Go back to Market Research
QAD supplies software for manufacturing operations, including ERP, supply chain planning, purchasing, inventory and factory execution. Its planning suite, formerly marketed as QAD DynaSys, covers demand forecasting, procurement, production, distribution and S&OP. Redzone supports frontline factory work, while ChampionAI adds assistants and automated workflows around enterprise applications. (1, 7)
Overall supply chain score: 5.2/10
| Dimension | Score |
|---|---|
| Supply chain depth | 5.4/10 |
| Decision and optimization substance | 5.8/10 |
| Product and architecture integrity | 4.8/10 |
| Technical transparency | 5.2/10 |
| Vendor seriousness | 5.0/10 |
QAD has substantial manufacturing functionality. Its public planning documents describe actual purchasing and production calculations, including supplier calendars, transport loading, finite capacity and supply variability. The governing process nevertheless remains a combination of forecast-led plans, inventory policies, configured priorities and planner intervention. The documentation does not establish a common economic objective that evaluates procurement, production and inventory actions across uncertain futures. (2, 3, 4)
The new AI proposition tackles a different problem: handling documents, obtaining supplier updates, answering questions and understanding old customizations. Those activities consume real labor, but their automation does not establish better replenishment or production decisions. QAD’s broader Manufacturing Intelligence layer, announced with planned NVIDIA integration, is expected in spring 2027. Its prospective capabilities should not be confused with the workflows already described in production. (6, 9, 10)
QAD versus Lokad
QAD’s center of gravity is the manufacturing operating environment: transactions, material records, production resources, factory teams and the processes connecting them. Lokad’s quantitative supply chain approach centers on the economic decisions made over those records. It builds executable logic that evaluates choices such as purchases, production and allocations against probabilistic demand and supply, including lead times.
This difference becomes concrete when a supplier is late. QAD’s Procurement Champion follows communications and updates the order workflow. The economically difficult questions begin with whether to expedite, substitute, split an order, change production or accept a shortage. Answering them requires comparing downstream consequences and uncertain alternatives. Administrative automation can improve the information available for that calculation; it does not supply the calculation. (6)
Lokad’s inventory optimization service assigns Supply Chain Scientists to build and maintain the data interpretation and decision logic. This makes ongoing model improvement an explicit part of the work. QAD distributes comparable responsibilities across its applications, customer configurations, implementation services and operating teams. Its recent customization-analysis work addresses an important part of that burden, but the public results chiefly concern migration discovery. (9, 14)
The relevant test for both vendors is whether the decision system improves economically as its understanding of the business changes. A technically successful migration, a faster forecast cycle or a more productive planner can leave the original decision policy wrong. Lokad also has to demonstrate returns after its own engineering and service costs; its economic framing is not a substitute for customer results.
What the planning software actually calculates
Manufacturing and purchasing have substantive models
The 2019 Production Planning document describes a hybrid of cost-based methods and heuristics, with material, labor, tooling and resource constraints. It accommodates different production stages and sites, overtime, outsourcing, stock building, campaigns and shelf life. This is a developed manufacturing model with useful decisions behind it. The public explanation does not identify the mathematical formulation or expose the way competing costs are combined. (2)
Procurement Planning is similarly specific. Requirements may come from reorder points, days of supply, forecasts or DDMRP. Order generation then respects supplier calendars, quantity increments and purchasing constraints. Look-ahead logic can bring future requirements forward to fill transport capacity. These features coordinate several products and periods; they also create economic choices that cannot be settled by utilization alone. (4)
A fuller truck may cost less to move per unit while committing cash earlier, increasing expiration exposure and crowding out a more valuable purchase. A feasible production plan may consume a scarce input on a low-margin item. QAD’s cost and constraint handling addresses parts of these problems. The missing public explanation is how the entire decision is valued when the future quantities, timings and alternatives are uncertain.
