Select your language

Process Intelligence in Practice (AI generated)
Process Intelligence in Practice (AI genrated)

Positioning: Process Intelligence within the BPM Lifecycle

Process intelligence refers to the data-driven analysis, control, and continuous improvement of business processes. It combines traditional Business Process Management (BPM) approaches with data-driven methods such as process mining and decision analytics (; ).

Within the BPM lifecycle (identification, modeling, implementation, monitoring, optimization), process intelligence primarily addresses the monitoring and optimization phases, thereby closing the gap between modeled processes and actual execution ().

The key point is that process intelligence is not a large-scale project, but an iterative improvement approach aligned with evidence-based management.

Step 1: Selecting an Appropriate Starting Process

From an information systems perspective, this is a problem–fit question: not every process offers sufficient data availability, standardization, or improvement potential.

  • High case volumes (statistical analyzability, economies of scale)
  • Low variance (standardization and automation potential)
  • Clear decision logic (formalizable in DMN)
  • High operational relevance (visible value for stakeholders)

These criteria align with the principle of purposeful artifact selection in Design Science Research ().

Step 2: Capturing the As-Is Process as a Socio-Technical System

The as-is process is not merely a sequence of activities, but a socio-technical system consisting of people, technologies, data, and rules.

  • People (case workers, decision practices)
  • Technologies (line-of-business systems, interfaces)
  • Data (forms, registry access)
  • Rules (laws, administrative regulations)

Data collection is typically qualitative (e.g., workshops), complemented by initial data analysis such as throughput times.

The goal is an as-is model that captures both the formal structure (BPMN) and actual practice (workarounds, informal rules).

Step 3: Weakness Analysis as a Data-Driven Diagnosis

Process weaknesses can be systematically identified along common BPM antipatterns ().

  • Waiting times (queueing issues, lack of prioritization)
  • Redundancies (lack of process integration)
  • Media discontinuities (insufficient system integration)
  • Inconsistent decisions (lack of formalization)
  • Routing errors (missing workflow logic)

From an information systems perspective, these issues are often rooted in data and system design rather than purely organizational shortcomings.

Step 4: Designing the To-Be Process as an Artifact

The to-be process represents an artifact in the sense of Design Science and integrates process, decision, and data logic ().

  • Process logic (BPMN)
  • Decision logic (DMN)
  • Data structures (FIM, registry models)

Key design principles include:

  • Explicit modeling of decisions to improve transparency and automation
  • Standardization of data as a prerequisite for interoperability
  • Automation of standard cases in terms of straight-through processing
  • Targeted escalation of exceptions (human-in-the-loop)

Step 5: Piloting as Evaluation

The pilot phase corresponds to the evaluation phase in Design Science Research and serves to validate the developed artifact ().

  • Validation of functionality
  • Assessment of goal achievement (time, quality)
  • Analysis of unintended effects

Typical metrics include throughput time, error rates, process variance, and user acceptance.

Step 6: Scaling and Data-Driven Process Management

Following a successful pilot, the process is transitioned into continuous process management in line with a closed-loop BPM approach ().

Process mining plays a central role as a monitoring, analysis, and control instrument ().

  • Detection of deviations between planned and actual processes
  • Identification of bottlenecks and variants
  • Data-driven decision support

In addition, governance structures are required:

  • Process owner (process accountability)
  • Data owner (data accountability)
  • Decision owner (accountability for decision logic)

Architecture Perspective: Integration over Fragmentation

Process intelligence only realizes its full potential within a consistent IT architecture.

  • Integration of line-of-business systems via interfaces (APIs, XÖV standards)
  • Use of central registers in line with the once-only principle
  • Consistent data models as a foundation (e.g., FIM)

Without such integration, optimizations remain local and fail to scale.

Case Example: Business Registration as a Minimal Viable Transformation

A city with approximately 80,000 inhabitants reduced the processing time for business registrations from 14 to 4 working days.

From an information systems perspective, the key measures were:

  • Elimination of redundant checks (process integration)
  • Introduction of explicit decision rules (DMN)
  • Automated routing (workflow logic)
  • Reduction of media discontinuities (end-to-end digital processing)

The improvement was achieved primarily through process redesign rather than additional technology.

Conclusion: Process Intelligence as Evidence-Based Public Sector Modernization

Process intelligence is a management approach that integrates process modeling, decision modeling, data standardization, and data-driven analytics.

Adoption starts pragmatically, while impact unfolds through systematic scaling and continuous improvement.

Related Articles

References

Publications

Social Networks