The future of production planning: From monolithic systems to adaptive best‑of‑breed

Insights from the FIR e. V. at RWTH Aachen study “Future of Production Planning” – with practical implications for ERP, APS, MES, and sustainability.

Production planning is under pressure: volatile supply chains, a shortage of skilled workers, batch size 1 manufacturing, and rapid technological innovation are only a few examples. A recent study (online survey, Feb–Jun 2025) shows: The future belongs to integrated IT landscapes, a clearly changing role for planners, and a systematic handling of partially conflicting objectives.

Starting point: More dynamics, more goal conflicts, more responsibility

The study paints a consistent picture: There is improvement potential across nearly all planning areas – particularly in tactical, plant‑ and network‑level sales and rough production planning. On the operational side, optimizing disposition is viewed as a decisive lever for efficiency and delivery reliability.

In many companies, ERP still dominates as the central planning anchor. At the same time, the findings show that complementary systems are gaining traction – especially APS for operational tasks – while Excel continues to fill gaps in some areas (e.g., network‑level rough production planning). The emerging future: Best of Breed architectures integrating ERP, APS, MES, and, where needed, SCM.

Wherever conflicting planning objectives must be balanced simultaneously (delivery reliability vs. inventory vs. setup costs vs. utilization), APS provides clear advantages over ERP standard logic – not as a replacement, but as a complementary decision engine with defined responsibilities.

IT system landscape: Yes to best‑of‑breed – but with Discipline

Status quo: ERP systems remain the central hub of enterprise IT, yet practitioners report medium or higher improvement potential across almost all planning domains, with tactical tasks (e.g., plant / network sales and rough production planning) standing out.

The surveyed companies show two tendencies:

• Large organizations prefer Best‑of‑Breed but deliberately rely on standard functionality to keep integration complexity manageable.

• Small and mid‑sized companies place more weight on customizations; the choice between Best‑of‑Breed and “one‑fits‑all” remains evenly split.

Implikation für die Architektur:

• ERP remains the place for master‑data governance, MRP/disposition and finance‑related processes.
• APS handles detailed planning and optimization (sequences, start times, resource assignment, setups) and feeds stable, executable plans back to ERP/MES.
• MES provides real‑time feedback (actual durations, shop‑floor reporting, disruptions) for short‑term corrections and continuous learning from plan deviations.
• SCM supports tactical cross‑site decisions when capacities and material flows must be coordinated across multiple plants.

In a Best‑of‑Breed environment, APS becomes the operational decision engine – provided interfaces are robust, master data is maintained properly, and responsibilities are clearly defined. Value is created not just by algorithms but through a well‑orchestrated interplay between ERP, APS, and MES.

Role of planners: How far automation will go in the next five years

The study reveals a clear trend: production planning is on the verge of a major automation shift. All planning activities – from gathering information to re‑planning – are already considered generally automatable today. However, respondents expect that this potential will expand significantly over the next five years and become widely operationalizable.

Automation is projected to increase most strongly in activities that are highly data‑driven and process‑standardized. This includes, above all, the collection, storage, and transmission of planning‑relevant information. For these tasks, the study indicates the highest future degree of automation, as they can be clearly defined, represented in a structured manner, and reliably executed by systems. Operational rescheduling will also see deeper automation – particularly in situations where many conditions must be considered simultaneously and algorithmic optimization already demonstrates strong capabilities.

Auffällig ist, dass die Rangfolge der Automatisierbarkeit über die Zeit hinweg konstant bleibt:

• Information‑related tasks are – and remain – highly automatable.
• Operational decision‑making processes are moving closer to a high degree of automation, supported by algorithmic capabilities.
• Strategic tasks, such as defining planning objectives and key parameters, intentionally retain low automation potential. They require human experience, judgment, and organization‑specific expertise.