The suite’s financial layer compares operational plans with budgets and exposes measures such as working capital and return on assets. Such reporting helps users judge scenarios, but a financial view of a plan does not establish that the numerical engine selected it by expected economic return. (5)
Variability enters the model, but the joint decision mechanism is unclear
QAD’s Demand Planning document explicitly includes forecast error, demand and supply variances, and lead time in inventory calculations. It also describes testing forecast accuracy at an offset corresponding to the decision time fence. These are relevant mechanisms: uncertainty is acknowledged, and forecasts can be evaluated at the time they were actually needed. (3)
The harder issue is the interaction between uncertainties. A late ingredient, variable yield and a demand surge can jointly change which production commitment is worthwhile. Safety-stock calculations based on variances do not establish a probability distribution over these combined outcomes, nor an economic ranking of the resulting alternatives. The public documents leave that connection unresolved.
The model’s segmentation rules and service-level targets also deserve scrutiny. They turn business preferences into inventory requirements before many downstream trade-offs are evaluated. A uniform target within a segment can hide large differences in marginal value, substitutability, expiration risk and customer consequences. More elaborate segmentation increases configuration work without necessarily repairing the economic approximation. (5)
Forecasting claims need the decision’s resolution
The historical demand brochure advertises forecast-accuracy improvement of as much as 90%. This is an improvement claim, not a statement that SKU forecasts achieve 90% accuracy. It still omits the metric, baseline, horizon and aggregation necessary to interpret the figure. Its demand-sensing label likewise supplies no comparative evaluation showing the contribution of additional signals at the relevant ordering deadline. (3)
For a manufacturer, the useful question is whether the revised forecast changes a purchase or production decision profitably. Aggregating products or extending time buckets can improve an accuracy percentage while erasing precisely the distinctions needed for that decision. QAD’s offset-testing feature is helpful, but the headline does not disclose results under that protocol.
ChampionAI: administrative automation and decision responsibility
QAD distinguishes a conversational assistant from background Champions. Its September 2026 release describes an accounts-payable workflow that reads invoices, matches them with orders and receipts, and routes discrepancies to people. Reported production figures include 90% automatic matching and 95% scanned-invoice extraction accuracy. These percentages concern document processing, not demand forecasting. (7)
The procurement product monitors order information, follows suppliers and escalates exceptions. Its potential cost-of-goods reduction is attributed to freeing buyers for strategic sourcing and negotiation. That is an indirect productivity proposition: it depends on how the released time is used and whether the resulting commercial decisions improve. A completed follow-up is not itself a reduction in procurement cost. (6)
QAD’s ChampionAI campaign nevertheless describes the customer’s team reviewing every action before it goes out. The January 2026 cloud program also supplies AI outputs on an as-is basis and requires appropriate human review before business reliance. The program covers manufacturing operations and supply chain cloud services; Connected Workforce has separate terms. (8, 14)
This leaves an important operational question behind the autonomy messaging: which classes of action are accepted automatically, which are sampled, and which require individual review? The answer determines whether labor scales with exceptional events or with the full transaction stream. It also determines who can correct a harmful decision policy.
