Workflow Automation in Manufacturing SMEs

Why Choosing the Right Process Is Key to ROI

Workflow automation is not an end in itself. In an operational setting, it must accomplish three things: reduce effort, shorten turnaround times, and increase process reliability. For this to succeed, a thorough pre-selection process is needed—one that pays off in multiple ways:

• Faster Results: Processes with a high degree of repetition and clear rules deliver visible benefits early on.
• Lower risk: Complex automation projects for special cases with little leverage can be specifically avoided.
• Clearer implementation logic: It becomes apparent early on whether a process is suitable for full automation, assisted automation, or—initially—only for standardization.

This is especially crucial for small and medium-sized businesses. Resources in IT and line organizations are limited; therefore, it shouldn’t be the loudest process that wins, but the one with the best balance of feasibility and impact.

Step 1: Quick preselection using a 2×2 matrix

The first stage serves as a preliminary filter. Each process is classified along two dimensions:

• Frequency: How often does the process occur?
• Process Complexity: How many exceptions, decisions, and variations does the process involve?

This leads to four possible areas and recommendations for action:

1. High Frequency, Low Complexity: Full Automation
   This is where the classic candidates for automation are found. The process occurs frequently, follows consistent rules, and has few exceptions—an ideal combination for direct automation. Typical examples include workflows in which data is fully available, approval rules are clearly defined, and no in-depth case-by-case evaluation is necessary.
2. High Frequency, High Complexity: Assisted Automation
   These processes occur frequently but cannot be fully captured in rules. The system can collect data, assign tasks, monitor deadlines, and make suggestions; however, the final decision remains with the human user. The goal is not fully automated “black-box” processing, but rather to provide a noticeable reduction in the workload for the business units.
3. Low Frequency, Low Complexity: Quick Wins
   Not every simple process needs to be automated. If a process occurs infrequently, clear standardization is often sufficient: templates, forms, checklists, or clearly defined responsibilities. The benefit here lies more in consistency than in the depth of automation.
4. Low frequency, high complexity: Decomposition. Taken as a whole, these processes are usually poor candidates for automation. The better approach: break the workflow down into subprocesses and specifically automate the rule-based steps. The insights gained from this approach are often more valuable than direct implementation, because while not every process is immediately suitable for automation, almost every process contains subcomponents that can be automated.

The detailed evaluation based on six criteria

Processes that appear promising in Stage 1 are evaluated in greater detail in Stage 2. To do this, each process is rated on a scale of 1 to 10 across six dimensions; the scores are then weighted individually and aggregated into a total score. The higher the score, the better suited the process is for workflow automation.

An overview of the six dimensions:

1. Frequency
   How often does the process run in daily, weekly, or monthly operations? The higher the frequency, the greater the impact of automation.
2. Rule Definition
   How clearly can the process be described? Are there clear “if-then” logic patterns, or do exceptions and informal decision-making processes dominate?
3. Technical Feasibility
   Is the relevant data available in digital form? Can the systems involved be integrated? Where are the data discontinuities?
4. Process Stability
   Is the process well-established, or is it currently undergoing a restructuring? Automating unstable processes ties up resources without delivering reliable results.
5. Organizational Maturity Level
   Are roles, responsibilities, and handoffs clearly defined? If it’s already unclear in the current state who makes which decisions, even a workflow won’t solve the problem.
6. Business Impact
   What specific benefits does automation offer? For example, in the form of time savings, error prevention, shorter turnaround times, or avoided downtime.

Sample weighting as a starting point:

These weightings are intentionally not a rigid model. A company facing a severe shortage of skilled workers may place greater weight on business impact, while one with a heterogeneous IT landscape may place greater weight on technical feasibility. What matters most is that the weighting is determined deliberately and discussed as a team.

Four Strong Case Studies for Small and Medium-Sized Manufacturing Companies

To make the framework more tangible, we’ll classify four typical processes that are common in manufacturing SMEs and that perform particularly well in terms of quality.

