AI Is Moving Supply Chains from Reactive Execution to Autonomous Orchestration

At SAPinsider Las Vegas 2026, Hernan De la Torre, Principal at KPMG US, presented the session “From Reactive to Autonomous: AI-Enabled Neural Supply Chain Orchestration in SAP,” during which he outlined how AI is reshaping supply chain operations, not by replacing existing systems but by introducing a new orchestration layer that coordinates decisions and execution across them.

This shift reflects a broader transition toward a business-led, AI-first, SAP-centric model of transformation, where value is not driven by isolated technology deployments, but by how decisions are orchestrated across the enterprise. As De la Torre noted, “Supply chains were built for efficiency, not for volatility. That must change.”

For SAP customers, the implication is clear. The next generation of SAP Transformations are no longer centered on implementing systems or optimizing individual processes. It is about enabling coordinated, real-time decision-making across the network.

From Linear Chains to Neural Orchestration

De la Torre framed the evolution of supply chain in four stages, each expanding the role of data, connectivity, and intelligence:

1. Digital Supply Chain (DSC): Linear, process-driven systems enhanced by visibility and reporting
2. Digital Supply Network (DSN): Interconnected environments using shared data and digital twins
3. Agentic Supply Networks (ASN): AI agents executing real-time actions across functions
4. Neural Supply Chain Orchestration (NSCO™): Human and AI co-orchestrating decisions and execution across the network.

Most organizations today are still operating between DSN and early ASN. The transition to NSCO introduces a fundamentally different operating model, where systems are not just connected, but continuously adapting, learning, and coordinating outcomes.

Orchestration Defines the Shift

Many organizations have already invested in automation through APIs, RPA, and batch processing. These approaches improve efficiency, but they remain bound within predefined workflows.

The shift described by De la Torre moves toward coordinated, multi-agent orchestration, where AI continuously senses changes in supply and demand, evaluates trade-offs, and initiates actions across systems.

Instead of processes being automated in isolation, they are actively managed across domains.

This is where supply chain becomes distinct. In finance, AI accelerates analysis. In supply chain, AI directly influences physical and operational outcomes, from inventory positioning to logistics execution.

SAP Remains the Execution Core

In this model, SAP, more than ever, remains the backbone of enterprise execution. Core systems such as SAP S/4HANA, IBP, and Ariba continue to serve as the system of record and the operational foundation that ensures integrity, control, and scale.

What changes is how organizations maneuver through that backbone. Instead of navigating transactions and dashboards, users increasingly engage through conversational and AI-driven interfaces, including Joule. These interfaces translate intent into coordinated actions across systems.

This introduces a new interaction model:

• SAP remains the execution layer
• AI becomes the coordination layer
• The user engages at the level of decisions, not transactions.

Architecture, not Algorithms, is the Constraint

While much of the market conversation focuses on models and agents, De la Torre emphasized that the real constraint is architectural readiness.

The NSCO™ model is not defined by a single capability, but by how multiple layers operate in proximity, each with a clear role in sensing, deciding, and executing across the network:

• Execution Core (SAP Backbone): SAP S/4HANA and core supply chain platforms remain the system of record and transaction engine
• Data & Signal Layer (Context): Internal enterprise data extended with external signals from suppliers, markets, IoT, and networks, increasingly enabled through SAP Business Data Cloud (BDC) and SAP’s recent investments in external data integration
• Agentic Layer: Domain-specific AI agents operating across planning, sourcing, manufacturing, and logistics
• Orchestration Layer (NSCO™ Core): Coordinating agents, decisions, and execution in real time across systems, ensuring interoperability and continuity
• Governance Layer: Enforcing policy, compliance, security, and trust across all actions
• Cognitive UX Layer: Interfaces such as Joule translating human intent into orchestrated execution.

The emergence of SAP BDC, combined with its recent strategic acquisition of Reltio, to access and harmonize external data is critical in this model. It extends NSCO beyond the enterprise boundary, allowing agents, whether inside or outside SAP, to operate on a shared, trusted data foundation.

If these layers are not aligned, orchestration breaks down. Agents may generate insights, but they cannot act reliably across fragmented systems or disconnected ecosystems.

From Decision Support to Supervised Autonomy

The transition to AI-driven supply chains is not binary. De la Torre outlined a maturity curve that moves from static ERP systems to fully adaptive, cognitive enterprises.

At intermediate stages, organizations begin to see task-level AI (forecasting, alerts), workflow-level automation, and agent-driven coordination across functions. At more advanced stages, supply chains become semi-autonomous, with AI managing execution while humans provide oversight, context, and governance.

Crucially, this is not about replacing people. As De la Torre emphasized, the role of AI is not to replace human decision-making, but to change how people engage with systems. “This is a vote of confidence for not replacing the individual, just how individuals are interacting with supply chain,” he said.

Human oversight remains central, particularly in environments where decisions carry operational, financial, or regulatory consequences. This becomes even more critical as AI begins to act beyond digital systems and into physical execution.

The introduction of “physical AI” was one of the more forward-looking aspects of the session. As De la Torre explained, it extends orchestration beyond digital systems into physical operations such as:

• warehouse execution aligned to real-time priorities
• production control based on supply constraints
• logistics coordination across yards and transportation networks
• closed-loop responses where detection triggers physical action.

In this model, AI goes beyond recommending decisions and begins to influence how work is executed in the physical world. This introduces new requirements around governance, safety, and trust, particularly in industries with complex or high-risk operations.

Starting with Targeted Use Cases

Despite the ambition of the NSCO model, De la Torre emphasized a pragmatic starting point. Organizations should not attempt full orchestration from the outset. Instead, they should focus on specific domains where AI can deliver immediate value and build momentum.

Planning and procurement are the most viable entry points, as both are data-intensive, exposed to volatility, and constrained by talent shortages. Manufacturing, by contrast, introduces additional complexity due to physical constraints and specialized systems.

This creates a sequencing challenge for organizations, not only where to begin, but how to scale in a controlled way as part of a broader, next-generation program delivery model. De la Torre’s recommendation is to build capability incrementally, then expand orchestration across domains. “Pick a lane and start developing agents to improve that function,” he said.

What This Means for SAPinsiders

Supply chain is becoming an orchestration problem. The next phase of SAP-driven supply chain transformation is not about optimizing individual workflows. It is about coordinating decisions and execution across systems, partners, and functions in real time.

Architecture readiness will determine AI success. Agent-driven supply chains depend on clean data models, integrated systems, and governance frameworks. Organizations with fragmented landscapes will struggle to move beyond isolated AI use cases and scale AI across the enterprise.

Human oversight remains central in autonomous systems. As supply chains move toward semi-autonomous execution, the role of the user shifts from operator to supervisor. Governance, trust, and accountability will become more important as AI takes on a larger role in execution.

Some or all of the services described herein may not be permissible for KPMG audit clients and their affiliates or related entities.