The engineering vacancy for the ChampionAI platform identifies MCP infrastructure, prompts, tool bindings and agent rollout as actual development work. Connecting language models to enterprise tools can remove interface friction. It also preserves the old workflow if the agent merely operates existing screens, messages and queues more efficiently. The public evidence does not show that this orchestration layer has replaced the planning suite’s underlying objective or uncertainty model. (11)
Accounts payable supplies another useful economic test. Earlier invoice completion can capture discounts, but paying earlier also consumes cash sooner. The right payment date depends on the discount, financing cost, contractual terms and supplier consequences. Automation should optimize that trade-off, rather than treat faster processing and shorter payment delays as interchangeable financial benefits. (7)
Technology history and delivery
| Period | Public position | Delivery and unresolved questions |
|---|---|---|
| 2016 | Cloud manufacturing software and selective in-memory processing | QAD explained why simulation and transactional history have different infrastructure needs. |
| 2018 | Completion of the Channel Islands modernization program | Web interfaces, REST services and an application platform were identified as delivered work. |
| 2019 | AI-powered DSCP | The release specified data cleansing, demand clustering and external-series forecasting; measured incremental economic results were absent. |
| 2020 | Reassessment of blockchain enthusiasm | A QAD executive revisited his 2017 enthusiasm and emphasized adoption and standardization barriers. |
| 2025 | Redzone productivity benchmark | A substantial operational dataset measured a combined deployment and coaching program. |
| 2026–2027 | ChampionAI, Lynx and Manufacturing Intelligence | Concrete workflow products coexist with a broader intelligence layer scheduled for the following year. |
QAD’s historical record contains more discrimination than a simple progression through fashionable labels. In 2016 it argued for in-memory computing in planning simulations while rejecting it as a universal answer for transactional history. That is a task-specific architectural argument. The same article valued customer control over upgrades; today’s cloud program gives QAD control over version deployment and requires customer testing. Cloud modernization has changed the distribution of control and work. (17, 14)
The 2018 Channel Islands account described a multi-year platform and interface program, with REST access and separate integration services. Those are identifiable engineering deliverables. The persistence of substantial customization discovery in 2026 shows that a modern platform does not automatically make an installed customer’s business logic portable or inexpensive to change. (19, 9)
The 2019 AI claim also had identifiable scope: filling missing hierarchical data, grouping similar demand histories and using external series in forecasting. The release’s broad intelligence vocabulary outran the demonstrated tasks, particularly its invocation of causation without an evaluation method. Those narrow ML features do not substantiate today’s claims of autonomous manufacturing judgment. (16)
Blockchain provides a different precedent. In 2020, QAD’s Tom Roberts publicly revisited his earlier enthusiasm and described the difficulty of getting participants to adopt common standards and pay for shared infrastructure. This was technology commentary, with an explicit reassessment of practical obstacles. It does not establish a failed QAD blockchain product. (18)
The September 2026 Manufacturing Intelligence announcement is more ambitious: a layer spanning systems that interprets events and initiates responses, supported by planned NVIDIA integration. Its expected market date is spring 2027. NVIDIA’s name supplies no evidence of QAD-specific improvements to planning methods, and the announcement offers no deployed benchmark of the combined system. Its economic claims remain a delivery commitment to follow. (10)
Software materials and architectural boundaries
| Product or team; evidence date | Documented materials | Consequence |
|---|---|---|
| DynaSys planning; 2019 product documents | In-memory planning model, embedded Qlik analytics and a data hub for ERP exchanges | Planning state, analytics and transaction records require consistent mappings and refreshes. |
| ERP engineering; vacancy accessed November 2026 | Progress OOABL and Java development | The established application runtime continues to require specialized maintenance. |
| ChampionAI platform; vacancy accessed November 2026 | MCP infrastructure, Kubernetes deployment work and LaunchDarkly prompt rollout; Python/FastAPI and PostgreSQL skills specified for the role | Agents add their own execution, configuration, evaluation and deployment lifecycle. |
| Lynx; September 2026 announcement | AWS Transform and Amazon Bedrock | Modernization depends on external model and transformation services as well as QAD’s interpretation of its applications. |
| Company-wide careers page; November 2026 | Java, JavaScript, AngularJS, C#, MySQL, MongoDB and Progress; cloud operations across AWS and GCP | The wider portfolio supports several technology families. Their presence does not identify a single customer deployment. |
These boundaries matter more than a raw component count. The planning engine needs a coherent interpretation of ERP quantities and dates; an agent needs permission to alter those records; an operational application must execute the accepted action. Each additional representation introduces mappings, release coordination and failure diagnosis.
