Overview
The NexLog OS Control Tower is the platform's decision layer. It consumes the event ledger (businessEvents) from WMS, TMS, OMS, YMS, Customs, ESG and Finance, applies the exception engine to deviations, triggers versioned playbooks and exposes a digital twin for what-if simulation.
It is a control tower that acts, not a panel that aggregates KPIs: it observes, decides, acts and learns on the same core of events and canonical data, without ETL.
Capabilities
The eight capabilities that make up the decision layer.
Executive Cockpit
Consolidated KPIs from WMS, TMS, OMS, YMS, ESG and Finance, with drill-down by tenant, channel and region.
Exception engine
Exception engine that detects deviations, triggers playbooks and escalates to a person when needed.
Playbooks
Versioned rules that automate decisions, with history, rollback and a full audit trail through the event ledger.
Digital twin
Digital replica of the logistics network, assets and orders that reflects the real state and enables what-if simulation.
Network design
Network optimization: DCs, routes, cross-dock, modal shift and the cost-to-service balance.
Capacity planning
Capacity planning by time window, dock, DC, route and workforce.
Peak readiness
Stress test and contingency plan for Black Friday, Christmas and large-scale promotions.
Operational GenAI
An assistant that drafts playbooks, explains exceptions in natural language, summarizes dashboards for leadership and suggests parameter adjustments under supervision.
Decision cycle
A control tower needs to observe, decide, act and learn. In NexLog OS, each step has its own component.
| Step | Component | What happens |
|---|---|---|
| Observe | Event ledger (businessEvents) | Receives canonical events from every module, correlated by correlationId and causationId. |
| Decide | Exception engine | Detects deviations and selects the applicable playbook. |
| Act | Playbooks | Automate the decision or escalate to the manager, who approves or rejects. |
| Learn | Event ledger and GenAI | Each decision is recorded in the ledger as an auditable event; GenAI learns from the transactional state. |
Data sources
Control Tower reads each module's canonical collections directly from the operational MongoDB, with no separate data warehouse for operations.
| Module | Collections | Used for |
|---|---|---|
| WMS | movements, inventoryCounts | Inventory accuracy and productivity. |
| TMS | shipments | OTD and transit tracking. |
| OMS | orders | Order completeness. |
| YMS | dockSchedules | Dock capacity and occupancy. |
| Customs | tradeDeclarations | Declaration status and Denied Party Check blocks. |
| ESG | carbonEntries | CO2e by region in the service footprint. |
| Finance | costAllocations, billing | Cost-to-serve per order. |
Consolidated KPIs
Unified OTIF
Consolidated metric of OTD (TMS), order completeness (OMS) and inventory accuracy (WMS).
Cost-to-serve
Total cost per order, allocated in costAllocations and billing.
Service footprint
Geographic coverage with service time, cost-to-serve and CO2e by region.
Roles and governance
Access follows the platform's ABAC authorization and multi-tenant roles (enum in UPPER_SNAKE_CASE).
| Role | Scope | In Control Tower |
|---|---|---|
TENANT_OWNER | Tenant | Executive who owns the account, with full control over the tenant. |
TENANT_ADMIN | Tenant | Manages users, plans and integrations. |
MANAGER | Module | Operations manager: approves exceptions and playbooks. |
OPERATOR | Module | Executes the WMS, TMS and YMS tasks generated by decisions. |
VIEWER | Read | Read-only, for analysts and auditors. |
Playbook and GenAI governance
Playbooks are versioned, with history and rollback. Every GenAI suggestion records the prompt, model and audit trail, and the user approves or rejects it before any change.
Control Tower as the decision layer of a supply chain OS
Dashboards are not control towers
Many companies call a control tower what is, in practice, a panel that aggregates KPIs. A real control tower needs to observe, decide, act and learn.
In NexLog OS, it consumes the event ledger (businessEvents) from every module, applies the exception engine to deviations, escalates versioned playbooks and exposes the digital twin for what-if simulation. Unified OTIF, cost-to-serve and service footprint sit on the same panel, without ETL.
Digital twin and what-if on the real state
With the unified MDM and the central event ledger, simulations run on the real state of the business. Network design tests new network scenarios, capacity planning sizes docks, routes and DCs before the peak, and peak readiness validates the operation against expected peaks.
As a result, capacity planning becomes a continuous exercise rather than an annual one.
GenAI applied to the supply chain
NexLog OS GenAI acts as a copilot: it suggests playbooks, explains exceptions in natural language, summarizes dashboards for leadership and adjusts parameters under supervision.
Because it lives on the same core of events and canonical data, it needs no parallel data lake: it learns from the transactional state, with native governance and traceability.