Respondents also emphasize that the path toward greater automation is hindered less by technical limitations and more by the effort required for data maintenance. Data quality, timeliness, and consistency are essential prerequisites for reliable automated planning outcomes. This creates a direct link to system selection: specialized planning tools, particularly APS systems, typically offer structured data pipelines, validation logic, and clear modeling approaches that make practical automation possible in the first place. An APS does not eliminate the need for data maintenance, but it institutionalizes it – ensuring that automation is not only technically feasible but also operationally stable.

Overall, the study underscores an important message: production planning in the coming years will become more automated, but not autonomous. Humans will remain central – though no longer primarily as “plan executors,” but as designers of target systems and overseers of automated decisions. Systems will take over repetitive tasks and complex, simultaneous optimizations, while planners will make strategic decisions, define quality standards, and orchestrate the interaction between ERP, APS, and MES environments.

Machine Learning (ML) and Large Language Models (LLMs): More precise planning and more efficient information work

The study indicates that artificial intelligence, particularly ML and LLMs, will gain substantial importance in production planning, though in different areas of application. While ML strengthens data‑driven analytics, LLMs primarily support information‑related processes.

Machine Learning: Greater impact in analysis and rescheduling

According to respondents, ML plays only a minor role today, but its importance is expected to increase significantly, by +53 percentage points, over the next five years. ML will be especially relevant for analyzing planning quality and for rescheduling activities, where more realistic forecasts and pattern recognition can enhance the robustness of decisions. A prerequisite for this remains a structured and precise data foundation – an aspect supported by specialized planning systems such as APS, which provide clear data models and consistent constraint representation.

Large Language Models: Accelerating the flow of information

LLMs unfold their potential wherever information needs to be condensed, explained, or communicated, while simultaneously filtering relevant insights from large datasets. Respondents expect their use to increase by 70 percentage points over the next few years. Tasks such as gathering and distributing information, as well as deriving improvement measures, are seen as particularly suitable. According to the respondents, LLMs do not make decisions themselves; instead, they support understanding of complex relationships and relieve planners of repetitive information-processing tasks.

Central bottleneck: Data maintenance

For both ML and LLMs, the greatest limiting factor is the effort required for data maintenance, ranking above acceptance or data‑privacy concerns. This underscores the importance of a structured system landscape. Systems such as APS provide a crucial foundation by mapping data flows and planning logic in a structured way, thereby increasing the reliability of data‑driven technologies.

Overall, companies expect ML and LLMs to make planning faster, more transparent, and more reliable – not through autonomous decisions, but through significantly improved information quality and decision support.

From control center to learning planning environment: A realistic target vision

The study results point to three defining characteristics of future‑proof production planning:

• Adaptive: The system learns from plan–actual deviations (ML‑based analytics) and continuously improves its solution approach. Based on predefined objectives, constraints, and guardrails, it iteratively searches for the best possible solution; planners define these boundaries and monitor the learning progress.
• Real‑time Integrated: Seamless data flows between ERP (demand/MRP/disposition), APS (optimization/scheduling), MES (feedback), and, where applicable, SCM – without system discontinuities.
• Multi‑criteria: In addition to delivery reliability, throughput time, and inventory levels, further criteria such as energy, material consumption, and waste are integrated as target criteria (transparency first, control impact thereafter).

Cultural prerequisites include trust in automated decisions, clear accountability for data quality, and transparency regarding the optimization logic. From a technical perspective, crucial enablers are robust interfaces (APIs) and and clearly structured, consistent master data that addresses the effort of data maintenance – which, according to respondents, remains the biggest obstacle. Without reliable master data, automation and optimization fail not because of the software used, but because of data quality.

Conclusion: Why Advanced Planning & Scheduling is becoming a necessity

The data points to a future of production planning that is integrated, adaptive, and multi‑criteria. ERP remains the foundation, APS becomes the decision‑driven complement for operational detailed planning, and MES provides the real‑time pulse. Technology supports planning – but it demands disciplined data work and clearly defined target systems. Organizations that establish a structured Best‑of‑Breed setup today, with APS as the core of the decision logic, make their planning more robust, transparent, and future‑proof.

Quelle: Center Integrated Business Applications / FIR Aachen GmbH

Study: Future of Production Planning