform.¹⁴ Software comparison sites emphasize a low-code design for NEXXE that lets users customize workflows and dashboards.¹³ Together, these indicate that GEP has implemented a microservices-based application tier with a low-code layer for composing UI and workflow components, consistent with current enterprise SaaS practice.
AI integration via Azure OpenAI. As noted, GEP’s MINERVA engine is integrated with Azure OpenAI to provide generative AI capabilities within the suite.⁹ This implies a service-oriented AI layer: application services call out to Azure OpenAI models for text generation, summarization and classification, while proprietary models run elsewhere in the stack.

What is missing is any low-level description of:

the programming languages and frameworks used (likely .NET/JavaScript, but not explicitly stated),
data modeling patterns beyond “running on Azure SQL Database”,
the internal design of QUANTUM and MINERVA (e.g., whether they use Kubernetes-hosted microservices, what ML libraries are employed),
how tenant isolation and multi-region deployment are handled.

Given GEP’s size and Azure reference status, it is reasonable to infer a technically competent, mainstream enterprise SaaS architecture, but there is no evidence of unusual or pioneering infrastructure akin to a custom DSL or event-sourced analytics engine. This is not a criticism – most enterprise buyers prefer conventional stacks – but it does mean GEP’s distinctiveness lies more in suite breadth than in unusual architectural innovation.

AI, machine learning and optimization: claims vs evidence

GEP’s marketing leans heavily on AI, ML and now generative AI. A critical review must distinguish between:

well-substantiated AI capabilities (where behavior and implementation are reasonably clear),
plausible but unproven claims (aligned with norms but not technically detailed),
ambiguous or potentially overstated claims (where “AI” could mask basic analytics).

Well-substantiated capabilities

Integration with Azure OpenAI. The MarketScreener/Microsoft article provides concrete evidence that GEP integrated OpenAI’s ChatGPT via Azure OpenAI Service, enabling conversational querying of procurement and supply chain data, process automation and decision support within GEP SOFTWARE.⁹ It explicitly states that generative AI capabilities are incorporated into GEP MINERVA’s AI/ML engine. This confirms real use of large language models (LLMs) for text-heavy tasks: querying, summarizing, and possibly document interpretation.
Wide deployment of AI-enhanced suite. The same article notes that GEP SOFTWARE is used in 120 countries and available in the Azure Marketplace, indicating that these AI features are embedded into a mature suite rather than being experimental add-ons.⁹

Plausible but weakly specified capabilities

Predictive analytics and ML inside SMART and NEXXE. NEXXE’s Azure Marketplace description refers to “advanced AI and ML to solve real-world supply chain problems”, covering demand sensing, risk prediction and anomaly detection.¹¹ Marketing and case studies describe predictive models for supplier risk, demand anomalies and logistics disruptions in general terms. It is highly plausible that GEP has built supervised ML models for classification and regression in these domains (e.g., demand uplift forecasts, risk scoring), but without details on features, model types or evaluation metrics, the technical sophistication is unknown. At a minimum, this appears to be standard enterprise ML rather than cutting-edge research.
Spend classification and data enrichment. Procurement suites commonly use ML to classify spend into taxonomies, deduplicate supplier records and recommend category mappings. Given GEP’s long history in spend analysis and multiple references to AI-powered classification in marketing, it is reasonable to assume such models exist, but again, specifics are not public.

Ambiguous or overstated claims

Most provocative, from a supply chain optimization standpoint, are claims around “inventory optimization”, “supply chain optimization” and “closed-loop planning” driven by AI. NEXXE’s materials mention inventory optimization, scenario planning and closed-loop planning, but do not specify:

whether inventory decisions are optimized over probability distributions of demand and lead time,
what objective is optimized (e.g., expected total cost, service level, profit),
whether optimization uses mathematical programming, heuristics, or rule-based scripts.

Given the absence of technical documentation, open-source models or peer-reviewed references, it is safest to assume that NEXXE implements a combination of rule-based heuristics and conventional forecasting, wrapped in a modern UI and enhanced with predictive ML and generative AI for analytics and collaboration. Until GEP publishes more technical details, claims of “AI-powered inventory optimization” should be treated as partially substantiated marketing language, not proof of advanced operations research.

In contrast, Lokad publicly documents probabilistic forecasting, specialized optimization algorithms and even academic work in differentiable programming; this transparency is precisely what is lacking in GEP’s AI story, making it difficult to judge how state-of-the-art GEP’s optimization truly is.

Deployment, roll-out and usage

Public case studies provide some insight into how GEP deploys its software:

Project-style implementations. A SMART case study for a global manufacturing company describes a deployment where GEP’s technology experts “helped the client deploy a full source-to-contract platform,” migrated three years of historical data, and rolled the system out to around 300 users, supported by dedicated GEP SMEs and account managers.¹⁴ This is consistent with multi-month implementation projects typical of enterprise S2P suites: data migration, configuration, training and change-management, rather than a pure self-service SaaS adoption.
Consulting and BPO integration. Umbrex emphasizes that GEP often combines its software with consulting and managed services, positioning itself as a transformation partner rather than a “software only” vendor.¹ This implies that successful deployments frequently include ongoing service contracts, with GEP staff embedded in or closely collaborating with client procurement and supply chain teams.
NEXXE use-cases. Case studies for NEXXE reference supply chain control tower implementations, inventory reductions and improvements in OTIF (on-time in-full) for large manufacturers and pharma companies.³ The anecdotes suggest that NEXXE is used to aggregate data from multiple ERPs and logistics systems, provide near real-time visibility, and coordinate responses to disruptions. However, they do not spell out the decision logic or how conflicting objectives (service vs. cost vs. risk) are arbitrated.

