Review of B2Wise, Demand Driven Planning Software Vendor

Last updated: November 4, 2026

Go back to Market Research

B2Wise supplies cloud software for materials replenishment, production scheduling, demand forecasting, and sales and operations planning. Its core follows Demand Driven Material Requirements Planning (DDMRP): inventory buffers, consumption signals, and priorities guide purchasing, production, and distribution. The company also provides methodology training and implementation support. CommWise adds AI-assisted tools for building applications, dashboards, and business workflows around the planning suite. 1 2 13

Overall supply chain score: 3.6/10

Dimension	Score
Supply chain depth	3.2/10
Decision and optimization substance	3.8/10
Product and architecture integrity	4.0/10
Technical transparency	3.8/10
Vendor seriousness	3.4/10

B2Wise produces operational recommendations and addresses concrete manufacturing constraints. Customer accounts report reductions in inventory, purchasing work, and planning meetings. Its central limitation is the decision criterion: buffer status and material flow stand in for an explicit economic comparison of competing actions under uncertainty. A clear priority can still be financially wrong.

The vendor’s criticism of forecast perfection is sensible, but buffers do not eliminate the need to anticipate the future. B2Wise itself uses forecasts to adjust buffers, projects requirements over lead times, and offers a third-party forecasting engine. The public method compresses uncertainty into averages, factors, profiles, and planner settings; it does not expose a calibrated joint model of future demand, supply delays, and operational consequences. 3 6 8

CommWise introduces a second concern. B2Wise has correctly warned that custom planning tools become expensive to maintain after implementation. It now promotes business users creating custom logic within hours. The current CommWise site acknowledges permissions, approvals, versions, and bounded applications. It does not explain enough about execution confinement, testing, or maintenance ownership to demonstrate that faster creation has solved the problems B2Wise previously identified. 10 13 14

B2Wise versus Lokad

B2Wise starts from a prescribed operating method. Practitioners select decoupling points, classify items, configure buffers, and use the resulting signals to prioritize orders. Scheduling adds bottleneck and capacity logic. The intended work remains planner-centered: the software structures choices and surfaces exceptions, while people monitor and adjust the operating model. 3 5

Lokad’s inventory optimization starts from the financial consequences of decisions. Purchases, transfers, allocations, and production orders compete according to their expected economic value across uncertain futures. Recurring decisions are automated, with Supply Chain Scientists maintaining and improving their logic. This is the design perspective developed in Introduction to Supply Chain.

The practical difference concerns scarcity. Two items can occupy similar red buffer zones while having very different margins, substitution options, expiry risks, and consequences of a shortage. Color and normalized penetration do not determine which item deserves the next unit of cash, material, or capacity. An economic model has to make that comparison explicit.

B2Wise reduces the operational problem to policies that planners can understand and maintain. This can stabilize daily work, but the simplification leaves the financial assumptions implicit. Lokad’s approach requires those assumptions to become part of the decision program. The relevant comparison is the economic quality and lifetime cost of the resulting decisions, including the labor needed to keep the model aligned with the business.

Decision logic: buffers are policies, not an economic objective

The Demand Driven Institute’s compliance criteria identify what the underlying method requires: decoupled lead-time calculations, buffer families, quantities derived from usage and lead times, dynamic adjustments, a net-flow calculation, and visible priorities. Compliance checks conformity to that method. The institute expressly distinguishes it from evidence that the vendor can implement the method successfully. 4

B2Wise extends those requirements considerably. Its materials page lists twelve additional buffer types beyond the three standard methods, alongside supplier alternatives, shelf-life handling, container building, minimum order values, and network rebalancing. Demand and supply variability factors are explicit inputs. These are meaningful operational capabilities. 3

Their breadth also makes the economics harder to infer. More buffer types create more choices about how an item should behave. A seasonal item with a minimum order value and two suppliers can require several interacting settings before the system produces a recommendation. Automatically tuning those settings only establishes an improvement if the objective captures the relevant revenue, costs, and risks.

The vendor emphasizes cash generation, yet the disclosed control logic mainly regulates flow, availability, and inventory. Faster flow is valuable when it releases capital or fulfills profitable demand. It is not a universal objective: using capacity for low-margin volume can displace a better opportunity; ordering sooner can increase expiry risk; reducing purchases can sacrifice profitable sales. The financial valuation of these alternatives cannot be recovered from a buffer color alone.