QAD explicitly places Champions on a platform separate from ERP or MES, including deployments against older and third-party systems. This permits incremental adoption. It also means that a shared brand is not evidence of a shared numerical model: the decision’s meaning must survive interfaces between applications with different histories and responsibilities. (7)
In-memory planning can shorten scenario calculations while retaining substantial memory and recomputation costs. Public material does not specify representative network sizes, workload-dependent runtimes or full refresh costs. The economic question is how much infrastructure and specialist work are needed to keep the decision system correct as the business changes.
Configuration maintenance and ownership cost
Customer-specific logic is central to QAD’s proposition and its ownership burden. The current Lynx page describes scanning customizations, explaining dependencies and transforming legacy logic into modern extensions. That work is more relevant to long-term productivity than adding another conversational interface: it attempts to reduce the effort required to understand and change the software. (23)
The launch account gives a concrete example. An unnamed contract manufacturer and packager had more than 1,000 customizations catalogued; nearly two-thirds received a migration path on the first pass. The engagement identified almost $500,000 in projected savings. The result concerns discovery, classification and a scoped modernization plan, rather than demonstrated completion of all changes or realized savings. (9)
Preserving a customization’s behavior is only one test. A rule may faithfully reproduce an economically poor policy. Useful modernization must also make it easier to diagnose that problem, alter the underlying objective or data interpretation, replay representative situations and release a correction safely. The published Lynx results do not yet trace that complete cycle.
The 2026 cloud program restricts newer-system changes to approved extension mechanisms, requires developer certification for customer access to development environments, and allows QAD to block incompatible changes. It separately excludes configuration changes, testing, data cleaning and upgrade projects from the cloud program’s standard scope. (14)
The resulting TCO includes subscription, integration, model upkeep, regression testing and the delay before useful changes reach production. Restricting changes can reduce uncontrolled modification, but the buyer still needs an affordable route for legitimate changes. An upgrade-safe extension can remain expensive to understand, test and replace.
Mayoly Spindler’s historical case illustrates the operating model. Its August 2019 case study describes a program begun in 2007, selection in 2008, sequential demand and production projects, more than 30 users and a monthly collaborative process. The implementation created a centralized planning discipline; it also required continuing participation across functions. (20)
Security must follow the agent’s authority
QAD’s public Security Overview describes network and user segmentation, security testing and shared responsibilities. It also assigns customers responsibilities for configuring virtual hosts, containers and application layers. The document does not map these responsibilities precisely to each current SaaS product. A buyer needs that mapping before it can know which mistakes the service prevents and which it expects the customer to prevent. (15)
The agent-platform vacancy identifies reviews of authorization and scope, behavioral tests and staged prompt rollout. These are relevant engineering activities. The public account still does not explain how runtime permissions limit a compromised agent, how tool calls are constrained independently of prompts, or how damaging actions are reversed. (11)
A supplier message or invoice is untrusted input. When an agent can read that input and update enterprise records, the security boundary must survive malicious content and incorrect model output. An approval step helps only if the reviewer sees the actual proposed effect and cannot be bypassed. Certifications and a general least-privilege policy do not establish these properties.
Productivity evidence and economic accountability
Redzone’s 2025 benchmark is unusually explicit about its population and intervention. It covers 1,500 plants that completed its 90-day deployment program during 2013–2024, using automated OEE records alongside coaching, leadership support and collaboration tools. QAD reports an average 26.4% productivity increase. The observation concerns that combined operating program. It cannot isolate the contribution of software or establish the value of Champions introduced later. (22)
OEE improvement can be valuable when it produces saleable output, reduces overtime or avoids capital expenditure. It can also create surplus production if demand and inventory decisions remain unchanged. A factory productivity measure therefore needs a financial bridge to the supply chain’s return.
