Manjunath Nagaraju, on January 07, 2026, 05:00 AM
Putting AI to work in SAP Digital Manufacturing
AI in manufacturing doesn’t need to start by shooting for the moon. For many plants, the first step is just to make better use of the data they already collect in SAP Digital Manufacturing (SAP DM) and the broader SAP landscape.
From the standpoint of an SAP system integrator, we see AI as another layer in the digital manufacturing stack. When applied to a foundation of well-governed processes and data, AI can simplify how people work with SAP DM, shorten investigation time on the shop floor, and connect operations more tightly to planning and supply chain decisions.
In this post, we’ll explore what’s available today in and around SAP DM, including:
- Joule, SAP’s AI copilot
- Custom agents
- How manufacturers can pragmatically apply these capabilities
What AI Looks Like in SAP DM Today
SAP DM already uses AI to improve productivity and quality on the shop floor. SAP focuses on AI and machine-learning scenarios that help manufacturers predict quality issues, gain clearer operational visibility, and respond more quickly to production changes.
At a high level, there are three main layers to consider:
Embedded AI Features in SAP DM
These include AI and advanced analytics that support:
- Early detection of quality issues
- Performance monitoring across lines, work centers, and shifts
- Predictive maintenance and anomaly detection by combining machine data with execution context
In most cases, these scenarios use data captured by SAP and processed through SAP BTP services or SAP Analytics Cloud.
Joule Inside SAP DM
Joule is SAP’s AI copilot that lets users ask questions in normal language instead of navigating menus or reports to access and synthesize information across SAP solutions. Users can open Joule directly from the SAP DM launchpad and ask contextural questions about production, orders, and KPIs. In SAP DM, Joule focuses exclusively on helping people find and interpret information.
Custom AI Agents Via Joule Studio
Joule Studio provides an “agent builder” for creating custom AI agents that are based on your organization’s processes and data. These agents can call predefined functions (like “get operation cycle times”, “retrieve open NFCs”, or “compare planned vs actual”) and automations, work across SAP and non-SAP systems, and are configured through a guided, low-code interface.
For manufacturing, this creates a path to move from “ask a question, get an answer” towards “describe and outcome, have an agent assemble the steps and systems needed to support it.”
From an SAP Partner perspective, these three layers are most valuable when the are aligned with well-defined process models for SAP DM and a realistic roadmap for AI adoption.
Practical Shop Floor Use Cases
Below are a few examples of how AI capabilities within SAP DM can be used today. The exact scope depends on your licenses, landscape, and data readiness, but the patterns tend to be consistent.
Accelerated Investigations for Production Supervisors
The challenge: Supervisors spend a non-trivial amount of time navigating multiple screens to understand what’s happening on a line and why a particular order, lot, or work center is underperforming.
With the help of AI they can:
- User Joule to ask questions like:
- “Show me the current OEE and scrap rate for Line 3 for this shift.”
- “Which SFCs are currently blocked on Work Center A10 and why?”
- Combine embedded AI in SAP DM with analytics to highlight unusual pattern, like:
- Sudden changes in cycle time for a specific operation
- Repeating defects tied to a material, tool, or work center
As far as system integration requirements to enable these scenarios, the work is to ensure that the underlying data model, master data, and event capture in SAP DM are adequate to enable Joule to provide trusted responses and actionable insights.
AI-Assisted Root Cause Analysis for Quality Engineers
The challenge: Quality engineers often need to manually correlate data from SAP DM, lab systems, equipment logs, and ERP to understand why a defect or excursion occurred.
With the help of AI they can:
- Use SAP DM’s AI features and connected analytics to flag quality excursions early, based on process data and inspection results
- Use Joule to:
- Pull together a narrative view of a specific quality event, including affected orders and operations
- Summarize historical patterns around similar defects
- Build a custom Joule agent that:
- Accepts a simple description of a defect or NC
- Collects relevant SAP DM data (e.g., operations, parameters, equipment, operators, time frames)
- Prepares a “first pass” investigation package for the engineer
SAP Partners can help design these agents so that they respect existing quality workflows and don’t bypass required approvals or regulatory guardrails.
Custom Joule Agents: Where to Start
Joule ships with ready-to-use agents for common business scenarios, and Joule Studio allows you to define your own agents using a guided workflow.
For SAP DM, we see a couple of good starting points for custom agents:
“Explain This” Agent
This agent could help supervisors and engineers understand what changed and why in a particular time window or work center. The agent would require the following inputs: time range, line, work center, or order to take actions like pulling KPIs, alarms, NCs, and parameter deviations to provide a narrative summarizing the changes.
Exception Coordination Agent
This agent would exist for the purpose of orchestrating the response when SAP DM detects repeat quality issues, failures, or delays. The agent could take then actions like:
- Checking related orders and customers
- Verifying inventory and alternatives
- Proposing tasks for quality, planning, and maintenance teams
As an SAP Partner, our role is to make sure agents are based on well-defined processes, work within an approved data scope, and are bound to appropriate security and authorization rules.
SAP Partner Support for SAP AI
AI capabilities are only as useful as the processes and data behind them. Partner support covers:
Data Readiness & Governance
Ensuring that master data, routing structures, and work center definitions in SAP Digital Manufacturing remain consistent is a foundational requirement for meaningful AI outcomes. When equipment integration and data capture are aligned and reliable, AI scenarios have the necessary inputs to produce trustworthy insights rather than unclear or contradictory recommendations.
Process Modeling & Templates
Mapping end-to-end manufacturing processes allows AI agents to understand the full operational context rather than responding to isolated transactions. By designing templates and process patterns that can be repeated across lines and sites, manufacturers create a consistent structure that supports scalable AI use instead of one-off implementations.
Landscape Integration
Successful AI scenarios in SAP DM require more than plant-level visibility. By integrating SAP DM with SAP S/4HANA, SAP Analytics Cloud, SAP BTP services, and relevant non-SAP systems, organizations ensure that AI can reference the right data sources and trigger downstream actions. In this coordinated landscape, AI-initiated responses such as quality alerts or maintenance activities can move beyond simple notifications into structured follow-up steps.
Guardrails & Change Management
AI agents must operate within defined controls so that required approvals, compliance checkpoints, and regulatory boundaries are not bypassed. Establishing these guardrails early helps maintain traceability and audit readiness. Alongside control measures, focused training and adoption support ensures that supervisors, engineers, and planners understand how to use the new capabilities and trust the recommendations being surfaced.
Getting Started With AI in SAP DM
If you’re considering AI in the context of SAP DM, you don’t need to tackle everything at once. A pragmatic approach might look like:
Identify the Highest Priority Use Cases
Start with specific questions like, “How do we reduce investigation time when a line underperforms?” or “How do we identify quality issues earlier in the shift?”
Assess Data & Process Readiness
Verify that the relevant data is captured and clean, and that the processes in SAP DM reflect how the work actually flows on the shop floor. Are they truly standardized?
Design a Proof-of-Concept (POC)
Use Joule Studio to build your first agent around a clear scenario, such as exception handling. Then connect it to existing systems through SAP BTP.
Iterate with Defined Metrics
Track outcomes such as time to decision on exceptions and use the results to guide the next phase of functionality and next wave of agent creation.
AI in SAP DM isn’t about replacing your MES or automating every decision. It’s about making it easier for people to see what’s happening, understand why, and use that information to coordinate a more efficient response.
When AI is based on a well-architected SAP DM implementation and thoughtful process design, it becomes another practical tool for improving flow, quality, and collaboration across the factory.
If these topics are relevant to your current roadmap, we invite you to connect with us.
Be the first to comment.