Scheduling adds finite capacity views, routing times, calendars, bottleneck control points, and production-line reassignment. B2Wise says orders can move between lines automatically when one is overloaded. That is a concrete automation claim, but balancing load is not the same as choosing the most valuable mix of orders. The public material does not disclose how the scheduler reconciles changeovers, margin, delays, and scarce components in one economic criterion. 5

The SAP integration page also describes projecting buffers to the full lead time and propagating the shape of average daily usage through the bill of materials. These mechanisms address anticipation and dependent demand. They undermine a literal reading of the claim that decisions use actual demand instead of forecasts: the replenishment policy still embodies a view of consumption before new supply arrives. 8

Forecasting and uncertainty: the aggregation problem

B2Wise’s forecasting page argues against endless forecast adjustment and introduces a threshold-based indicator intended to highlight changes that affect the plan. It identifies Forecast Pro as the forecasting engine and describes selecting an aggregation level before disaggregating results to individual locations and parts. The public page does not give the threshold indicator’s formula or demonstrate its relationship to incremental profit. 6

Reducing unnecessary forecast edits can save labor. However, whether a forecast change crosses a planning threshold depends on the existing policy. A poor policy can be insensitive to economically important changes. Stability is therefore a property to evaluate, not evidence that the decisions are good.

The claim that aggregate forecasts are more accurate is particularly weak as a supply-chain argument. If total demand is forecast correctly but assigned to the wrong warehouse, aggregate accuracy can be perfect while stock is unavailable where it is needed. Disaggregation restores a local number; it does not restore information that aggregation removed. Forecast evaluation must return to the item, location, and horizon of the actual purchase or production decision.

B2Wise also recognizes that supplier performance changes buffer needs. Its materials documentation adjusts buffers as demand, variability, and lead times change. This is stronger than ignoring unreliable supply, but a variability factor and an updated lead time do not constitute a probability distribution of future arrivals. They also do not establish dependencies between demand surges, common suppliers, capacity losses, and shared materials. 3

The distinction matters when buffers compete for the same resources. Protecting each item’s service independently can consume more cash or capacity than the business has. The public record does not explain how B2Wise values the competing shortage risks jointly. Buffer simulations and alternate profiles may help a planner explore the problem; they do not by themselves solve it.

Technology roadmap and continuity of claims

Period	Public record	Assessment
2016–2018	The company gives 2016 as its founding year; its later retrospective dates the first B2Wise projects to 2018.	Company formation and initial product delivery are distinct milestones. 2 10
2019	A vendor user-forum post promises developments in forecasting, machine learning, S&OP, and supplier management over the following six months.	The machine-learning narrative predates generative AI; its promised implementation detail was limited. 20
November 2023	A CEO retrospective announces a completed AWS serverless backend rewrite and a new interface planned for January 2024. It also promotes a training simulation as proving demand-driven planning always superior.	The rewrite is a concrete engineering claim. The claimed universal proof from a training game is not a credible validation method. 18
March 2024	A technical product article identifies AWS serverless, Forecast Pro, segmentation, and buffer-setting AI.	The forecasting component is named and attributable; integration is not evidence of proprietary forecasting research. 19
March–April 2025	One article warns about the maintenance burden of configurable planning systems; the next month’s webinar promotes a one-day S&OP build and disposable software.	Development speed does not resolve the maintenance problem identified by the vendor itself. 10 12
January 2026	Sundance Growth’s investment supports expansion and CommWise. B2Wise calls itself a training company before a software company.	The commercial model continues to emphasize practitioner education and operating-method adoption. 11
August–September 2026	CommWise is formally launched, with customer availability from September 1 and a challenge to build working solutions within 48 hours.	Earlier demonstrations preceded the announced general customer availability. Long-term maintenance results cannot yet be inferred from the launch. 13

There is follow-through behind parts of the roadmap: current pages still identify the forecasting engine, describe scheduling functions, and link to an operating CommWise platform. The problem is the size of the conclusions attached to those developments. An integrated forecasting library does not validate generalized AI; serverless hosting does not establish security or unlimited scaling; a training game cannot establish universal economic superiority.

