GenXChains replaces fragmented spreadsheet planning with a single workflow-driven platform that aligns demand, supply, inventory and financial outcomes. Forecasts are generated, plans are submitted and approved, and every change is attributed — so the number your executives see has a traceable owner.
Role-based access for eight planning roles · REST API with OpenAPI & Postman collection
Most planning organizations do not lack data. They lack one version of it, and a process that forces agreement before the month closes.
Demand sits with sales, supply with operations, inventory with the warehouse. Reconciling them is the job, and it consumes the cycle.
Three teams arrive at the review with three numbers and no shared definition of which is current, or who changed what.
Weeks of preparation for a meeting whose decisions are recorded in slides, then lost before the next cycle begins.
"What if demand runs 15% high?" is answered by intuition because building the alternative plan takes longer than the meeting.
Stockouts, capacity constraints and KPI breaches surface after they have already cost something.
GenXChains puts all four planning domains on a shared data model with role-based approvals and a complete audit trail.
The S&OP cycle is modelled explicitly in the platform. Each step has an owner, a gate, and a recorded decision — so the process advances only when the prior step is genuinely complete.
Admins maintain products and categories; master data is agreed before planning starts.
AdminStatistical forecasts are generated, adjusted with judgement, then submitted for approval.
Demand plannerCapacity and constraints are modelled and gap analysis is run against the approved demand plan.
Supply plannerInventory health is reviewed and finance runs scenarios to quantify the tradeoffs.
Inventory · FinanceLeadership reviews KPIs, approves the final plan or scenario, and the decision is recorded.
ExecutiveEach module is a first-class part of the platform rather than a bolt-on report. They share the same products, periods and plan versions, so a change in one is visible in the rest.
Executive overview with summary KPIs, open alerts and current S&OP cycle status.
Product and category master data — the shared vocabulary every other module plans against.
Create, adjust, submit and approve demand plans with forecast, adjusted, consensus and actual quantities.
Capacity and constraint modelling, lead times and cost per unit, with gap analysis against demand.
Inventory health, safety stock and reorder points, policy exceptions and recommendation runs.
Seven models, automatic selection by history length, accuracy metrics and anomaly detection.
What-if simulations with side-by-side comparison — including best-case and worst-case variants.
The five-step workflow with owners, gates and step advancement under governance.
KPI dashboards with trends, targets and alerting when a metric moves off target.
Every status transition and material change recorded with actor, entity and before/after values.
Agentic recommendations against the supply plan, with approval, revision and publish workflow.
Canonical event ingest with idempotency, replay, out-of-order detection and dead-letter handling.
Forecasting is a strategy library, not a single black box. GenXChains selects a model based on how much history a product actually has, and reports the accuracy so a planner can disagree with evidence.
| Model | Best suited to | Typical use |
|---|---|---|
| Moving average | Short, noisy history | New products and low-volume SKUs |
| Exponential smoothing | Level with mild trend | Stable repeat demand |
| EWMA | Recency-weighted series | Demand that has recently shifted level |
| Seasonal naive | Strong, stable seasonality | Seasonal ranges with several cycles of history |
| ARIMA | Autocorrelated series | Medium history with structure worth modelling |
| Prophet | Trend plus multiple seasonalities | Long history with holidays and changepoints |
| LSTM | Non-linear patterns | Long history where simpler models underfit |
The factory picks a strategy from the length and shape of the series, so planners are not asked to be statisticians to get a defensible baseline.
Outliers are flagged by index rather than silently smoothed, so a spike is a question for a planner instead of a distortion in next month's plan.
Forecast runs are persisted with their diagnostics and audits, making model-versus-actual reviewable after the fact.
Scheduling recommendations are produced by a team of specialised agents, and every recommendation carries an explanation, an impact estimate and a revision history. Nothing reaches the floor without a human approving it.
Proposes schedule changes against the current supply plan and production events.
Enforces capacity, lead-time and policy limits so proposals stay feasible.
Weighs objectives configured per site rather than applying one fixed cost function.
Runs the what-if before the change is published, and persists the run for later inspection.
Detects machine-down and out-of-order events and raises them for triage with severity.
States why a recommendation was made, in language a scheduler can challenge.
Safety stock and reorder points are treated as decisions with owners and evidence, not as constants someone set two years ago.
Inventory is where demand error and supply constraint actually collide. When policy is stale, the S&OP plan is agreed against buffers that no longer reflect the business — and the gap shows up as expedites and write-offs rather than as a planning decision.
GenXChains puts inventory review inside the reconciliation step, so buffers are re-agreed in the same cycle that sets demand and supply.
GenXChains is a planning layer, not a replacement for your ERP or WMS. It ingests master data, inventory positions and demand actuals, and publishes agreed plans back.
Documented integration blueprint covering master data, inventory and demand actuals.
Mapped against the same canonical data-source matrix used across ERP integrations.
Ingest and publish paths follow one integration contract rather than per-system bespoke code.
OpenAPI schema plus a maintained Postman collection and environment for direct integration.
Events carry idempotency keys, so a retried delivery does not double-count. Out-of-order events are detected against a reorder grace window and marked for triage instead of quietly rewriting state. Failed events retain retry budget and dead-letter context, and any event can be replayed.
Access is scoped to the job. Approvals are gates rather than formalities, and the audit log is written as part of the transaction — not reconstructed afterwards.
Reviews KPIs, approves plans and scenarios, makes the final call.
Runs the cycle, assigns owners and advances the steps.
Generates forecasts, adjusts demand plans, submits for approval.
Builds supply plans, manages capacity and runs gap analysis.
Owns inventory health, safety stock and reorder policy.
Runs scenarios and quantifies financial impact.
Manages users, products and categories.
Read-only visibility for stakeholders who consume rather than plan.
Business rules are backed by check constraints and unique business keys, so invalid state is rejected by the store, not only by the UI.
Every schema change ships as a reviewed migration, making environments reproducible and upgrades auditable.
The application refuses to start in production with a default signing key or an unmanaged schema.
Priced per planning seat. Viewers are unlimited on every plan — the people who consume the plan should never be a reason not to adopt it.
One product family, one cycle, to prove the process before committing.
The full governed cycle for a single planning organization.
Multi-site planning with ERP integration and agentic scheduling.
No. GenXChains is a planning layer that sits above your systems of record. It ingests master data, inventory positions and demand actuals, and publishes the agreed plan back. Integration blueprints exist for Epicor Kinetic, SAP and Oracle, and everything is reachable over a documented REST API.
Less than you would need for a single-model tool. Model selection is driven by series length, so short-history products get a moving average or exponential smoothing while long-history products get ARIMA, Prophet or LSTM. You get a defensible baseline on day one and better models as history accumulates.
No. Agentic recommendations enter a pending-approval state and require a human to approve, reject or publish them. Every recommendation carries an explanation and an impact estimate, revisions are chained, and published schedules are snapshotted for replay.
Plans carry an explicit version and a unique business key across product, period, region and channel. The database rejects conflicting versions rather than accepting a silent overwrite, and the audit log records who changed what.
Status transitions are constrained — a plan moves through draft, submitted and approved, and invalid transitions are rejected. Permissions are scoped by role, so submitting and approving are genuinely different rights held by different people.
Yes. GenXChains is a containerized FastAPI backend with a React front end and a PostgreSQL database, with schema changes shipped as reviewed migrations. It runs in your cloud or on your own cluster.
Bring a single product family and one month of history. We will stand up the cycle, generate the baseline forecast, and show you the gap analysis against your current plan.