The planning case evidence is less complete. QAD’s anonymous frozen-foods case reports roughly $500,000 less inventory and a 17% forecast-accuracy improvement after a nine-month implementation combining planning software and S&OP changes. It omits a full cost denominator and an attribution method. Mayoly Spindler supplies named operational testimony, but its reported improvements likewise do not establish net economic return after continuing planning effort. (20, 21)
QAD’s public cloud terms define service availability and support obligations much more concretely than economic decision quality. The buyer must connect the project to a baseline, a measurement period, full costs and a bounded remediation process. Otherwise, the vendor can deliver working software and the customer can still be left owning an unprofitable planning policy. (14)
Supply chain score
Supply chain depth
| Criterion | Score |
|---|---|
| Economic framing | 6/10 |
| Decision end-state | 7/10 |
| Conceptual sharpness on supply chain | 6/10 |
| Freedom from obsolete doctrinal centerpieces | 4/10 |
| Robustness against KPI theater | 4/10 |
Average: 5.4/10.
Economic framing — 6/10. Procurement and manufacturing calculations include real costs and resource choices. Financial planning extends the view beyond service levels. The missing link is a common monetary objective that evaluates decisions across uncertain demand and supply outcomes. (2, 5)
Decision end-state — 7/10. The planning suite generates purchasing, distribution and production plans with operational constraints. These are actionable outputs. Acceptance of trade-offs, policy configuration and exception handling still require a substantial surrounding process. (1, 4)
Conceptual sharpness on supply chain — 6/10. Supplier calendars, material dependencies and manufacturing stages give the product a detailed operational vocabulary. The public reasoning is weaker where a feasible plan must become an economically justified commitment under interacting uncertainties. (2, 4)
Freedom from obsolete doctrinal centerpieces — 4/10. S&OP coordination, service-level segmentation and buffer-oriented replenishment remain central. These methods can organize work while concealing the marginal cost of individual decisions behind targets and categories. (4, 5)
Robustness against KPI theater — 4/10. The forecast-improvement headline lacks a reproducible measurement definition. Redzone provides a clearer operational evaluation, but neither a better accuracy percentage nor higher OEE directly establishes supply chain profitability. (3, 22)
Decision and optimization substance
| Criterion | Score |
|---|---|
| Probabilistic modeling depth | 4/10 |
| Distinctive optimization or ML substance | 6/10 |
| Real-world constraint handling | 8/10 |
| Decision automation and labor productivity | 6/10 |
| Resilience under real operational complexity | 5/10 |
Average: 5.8/10.
Probabilistic modeling depth — 4/10. Inventory calculations acknowledge demand and supply variability. The documentation does not expose a joint stochastic projection feeding the selection of purchases and production commitments. This leaves the value of correlated adverse outcomes unclear. (3)
Distinctive optimization or ML substance — 6/10. Detailed planning functions and identified learning tasks give the product more substance than an unexplained AI label. The public evidence does not show how the methods outperform simpler alternatives economically or how the newer agents improve the numerical engine. (2, 16)
Real-world constraint handling — 8/10. The product represents meaningful manufacturing and procurement restrictions, including coupled transport, material and calendar effects. Maintaining those representations remains implementation work, and their breadth does not resolve uncertainty by itself. (2, 4)
Decision automation and labor productivity — 6/10. Document handling and supplier coordination can remove repetitive work. Campaign and contract language retain human review, while public labor figures do not account comprehensively for monitoring, exceptions and ongoing configuration. (6, 8, 14)
Resilience under real operational complexity — 5/10. Customizations and Lynx provide mechanisms for changing business logic. Public examples stop short of tracing a failed planning assumption through diagnosis, reformulation of objectives or actions, historical replay and safe release of a lasting correction. Faster scenario selection does not establish that improvement process. (9, 23)
Product and architecture integrity
| Criterion | Score |
|---|---|
| Architectural coherence | 5/10 |
| System-boundary clarity | 6/10 |
| Security seriousness | 3/10 |
| Software parsimony and ownership cost | 4/10 |
| Compatibility with programmatic and agent-assisted operations | 6/10 |
Average: 4.8/10.
