Process intelligence combines process mining, workflow orchestration and a decision gate into one enterprise architecture that shows how work actually runs, coordinates what happens next, and keeps automation and AI agents inside auditable boundaries. This page collects my work on process intelligence and workflow orchestration: the independent landscape, the architecture guides, unified orchestration by industry, and how processes connect to data integration and trusted agentic AI.

From business process management to process intelligence

Business process management has existed as a discipline for decades, built on process models designed before the process ran. Process intelligence shifts the focus to runtime reality. The analyst market reflects the change. Gartner replaced its Magic Quadrant for Process Mining with a Magic Quadrant for Process Intelligence Platforms in May 2026, scoped to mining, task mining, modeling and monitoring. Forrester covers the orchestration side separately as adaptive process orchestration. I anchor both under process intelligence, because an enterprise needs all three jobs in one architecture.

• Process mining observes how processes actually run, based on event logs from ERP, CRM and other systems. It surfaces bottlenecks, rework loops and the variants that quietly bypass controls.
• Workflow orchestration coordinates what happens next across systems, people and AI agents, with state, retries, approvals and recovery.
• The decision gate enforces what each participant, human or AI, may do. What passes the rules proceeds, and what does not is escalated or rejected with a documented reason.

The three jobs feed each other. Mining finds the bottleneck, for example manual approvals that queue simple cases behind difficult ones. Orchestration routes each case to the right path, the gate decides which cases may run automatically, and every decision becomes a new event log that mining reads in the next cycle.

Unified orchestration across data, infrastructure, applications and business processes

Unified orchestration means one control plane for data pipelines, infrastructure automation, applications and business processes, with AI agents governed as steps across all four. Most enterprises run a separate orchestrator per team today, from job schedulers and data pipeline tools to BPM engines and infrastructure automation. Each one works inside its own silo, and the processes that cross silos run on scripts, tickets and manual handoffs. A shared control plane gives those cross-silo processes one place for dependencies, approvals, audit and recovery.

Workflow orchestration sits above data integration and complements it. Event streaming with Apache Kafka and APIs move data in milliseconds, and a Kafka event can trigger a workflow that runs for seconds, hours or days, waits for a human approval, and calls an AI agent as one governed step. Hard real-time control stays in PLCs and stream processing, while orchestration handles the long-running coordination around it. Together with data integration and trusted agentic AI, process intelligence forms what I call the Trinity of modern data architecture.

This is how the topics on this site fit together. The Process Intelligence Landscape covers mining and orchestration vendors. The data integration page covers the data underneath, the trusted agentic AI page covers the agents that act inside governed processes, and the industrial IoT page covers orchestration from the plant floor to the enterprise.

The Process Intelligence Landscape

The map below places process orchestration and process mining vendors in two bands, from platform-native to agnostic, with bubble size for market reach and darker fill for stronger agentic AI capability.