Most enterprises do not have one factory, one ERP instance, or one AI use case. They have dozens of plants across multiple countries, a mix of SAP ECC and S/4HANA systems, legacy SCADA platforms, IoT sensor networks, and a growing list of AI workloads — from predictive maintenance to quality inspection to supply chain optimisation. The challenge is not building one AI model. It is deploying AI across a complex, distributed enterprise without creating an unmanageable web of point-to-point integrations. SAP Business Technology Platform (BTP) solves this through a hub-spoke architecture: a central AI hub that orchestrates models, data, and governance, connected to edge spokes at each plant or facility that execute AI locally where decisions happen. Oxmaint fits natively into this architecture as the maintenance spoke — capturing field execution data at every site and syncing bi-directionally to the SAP hub. Start your free trial and plug Oxmaint into your SAP BTP architecture from day one. Or schedule a demo to see how Oxmaint operates as the maintenance execution spoke in your enterprise AI strategy.
Enterprise AI ArchitectureSAP BTP Connectivity for AI: Hub-Spoke Architecture Strategy (2026 Guide)Strategic Guide · 8 min read
350+
AI features now available on SAP BTP — including Joule Agents, Knowledge Graph, and Generative AI Hub
2,400+
Joule skills deployed across SAP BTP — powering agentic workflows in procurement, maintenance, and finance
40%
Enterprise applications will feature AI agents by end of 2026 — up from less than 5% in 2025 (Gartner)
4x
Impact achieved by enterprises with unified data and governance foundations vs fragmented AI deployments
Oxmaint: The Maintenance Spoke in Your SAP AI Architecture
Oxmaint captures field execution data at every facility — work orders, condition scores, parts consumption — and syncs it bi-directionally to your SAP BTP hub. Every plant gets mobile-first maintenance AI. Your enterprise gets unified, governed data.
What Is Hub-Spoke AI Architecture?
Hub-spoke architecture is an enterprise design pattern where a central platform (the hub) manages AI models, data governance, security, and orchestration — while distributed sites (the spokes) execute AI locally for real-time, on-site decisions. Instead of every factory building its own AI stack, or every AI workload routing through a single cloud bottleneck, the hub provides shared services and the spokes provide localised execution. SAP BTP serves as the natural hub for SAP enterprises, and Oxmaint operates as the maintenance execution spoke at each facility.
SAP BTP Hub
AI Core · Knowledge Graph · Gen AI Hub · Governance
Plant A
Edge AI · Oxmaint · IoT
Plant B
Edge AI · Oxmaint · SCADA
Power Plant
Edge AI · Oxmaint · DCS
Water Utility
Edge AI · Oxmaint · PLC
HQ / FM
Analytics · Dashboards
Remote Sites
Offline · Mobile · Sync
Why Hub-Spoke Beats Point-to-Point for Industrial AI
✗ Every system connected to every other system
✗ N² integration complexity — 10 systems = 90 connections
✗ No central governance — each site manages its own AI
✗ Model updates deployed manually to each location
✗ Data silos — no unified view across plants
✓ Central hub orchestrates; spokes execute locally
✓ N+1 complexity — add a spoke, not a web
✓ Unified governance, security, and compliance
✓ Models trained centrally, deployed to all spokes at once
✓ Enterprise-wide data layer — portfolio-level intelligence
SAP BTP: The AI Hub Components
SAP BTP is not a single product — it is a layered platform that provides the shared AI infrastructure your spokes depend on. Here are the four layers that make it the natural hub for enterprise AI architecture.
01
AI Core & Generative AI Hub
The control plane for all AI workloads. Deploy, manage, and orchestrate foundation models (GPT, Claude, Gemini, Mistral) through a single governed interface. Train models centrally, push optimised versions to spokes. Full lifecycle management with metering, versioning, and compliance.
02
SAP Knowledge Graph & HANA Cloud
The semantic intelligence layer. 452K ABAP tables and 7.3M fields connected into a relationship network AI can reason over. HANA Cloud provides vector search for RAG, graph traversal for relationship reasoning, and in-memory analytics — all in one multi-model engine.
03
Integration Suite & Connectivity
The data plumbing. Pre-built connectors to S/4HANA, ECC, Ariba, SuccessFactors, and 170+ third-party systems. OData, REST, RFC/BAPI, and event-driven integration. Every spoke — including Oxmaint — connects through governed, auditable data flows without custom middleware.
04
Joule Agents & Orchestration
The action layer. Joule Agents execute multi-step workflows autonomously — procurement, maintenance planning, production scheduling — while respecting business rules and human approval gates. Multi-agent collaboration across departments, all grounded in the Knowledge Graph.