Architectural coherence — 5/10. DSCP describes an integrated planning model, while Champions operate across separate enterprise applications. Consistent data exchange is developed, but a unified computational objective across the portfolio is not established. (5, 7)
System-boundary clarity — 6/10. Product descriptions distinguish planning, transactions and agent workflows reasonably well. The cross-system intelligence announcement extends the intended boundary substantially before demonstrating how its decisions will be reconciled with existing application logic. (1, 10)
Security seriousness — 3/10. Security and engineering materials identify controls and testing responsibilities. They leave agent authority, extension containment and compromise recovery insufficiently inspectable. The buyer cannot derive the potential damage from a compromised tool or input from the public architecture. (11, 15)
Software parsimony and ownership cost — 4/10. Several application generations, integration layers and customer extensions must be maintained. Lynx targets a real source of cost, but its current evidence does not establish a lower full lifecycle burden for planning-model improvements. (9, 13, 14)
Compatibility with programmatic and agent-assisted operations — 6/10. Platform services and agent tooling provide programmatic access. Durable productivity still depends on whether agents can help inspect, test and improve executable decision logic without reproducing the manual review workload. (11, 19)
Technical transparency
| Criterion | Score |
|---|---|
| Public technical documentation | 6/10 |
| Inspectability without vendor mediation | 5/10 |
| Portability and lock-in visibility | 4/10 |
| Implementation-method transparency | 6/10 |
| Evidence density behind technical claims | 5/10 |
Average: 5.2/10.
Public technical documentation — 6/10. Historical product documents explain numerous calculations and operational boundaries. Current public material is stronger on application scope and AI workflows than on the planning engine’s formulation, uncertainty model and reproducible performance. (2, 3, 7)
Inspectability without vendor mediation — 5/10. An outsider can reconstruct substantial planning behavior and identify active engineering areas. The decisive numerical objectives and runtime safeguards still require vendor explanation rather than independent inspection. (4, 11, 15)
Portability and lock-in visibility — 4/10. Integration services expose application data, but migration also requires preserving customer-specific rules and dependencies. Lynx’s own proposition illustrates how much business meaning can remain embedded in an established installation. (19, 23)
Implementation-method transparency — 6/10. Customer cases and public cloud terms identify phased deployment, participation and maintenance responsibilities. They leave the recurring effort needed for a particular customer’s economic model difficult to estimate before detailed scoping. (14, 20)
Evidence density behind technical claims — 5/10. Established planning documents and operational cases provide substance. Broad agentic intelligence claims have thinner validation, and the most ambitious cross-system proposition remains scheduled for future delivery. (10, 21, 22)
Vendor seriousness
| Criterion | Score |
|---|---|
| Technical seriousness of public communication | 5/10 |
| Technology roadmap integrity | 6/10 |
| Conceptual sharpness | 5/10 |
| Economic accountability | 3/10 |
| Defensibility in an agentic-software world | 6/10 |
Average: 5.0/10.
Technical seriousness of public communication — 5/10. Specific engineering and workflow descriptions coexist with expansive systems-of-action claims. The promotion often moves from task automation to manufacturing intelligence without demonstrating the intervening decision mechanism. (7, 10, 11)
Technology roadmap integrity — 6/10. The history includes delivered platform work, identifiable ML tasks and explicit reconsideration of blockchain enthusiasm. The current NVIDIA-backed vision remains an unfulfilled future commitment whose contribution must be measured after release. (16, 18, 19, 10)
Conceptual sharpness — 5/10. QAD understands the practical burden of manufacturing and legacy logic. It less clearly separates accurate records, efficient workflows and economically sound decisions; each requires a different success test. (6, 9)
Economic accountability — 3/10. Published results rarely show returns net of full implementation and maintenance costs. AI review obligations and separate change services leave the customer with substantial decision-quality responsibility, without a similarly explicit public economic remediation commitment. (14, 21)
Defensibility in an agentic-software world — 6/10. Manufacturing semantics, established applications and accumulated integrations are valuable assets. Lynx addresses improvement productivity, but MCP orchestration and conversational access offer a weaker advantage unless they enable sustained improvements to the underlying decision logic. (11, 23)
Sources
[1] QAD — Supply Chain Planning
Supply Chain Planning. Undated; accessed November 4, 2026. Current planning scope and product functions.