The 2025 maintenance critique is particularly pertinent. It explains that the hard work begins when real users encounter conditions the original design missed, and that skilled support resources can become difficult to retain. The August 2026 launch answers with a division between a stable standard core and customer-specific applications around it. That boundary may reduce changes to the core, but custom logic at the edge can still alter orders, priorities, and business data. Moving the logic does not remove the obligation to test and maintain it. 10 13

B2Wise’s September 2026 AI article explicitly favors assisting human decisions over replacing planners. That choice limits the productivity ambition: improving each planner’s tools leaves recurring reviews in place when the larger opportunity is to automate the decisions themselves. 2 15

Bill of software materials

Layer	Public evidence	Consequence
AWS application hosting	The November 2023 rewrite announcement and March 2024 product article identify an AWS serverless backend.	Managed infrastructure shifts some operational work to AWS, while application behavior and resource consumption remain B2Wise’s responsibility. 18 19
Forecast Pro	The current demand-forecasting page identifies the external engine.	Forecasting depends partly on another vendor’s methods, releases, and integration contract. 6
Application languages and frameworks	B2Wise’s undated OfferZen employer page lists JavaScript, Node.js, Python, Go, C#, TypeScript, jQuery, and .NET Core.	The recruiting inventory spans several runtime families; it does not establish a single current deployment composition. 17
Data and compute components	The same employer page lists MS SQL, Apache Spark, S3, EC2, Lambda, and Azure.	The dated AWS rewrite needs to be distinguished from this broader inventory; the current role of each component is not documented. 17
Delivery and operation	The employer page lists Docker, Terraform, Prometheus, GitHub, and npm, with development, UAT, and production release stages.	Infrastructure, dependencies, monitoring, and release testing contribute to continuing software cost. 17
ERP exchange	The integration page describes files, secure FTP, mapping, validation, order confirmation, and APIs.	Order execution crosses several transformation and approval boundaries. 7
CommWise applications and connector	The product site describes an application library and approved AI connections; the status service identifies an MCP connector.	Generated applications and their assistant-facing interface add a separate maintenance and authorization surface. 14 23

The undated employer page lists earlier framework versions and customer-scale figures. It does not establish which components survived the AWS rewrite or belong to current deployments.

The more important source of technical mass may be customer-specific logic. B2Wise’s own history describes a growing product team and continuous revisions long after the first successful implementations. CommWise makes creating additional applications cheaper, but each application that influences production data still needs an owner, tests, change control, and a retirement path. A low initial build cost can coexist with a high lifetime maintenance cost.

Configuration work, agents, and security boundaries

B2Wise’s February 2024 improvement article places continuing attention on people, processes, data, and system parameters. It calls for repeated review of buffer performance and adjustment of operating settings. The materials page adds reason codes for overrides and histories of buffer and net-flow behavior. Those records can help investigate decisions; they also confirm that configuration maintenance is a recurring part of the operating model. 3 9

The labor has several layers: maintaining item and supplier data, selecting profiles and decoupling points, resolving exceptions, approving orders, reviewing policies, and now maintaining generated applications. The homepage’s 70% planner-efficiency claim lacks a published denominator and method. It cannot establish the change in total labor across customer staff, consultants, support, and application maintenance. 1

The standard integration workflow still says users review and confirm order quantities before export to ERP. The connector automates transport after that confirmation. This is a meaningful boundary: automated transfer of a human-approved order is different from automated economic selection of the order. 7

CommWise’s MCP connector lets an AI assistant interact with web applications. That interface is inexpensive to add relative to engineering a robust decision policy. Its existence therefore carries little weight as evidence of supply-chain intelligence. Building more dashboards or reproducing a planner’s sequence of app interactions can preserve the same labor-intensive operating method beneath a new interface. 23

The present CommWise page does identify governance features: approved AI connections, scoped tools, permissions, approvals, and reusable versions. These are relevant product commitments. The public description does not specify the execution model for generated logic, isolation between tenants, restrictions on outbound connections and secrets, or the checks that prevent an application from corrupting production data. 14

This omission is especially consequential when the builder targets business practitioners. The risk does not depend on whether they type Python or describe the same computation to an LLM. Once generated logic can access live business data, mistakes and compromised dependencies need containment by construction. Human approval is insufficient if the approver cannot inspect the resulting program’s behavior.