The Spoke: Where AI Meets the Factory Floor
The hub provides governance, models, and orchestration. But AI only creates value when it executes where work happens — on the plant floor, at the pump station, in the technician's hands. Each spoke runs a local execution stack: edge AI for real-time detection, a CMMS like Oxmaint for maintenance workflows, IoT gateways for sensor data, and local storage for offline operation. The spoke operates independently when disconnected and syncs to the hub when connected.
Oxmaint CMMS
Mobile work ordersParts trackingCondition monitoringOffline execution
Edge AI Processing
Anomaly detectionVision inspectionVibration analysisSmall language models
IoT & Sensor Layer
OPC-UA / MQTTSCADA integrationPLC connectivitySensor gateways
Deploy Oxmaint as the Maintenance Spoke at Every Facility
Oxmaint plugs into your SAP BTP hub-spoke architecture natively. Every site gets mobile-first maintenance execution with offline capability. Every work order, parts transaction, and condition score syncs to your enterprise AI layer automatically.
Deploying Hub-Spoke AI: A 4-Phase Approach
Phase 1
Establish the Hub
Provision SAP BTP with AI Core, Integration Suite, and HANA Cloud. Define governance policies, data ownership rules, and model lifecycle standards. This is the foundation — invest once, share across all spokes.
Phase 2
Deploy Pilot Spoke
Select one facility. Deploy Oxmaint for maintenance execution, connect IoT sensors, and activate edge AI for condition monitoring. Validate bi-directional data sync to SAP PM, MM, and FI/CO through the hub.
Phase 3
Train & Optimise Centrally
Use aggregated data from the pilot spoke to train AI models in the hub — anomaly detection, RUL prediction, parts demand forecasting. Optimise models using real operational data, then push refined versions back to the spoke.
Phase 4
Scale Across Sites
Replicate the spoke stack to additional facilities. Each new spoke inherits the hub's governance, models, and integration — no per-site AI buildout. Oxmaint deploys at each site in 4–8 weeks with full SAP sync from day one.
What Flows Between Hub and Spoke
Trained AI models & updates
Equipment master data & hierarchies
Maintenance plans & PM schedules
Material master data & BOMs
Governance rules & compliance policies
Work order completions & confirmations
Sensor readings & condition scores
Parts consumption & goods issues
Anomaly detections & failure codes
Training data for model improvement
Why This Architecture Wins
Scale Without Complexity
Adding a new plant means adding a spoke — not rebuilding integrations. Oxmaint deploys at each new site in weeks, inheriting the hub's governance, models, and SAP connectivity automatically.
Governed by Default
Security policies, data masking, content filtering, and compliance rules are defined once at the hub and enforced across every spoke. No per-site governance scramble.
Fast Where It Matters
Time-critical AI (anomaly alerts, defect detection) runs at the spoke with sub-5ms latency. Portfolio analytics and model training run at the hub with massive compute. Each workload runs where it performs best.
Works Offline
Spokes operate fully offline — Oxmaint captures work orders, parts, and readings without connectivity. When the link to the hub returns, everything syncs. No data loss, no operational disruption.
Frequently Asked Questions
What is SAP BTP and why does it matter for AI?
SAP Business Technology Platform (BTP) is SAP's cloud platform that provides AI Core, the Generative AI Hub, Knowledge Graph, HANA Cloud, and Integration Suite as shared services. For industrial AI, it serves as the central hub that orchestrates AI models, governs data flows, and connects distributed facilities — eliminating the need for per-site AI infrastructure.
What is hub-spoke architecture for enterprise AI?
Hub-spoke architecture centralises AI governance, model management, and data orchestration in a single platform (the hub), while distributing execution to local sites (the spokes). Each spoke runs AI locally for real-time decisions and syncs data back to the hub for portfolio-level intelligence. This avoids the N² complexity of point-to-point integrations.
How does Oxmaint fit into SAP BTP architecture?
Oxmaint operates as the maintenance execution spoke at each facility. It captures work orders, parts consumption, condition scores, and technician data on mobile devices — even offline. Through a certified SAP connector, all data syncs bi-directionally to SAP PM, MM, and FI/CO through the BTP hub. Each site gets full maintenance AI capability; the enterprise gets unified data.
Start your free trial and deploy Oxmaint as your first maintenance spoke this month.How quickly can we add a new spoke (plant/facility)?
Does this work with SAP ECC and S/4HANA?
Yes. Oxmaint's certified connector supports both SAP ECC and S/4HANA environments. SAP BTP's Integration Suite also provides pre-built connectors for both versions — making the hub-spoke architecture viable whether you have fully migrated to S/4HANA or are still running ECC at some sites.
Schedule a demo to confirm compatibility across your SAP landscape.
One Hub. Every Spoke. Unified Maintenance Intelligence.
SAP BTP provides the enterprise AI hub. Oxmaint provides the maintenance spoke at every facility. Together, you get scalable, governed, edge-ready maintenance AI that syncs to SAP — across your entire portfolio. Deploy in weeks, not quarters.