Production Planning. 2019; accessed November 4, 2026. Historical manufacturing constraints, cost-based methods and heuristics.
Demand Planning. 2019; accessed November 4, 2026. Forecasting, variance inputs, offset testing and accuracy claims.
[4] QAD — Procurement Planning
Procurement Planning. 2019; accessed November 4, 2026. Purchasing policies, order constraints and transport look-ahead.
[5] QAD — Demand & Supply Chain Planning
Demand & Supply Chain Planning. 2019; accessed November 4, 2026. Suite model, financial views and technical components.
[6] QAD — Procurement Champion
Procurement Champion. Undated; accessed November 4, 2026. Current supplier follow-up and exception workflow.
[7] QAD — Champion Assist and Accounts Payable Champion
Champion Assist and Accounts Payable Champion. 2026-09-22; accessed November 4, 2026. September release; invoice automation and separate Champion platform.
ChampionAI campaign. Undated; accessed November 4, 2026. Campaign claims and action-review language.
[9] QAD — Lynx Champion announcement
Lynx Champion announcement. 2026-09-22; accessed November 4, 2026. Modernization mechanism and projected customer savings.
[10] QAD — Manufacturing Intelligence announcement
Manufacturing Intelligence announcement. 2026-09-22; accessed November 4, 2026. Cross-system vision and expected spring 2027 availability.
[11] QAD — Software Engineer, AI Agent Platform
Software Engineer, AI Agent Platform. Undated; accessed November 4, 2026. Employer-authored platform responsibilities and required engineering skills.
[12] QAD — Progress Developer II
Progress Developer II. Undated; accessed November 4, 2026. Employer-authored ERP development responsibilities.
[13] QAD — Careers technology overview
Careers technology overview. Undated; accessed November 4, 2026. Company-wide technology families; product assignments are not supplied.
[14] QAD — Cloud Program Document
Cloud Program Document. 2026-01-13; accessed November 4, 2026. Version dated January 13, 2026; cloud scope, maintenance and AI responsibilities.
Security Overview. Undated; accessed November 4, 2026. Shared responsibilities and public security description.
[16] QAD — DSCP 2019 release announcement
DSCP 2019 release announcement. 2019-12-02; accessed November 4, 2026. Contemporary QAD announcement distributed by Business Wire; republished by Nasdaq.
[17] QAD — QAD Wins the Stratus Award
QAD Wins the Stratus Award. 2016-07-21; accessed November 4, 2026. Historical architectural choices and upgrade-control position.
[18] QAD — Is Blockchain a Dirty Word?
Is Blockchain a Dirty Word?. 2020-07-07; accessed November 4, 2026. Executive reassessment of blockchain adoption barriers.
[19] QAD — Channel Islands Initiative
Channel Islands Initiative. 2018-10-18; accessed November 4, 2026. Delivered platform, interface and integration work.
[20] QAD — Mayoly Spindler case study
Mayoly Spindler case study. 2019-08; accessed November 4, 2026. Case study updated August 2019; named customer history and operating process.
[21] QAD — Frozen foods manufacturer case study
Frozen foods manufacturer case study. Undated; accessed November 4, 2026. Anonymous planning implementation and reported operational results.
[22] QAD — 2025 Redzone productivity benchmark
2025 Redzone productivity benchmark. 2025-07-30; accessed November 4, 2026. Vendor study methodology and measured factory-program outcomes.
Lynx Champion. Undated; accessed November 4, 2026. Current customization analysis and transformation proposition.