Overview

Top-level synthesis of the BPM knowledge base. Updated when a source meaningfully shifts the global picture.

Scope

This wiki synthesises work across:

Process Frameworks & BPM — foundational BPM theory, lifecycles, BPMN.
Process Mining — event-log-based process discovery & analysis.
Predictive & Prescriptive Process Monitoring (PPM / PrPM) — forecasting next activity, remaining time, outcomes; runtime interventions.
Agentic BPM (APM) — autonomous-agent-driven process execution and orchestration.
Task Analysis & JTBD — decomposing work into tasks and jobs-to-be-done.
Literature Review Methodology — systematic review practice, incl. LLM-assisted review.
AI Capabilities & Adoption — how AI is reshaping knowledge work and tasks.
Philosophy of Science — theories of causation, limits of RCTs, methodological pluralism (PHI403 Anjum & Rocca) — meta-layer reframing the BPM / PPM / APM programs.
Riess research line — primary sources by the wiki owner (entities/mike-riess), see syntheses/riess-research-arc.

Current synthesis

The first ingested source — sources/2026-calvanese-agentic-bpm-manifesto — establishes Agentic BPM (APM) as an organising paradigm across these domains. APM's framework lets us place the other corpora in a coherent structure:

Traditional BPM & Process Mining → provide the operational and empirical foundation APM extends.
PPM → becomes one enactment capability an APM agent may invoke (the "predictive monitoring techniques" the manifesto cites for time-to-completion estimation).
XAI / causal process analytics → materialises the APM capability of explainability.
Self-adaptive systems / RL for BPM → materialises concepts/self-modification.
LLMs & conversational interfaces → materialise concepts/conversational-actionability, though the APM position is that LLMs alone are insufficient.
Task Analysis & JTBD → provide agent-level goal/outcome decomposition that maps onto the APM agent's intentional model.
Literature Review Methodology → meta-layer; governs how this wiki itself should select and evaluate sources.

Working thesis (v1)

The APM Manifesto supplies a vocabulary (frames, process-awareness, framed autonomy, PRA loop) that can be used as the backbone ontology for this wiki. Every subsequent ingest should ask: which of the four APM capabilities (framing / explainability / actionability / self-modification) does this source advance, and through what mechanism?

The intellectual lineage ABPMS (2023) → APM (2026) is the key narrative axis of the wiki — see syntheses/abpms-to-apm-evolution for the side-by-side. Both manifestos come from largely the same author collective; APM is a continuation, not a competing paradigm.

Philosophy of Science as critical friend

Every BPM / PPM / APM method encodes an implicit theory of causation: correlation → Hume (concepts/regularity-theory-of-causation); intervention → Woodward (concepts/interventionist-theory-of-causation); mechanism → Russo-Williamson (concepts/mechanisms-in-science); disposition → Anjum/Mumford (concepts/dispositionalism). The PHI403 course by Anjum & Rocca supplies the vocabulary to name those commitments and diagnose where a chosen method's symptom of causation will fail. This matters operationally: prescriptive process monitoring makes strong interventionist causal claims that PPM benchmarks alone cannot validate (an ecological fallacy); deep-learning PPM inherits the external-validity problems Cartwright & Hardie diagnose for policy RCTs (it-worked-there, will-it-work-here); probabilistic PPM output is ambiguous between frequentist, credence, and propensity readings that imply different actionability. Philosophy of Science therefore acts as a critical friend (Anjum & Rocca's phrase) to the operational programs — not a separate domain.

Open questions

index — full catalog
log — chronological history

Backlinks