Processing of Incoming Invoices, Including the Approval Workflow, in Financial Accounting

In medium-sized industrial companies, incoming invoices are still frequently processed manually: Invoices arrive by mail or email, are printed out, manually posted to accounts, forwarded via internal mail, and eventually entered into the ERP system. This process, which involves a change in media, ties up resources in the accounting department and is prone to errors; at the same time, it is highly amenable to standardization.

What makes him so strong?

• High volumes and regularity: Incoming invoices arrive continuously, often numbering in the hundreds or thousands per month. Even small efficiency gains per invoice add up to a noticeable reduction in workload.
• Clear, verifiable rules: Verification steps such as “Is the supplier known?” and “Is the invoice amount within the budget limit?” can usually be clearly defined as rules. Value thresholds, cost centers, supplier status, or deviation tolerances (e.g., ±5%) are ideal criteria for automated decisions.
• Direct Impact on Transparency and Liquidity: An automated workflow reduces processing times, prevents lost invoices, and improves visibility into outstanding liabilities. Discount periods can be taken advantage of more consistently, and late fees can be avoided.
• Strong technical foundation: Many data points are already available: supplier master data, purchase orders, goods receipts, cost center structures, and authorization schemes are all stored in the ERP system. Modern OCR and capture solutions can extract invoice data in a structured format from PDFs and scans and transfer it to the workflow.

Depending on the complexity of the supplier base and the exception logic, the spectrum ranges from largely fully automated (“no-touch invoice”) to highly assisted automation, in which the system automatically processes standard cases while the accounting department focuses on exceptions. In both scenarios, the potential for boosting productivity and process reliability is particularly high for medium-sized industrial companies.

Handling Complaints and Nonconformities in Quality Management

Complaints and quality deviations typically follow a formalized process: recording, evaluation, assignment, tracking of corrective actions, feedback, and documentation. While the technical root cause analysis remains the responsibility of the relevant department, everything else involved is a classic workflow use case.

What makes him so strong?

• Robust workflow logic: Many companies already use defined quality processes (8D, CAPA, etc.) in this area.
• Clear deadlines and the need for escalation: That’s exactly what a workflow is designed for.
• High level of transparency: Status, responsible parties, and pending tasks are visible at a glance.
• Tangible benefits: Response times are reduced, tracking becomes more reliable, and compliance improves.

Since not every technical evaluation can be automated, this process is particularly well-suited for assisted automation. The system controls the workflow, and a human makes the substantive decision.

Creating and Modifying Material Master Data

Material master data is an often-underestimated bottleneck in everyday industrial operations. New materials or changes must pass through design, purchasing, work planning, scheduling, and controlling. However, without a well-defined workflow, this leads to follow-up questions, data gaps, and delays that directly impact planning, procurement, and production.

What makes him so strong?

• Clearly defined handoffs: The steps follow a logical sequence with clearly defined responsibilities.
• Significant organizational impact: The workflow fosters accountability across departmental boundaries.
• Significant Business Impact: Incorrect or incomplete master data has a direct impact on downstream processes.
• Good process stability: Especially in the master data environment, well-defined standards retain their value particularly well.

This process isn’t a spectacular digital topic, but it’s extremely effective from an operational standpoint. That’s exactly what makes it so attractive to small and medium-sized businesses.

How to Create Your Own Automation Roadmap

A simple but effective approach can be derived from the framework:

Identify potential processes
. Ask specific questions in Purchasing, Quality, Work Planning, Maintenance, and Supply Chain about processes that involve a high degree of manual work, frequent follow-up inquiries, or long lead times.

Categorize Processes at a High Level
Use the frequency-complexity matrix to identify initial candidates and weed out obviously unsuitable processes early on.

Evaluate Processes in Detail
Evaluate your shortlist across the six dimensions and define your weightings deliberately, not haphazardly.

Starting with a clear winner
Don’t choose the most prestigious process, choose the most sensible one. Early successes build trust and facilitate further scaling.