Compared with Lokad’s deployment model – iterative Envision script development led by supply chain scientists – GEP’s roll-outs look closer to classic enterprise platform implementations: larger cross-functional projects, deeper process standardization, and a heavier focus on governance, training and change-management.

Commercial maturity and client base

Multiple sources agree that GEP is commercially mature, particularly in procurement:

Umbrex and Everipedia cite hundreds of clients, including large enterprises across many sectors.¹²
Spend Matters’ vendor snapshot (2019) positions GEP as a “top-tier” S2P vendor with significant market presence and continued investment in SMART.³
Gartner’s 2025 Magic Quadrant for Source-to-Pay Suites places GEP in the Leaders quadrant, implying both completeness of vision and ability to execute in S2P.⁹

For supply chain, the picture is more nuanced:

NEXXE appears on Azure Marketplace and in customer stories, but there is no dedicated Gartner Magic Quadrant or Forrester Wave specific to NEXXE as a supply chain planning tool.
Case studies reference meaningful improvements (inventory reductions, OTIF gains), yet they lack enough specificity to discriminate between improvements stemming from better visibility and process discipline versus those from genuinely superior optimization algorithms.

From a skeptical standpoint, it is fair to say that GEP is an established player in procurement technology, and an emerging – but not yet clearly benchmarked – player in supply chain planning technology.

Assessment of technical merit and uniqueness

Answering the user’s key questions:

What does GEP’s solution deliver, in precise technical terms?

In procurement (GEP SMART): A cloud-hosted, Azure-based S2P application suite that manages master data, transactional documents and workflows across spend analysis, sourcing, contracts, supplier management, purchase-to-pay and invoicing. It stores structured data primarily in Azure SQL Database, uses web application front-ends and exposes process logic configurable via low-code tools. AI is used for classification, recommendations and generative assistance (e.g., chat, summarization).
In supply chain (GEP NEXXE): A supply chain visibility and collaboration platform aggregating data from ERPs, WMS and logistics systems; offering real-time dashboards, alerts and scenario analysis; and providing some degree of demand and inventory planning supported by generic ML and optimization routines. It is architected as a microservices-based, low-code platform on Azure, integrated with the MINERVA AI engine and Azure OpenAI for analytics and conversational features.
Cross-suite (QUANTUM & MINERVA): A low-code environment (QUANTUM) for building applications on top of SMART/NEXXE and an AI/ML engine (MINERVA) that centralizes predictive models and generative AI services.

Through what mechanisms and architectures does it achieve these outcomes?

Infrastructure: Azure SQL Database elastic pools, Azure Marketplace deployment, microservices and low-code UI frameworks, as evidenced by Microsoft Azure references, Azure Marketplace listings and engineer profiles.¹⁰¹¹¹³¹²¹⁴
Analytics and AI: Proprietary ML models (details undisclosed) for predictions and classifications; Azure OpenAI-hosted LLMs for conversational interfaces and text processing; QUANTUM low-code tooling for embedding these into workflows.¹⁵⁹
Decision logic: Business rules, heuristics and scenario-analysis tools implemented within SMART and NEXXE; some degree of optimization for inventory and planning is claimed but not technically specified. There is no public evidence of fully probabilistic end-to-end models, advanced combinatorial solvers or differentiable programming in the Lokad sense.

How state-of-the-art is GEP’s technology?

In infrastructure and application design, GEP appears up-to-date but not exceptional: Azure-native, microservices, low-code, generative AI via Azure OpenAI. This is consistent with current best practice among serious enterprise SaaS vendors, but not uniquely advanced.
In procurement process digitization, SMART is competitively strong and mature, as evidenced by Gartner Leader positioning and long-standing analyst coverage.⁹³
In supply chain decision optimization, public information suggests GEP is behind the frontier defined by specialized optimization vendors. There is scant evidence of rigorous probabilistic forecasting, advanced stochastic optimization, or transparent, mathematically-grounded decision engines. NEXXE’s strengths appear to be visibility, collaboration and analytics rather than groundbreaking optimization algorithms.

Commercial maturity

GEP is commercially mature in procurement (decades in market, large enterprise base, analyst recognition) and emerging but less validated in sophisticated supply chain planning. Its solutions are best understood as broad enterprise applications with AI-enhanced workflows, not as specialized quantitative engines.

Conclusion

GEP is a substantial, Azure-native vendor whose core strength lies in unifying procurement processes across global enterprises through GEP SMART and complementing this with supply chain visibility and collaboration via GEP NEXXE. The architecture is modern and credible: Azure SQL Database elastic pools, microservices, low-code UI, and a cross-suite AI layer leveraging Azure OpenAI. Its commercial position in S2P is well established and independently validated.

From a technical, optimization-centric perspective, however, GEP’s public story is much thinner. While MINERVA, QUANTUM and NEXXE are marketed as AI-powered and optimization-driven, available evidence points primarily to predictive analytics, generative assistance and workflow-centric tooling, not to deeply specified probabilistic models or advanced optimization engines. Supply chain decision-making appears to be a mix of rule-based heuristics, conventional forecasting and scenario planning, with AI improving insight and automation rather than fundamentally redefining the decision mathematics.

Relative to Lokad, GEP is a broad suite vendor with AI-augmented enterprise workflows, whereas Lokad is a narrow but deep optimization platform built around probabilistic forecasting and custom decision models. For organizations evaluating technology specifically to push the frontier of quantitative supply chain optimization, GEP’s supply chain stack currently lacks the technical transparency and evidence that would justify treating it as state-of-the-art in that niche. For organizations seeking a single vendor for procurement transformation with reasonable supply chain visibility and some AI-driven analytics, GEP’s offering is credible and mature – but should be understood as a process platform, not a pure optimization engine.