The Trust Center describes policies, reports, and controls including audit logging, least privilege, and multi-factor authentication. Its public FAQ gives examples such as HTTPS, managed firewalls, segregated guest networks, and encrypted Windows laptops. Those measures address particular exposure paths; they do not explain the containment of generated planning applications. Certification and audit badges cannot supply the missing architectural answer. 16

Customer results and responsibility for outcomes

The September 2026 Blanchon case reports a 15% reduction in raw-material inventory value, a 60% reduction in purchase-order volume, and service maintained above 99%. It describes full deployment in France within six months, using a B2Wise connector to Sage X3. These figures support both inventory improvement and fewer purchasing transactions. The service figure is an operational service metric, not a forecast-accuracy result. 21

Fewer purchase orders can reduce repetitive work, but the case does not measure total planner hours or the support and coaching required to maintain the model. It explicitly identifies continuing support and iteration as part of the implementation. The six-month deployment also gives a more concrete scope than the investment release’s broad promise of results within ninety days: initial benefits and completion of a deployment are different milestones. 11 21

The February 2025 Goizper webinar page reports sales rising 30%, planning-meeting time falling 80%, and service exceeding 90%. A named manufacturing leader participates in the account. Meeting time is a specific productivity measure, but it covers only one part of planning work; sales growth is not a measure of the software’s incremental profit. The page does not disclose a controlled attribution or all-in cost calculation. 22

The responsibility model deserves equal attention. Training and configuration genuinely affect results, but they also provide an easy explanation for any disappointing outcome. B2Wise’s materials explicitly monitor whether planners follow recommendations, and its 2026 launch assigns the business user ownership of decisions. The public cases do not show the reciprocal analysis: how the vendor identifies recommendations that were followed correctly yet lost money, and how it corrects the underlying policy. 3 13

A credible economic assessment must include both kinds of failure. Otherwise successes validate the methodology, while failures can always be attributed to data, training, adoption, or customer judgment. The current evidence reports benefits without closing that accountability loop.

Supply chain score

The 25 criteria carry equal weight. Each dimension is the arithmetic average of its five sub-scores.

Supply chain depth

Criterion	Score
Economic framing	3/10
Decision end-state	6/10
Conceptual sharpness on supply chain	3/10
Freedom from obsolete doctrinal centerpieces	1/10
Robustness against KPI theater	3/10

Average: 3.2/10.

Economic framing — 3/10. Cash generation is prominent in the positioning, but the disclosed decision mechanism centers on flow, buffer status, and policy settings. No public formulation shows how competing purchases or production orders are ranked by their expected economic contribution. 1 3

Decision end-state — 6/10. The software produces purchasing, manufacturing, and distribution recommendations and supports ERP order transfer. These are operational outputs. Human confirmation and policy maintenance remain substantial parts of the process. 7

Conceptual sharpness on supply chain — 3/10. B2Wise recognizes forecast instability and physical constraints. Its opposition between actual demand and forecasting obscures the anticipatory assumptions already built into buffers and lead-time projections. 3 8

Freedom from obsolete doctrinal centerpieces — 1/10. DDMRP profiles, segmentation, visual priorities, tactical parameter review, and certification form the product’s center. The method substitutes a different set of planning proxies for an explicit economic decision model. 4 9

Robustness against KPI theater — 3/10. Customer cases provide concrete inventory and workload measures. Aggregate forecast-accuracy positioning, broad efficiency claims, and emphasis on following signals can still reward favorable indicators without establishing better net economics. 1 6 21

Decision and optimization substance

Criterion	Score
Probabilistic modeling depth	2/10
Distinctive optimization or ML substance	3/10
Real-world constraint handling	6/10
Decision automation and labor productivity	4/10
Resilience under real operational complexity	4/10

Average: 3.8/10.

Probabilistic modeling depth — 2/10. Variability factors, changing lead times, forecasts, and simulations are documented. They do not expose calibrated distributions of future demand and supply or their joint use in selecting actions. 3 6

Distinctive optimization or ML substance — 3/10. B2Wise identifies its forecasting supplier and describes domain-specific buffer extensions. AI claims about selecting profiles, cleaning data, and adjusting settings lack public algorithms and outcome validation sufficient to establish a distinctive optimizer. 3 6 19

Real-world constraint handling — 6/10. Routing times, finite capacity, line reassignment, minimum orders, alternative suppliers, shelf life, and container building are concrete. The unresolved issue is how their interactions are valued and selected economically. 3 5

Decision automation and labor productivity — 4/10. Blanchon reports fewer purchase orders and Goizper shorter planning meetings. These are meaningful gains. The vendor’s operating model still retains human review and continuing settings work, with no comprehensive labor accounting. 7 21 22

Resilience under real operational complexity — 4/10. Multiple buffer types and manufacturing controls accommodate operational differences. The increasing number of policies and custom applications also creates coordination work; public evidence does not quantify how that burden scales with exceptions and changing constraints. 3 5 9 14 19

Product and architecture integrity

Criterion	Score
Architectural coherence	5/10
System-boundary clarity	5/10
Security seriousness	3/10
Software parsimony and ownership cost	3/10
Compatibility with programmatic and agent-assisted operations	4/10

Average: 4.0/10.

Architectural coherence — 5/10. The planning suite follows a consistent operating method, with an identifiable forecasting engine and ERP exchange model. CommWise’s separate application layer creates a new boundary for data, permissions, and business logic whose integration is only partly documented. 6 7 14

System-boundary clarity — 5/10. The standard file, validation, confirmation, and ERP-export workflow is described. CommWise distinguishes the shared core from customer extensions, but the runtime and production-data permissions of those extensions remain unclear. 7 13

Security seriousness — 3/10. The public record mentions practical access and transport controls alongside governance features. It does not demonstrate the isolation and restricted execution needed when non-specialists create applications around live planning data. Compliance labels do not resolve that gap. 14 16

Software parsimony and ownership cost — 3/10. The recruiting inventory spans several technologies, while the operating method adds ongoing configuration, training, and support. Rapidly generated applications risk adding maintenance obligations faster than the business can validate their value. 9 10 17

Compatibility with programmatic and agent-assisted operations — 4/10. ERP APIs and the CommWise MCP connector expose automation surfaces. The public implementation remains organized around planner work and app building rather than a documented program that autonomously selects and evaluates economic decisions. 7 23

Technical transparency

Criterion	Score
Public technical documentation	4/10
Inspectability without vendor mediation	3/10
Portability and lock-in visibility	3/10
Implementation-method transparency	5/10
Evidence density behind technical claims	4/10

Average: 3.8/10.

Public technical documentation — 4/10. Product pages disclose planning inputs, operational objects, and integration steps. Public material is much thinner on API contracts, optimizer objectives, forecast validation, and generated-application execution. 3 5 7

Inspectability without vendor mediation — 3/10. The DDI criteria explain the baseline method, and B2Wise describes extensions. An outsider cannot reproduce the proprietary buffer selection, AI behavior, or threshold indicator from the published material. 4 6

Portability and lock-in visibility — 3/10. Input files and order exports are legible. Public evidence does not establish portable exports of operating policies, custom app logic, permissions, and histories sufficient to run the same process independently. 7 14

Implementation-method transparency — 5/10. B2Wise openly describes education, transformation, data preparation, support, and repeated parameter improvement. Customer cases identify connectors and implementation scope. Continuing effort and costs are much less quantified. 9 11 21

Evidence density behind technical claims — 4/10. Named components and customer outcomes support the existence of functioning planning software. They provide weaker support for broad AI, scalability, and generalized productivity claims, particularly for the newly launched application-building layer. 13 17 19

Vendor seriousness

Criterion	Score
Technical seriousness of public communication	3/10
Technology roadmap integrity	4/10
Conceptual sharpness	3/10
Economic accountability	4/10
Defensibility in an agentic-software world	3/10

Average: 3.4/10.

Technical seriousness of public communication — 3/10. Operational detail is mixed with universal claims about the superiority of the methodology and the benefits of flow. A training simulation cannot establish those claims across real supply chains. 3 18

Technology roadmap integrity — 4/10. Forecasting integration, AWS modernization, and CommWise show identifiable developments. The transition from warning about custom-software maintenance to promoting rapid app creation leaves an important engineering question unanswered. 10 13 18

Conceptual sharpness — 3/10. The vendor is clear about planner-centered methods, but the financial meaning of its control signals remains implicit. Rejecting forecast perfection does not remove the need to model uncertainty or compare decisions economically. 2 3 6 8

Economic accountability — 4/10. Blanchon and Goizper provide measurable operational results. Public evidence does not establish net returns after recurring labor or a balanced process for identifying costly software recommendations when users followed them correctly. 21 22

Defensibility in an agentic-software world — 3/10. Manufacturing details and implementation experience have substance. Training, dashboards, and conversational app creation offer weaker protection as software creation becomes cheaper. The harder advantage would be validated economic decisions with much less total human work. 5 9 14 21 23