Process Discovery
A set of techniques—both manual and automated—used to identify, analyze, and visually reconstruct how business processes actually operate based on real data and observations.
Read summarized version with
What is Process Discovery?
Process discovery is essentially how organizations figure out what's truly going on in their workflows. Not the idealized version from policy manuals, but how work actually gets done day-to-day. Teams use a combination of interviews, direct observation, and increasingly, automated process discovery tools like process mining software to piece together an accurate picture of their current operations before attempting any improvements.
Here's the tricky part: most processes are basically invisible. Work moves through software systems, bounces between departments, and involves countless small judgment calls that no one ever wrote down. Sure, process documentation might exist somewhere, but it often describes what should happen rather than what does happen. Process discovery closes that gap by capturing the "as-is" state through evidence, not guesswork.
Why does this matter so much? Because you really can't fix what you can't see. Organizations that skip the discovery phase and dive straight into optimization tend to solve the wrong problems. When you surface how work actually flows through the business, including all those undocumented workarounds, hidden bottlenecks, and process variations people have invented, you get a foundation for improvements that have a real chance of sticking. This understanding becomes the starting point for effective process mapping.
Key Characteristics of Process Discovery
- Evidence-Based: Rather than trusting what people think happens, discovery relies on data from system logs, direct observation, or structured interviews to reveal actual process execution.
- Current State Focus: The goal is understanding how things work today (the "as-is" state), not designing some ideal future process.
- Multiple Techniques: Blends manual methods like interviews and Gemba walks with automated approaches like process mining and task mining for more complete coverage.
- Variation Detection: Effective discovery surfaces not just the main process flow, but also the edge cases, workarounds, and exceptions that teams have cobbled together over time.
- Objective Analysis: Automated process discovery strips away subjective bias by reconstructing workflows from system data rather than depending on memory or assumptions.
Types of Process Discovery
Manual Discovery
The traditional approach involves sitting down with workers for interviews, observing work as it happens, and running workshops where people map processes together. These methods capture institutional knowledge and context that systems simply cannot record. The catch? They take a lot of time, can be influenced by personal bias, and may completely miss digital workflows buried inside software systems.
Automated Process Discovery
Modern automated process discovery pulls data directly from system event logs and digital workflows. Process mining software reconstructs the actual sequence of activities based on timestamps and case identifiers. Task mining takes this further, capturing user interactions on desktops to fill in gaps where system logs fall short. Combined, these process discovery tools deliver faster, more accurate, and more thorough views of how work really happens.
Process Discovery Examples
Example 1: Insurance Claims Processing
Picture an insurance company trying to reduce claims processing time but struggling to pinpoint where delays actually occur. By running process mining on their claims management system, they uncover that 35% of claims loop back to an initial review stage. Nobody had documented this pattern before. The automated process discovery shows that missing documentation triggers these loops, which points to an obvious fix: collect better data from claimants upfront.
Example 2: IT Service Desk
An IT department decides to use task mining to see how technicians actually resolve tickets. What they discover surprises them: technicians have developed their own unofficial workarounds for common problems, solutions that never made it into any knowledge base. By documenting these discovered processes and making them the official approach, the department cuts average resolution time by 25%.
Process Discovery vs Process Mapping
These two disciplines complement each other but tackle different problems.
| Aspect | Process Discovery | Process Mapping |
|---|---|---|
| Purpose | Find out how processes actually work today | Visualize and document known processes |
| Starting Point | Unknown or undocumented workflows | General understanding of the process |
| Methods | Interviews, observation, process mining, task mining | Flowcharts, swimlane diagrams, SIPOC |
| Output | Data-driven understanding of current state | Visual diagrams showing process flow |
| When to use | Before optimization, when you suspect documentation is outdated | After discovery, to communicate and analyze findings |
How Glitter AI Helps with Process Discovery
Glitter AI takes a refreshingly practical approach to process discovery. Instead of requiring complex process mining infrastructure or scheduling endless interview sessions, it captures real processes as employees actually perform them. When someone uses Glitter to record their workflow, they're automatically creating discovered process documentation without even thinking about it.
This bottom-up approach means teams can surface how work really happens without launching massive IT projects. Each recorded workflow becomes concrete evidence of actual process execution, complete with screenshots, annotations, and step sequences. Organizations can then use these captured processes to spot variations between team members, identify inefficiencies, and build documentation that reflects reality rather than wishful thinking.
Frequently Asked Questions
What is process discovery?
Process discovery is a set of techniques used to identify and understand how business processes actually operate. It uses interviews, observations, and automated tools like process mining to reveal the current state of workflows based on real data rather than assumptions.
What is the difference between process discovery and process mining?
Process mining is one technique used within process discovery. Process discovery is the broader practice of understanding how processes work, while process mining specifically refers to using event log data from IT systems to automatically reconstruct and visualize those processes.
What are the main types of process discovery?
The two main types are manual discovery (interviews, observations, workshops) and automated discovery (process mining from system logs, task mining from desktop interactions). Most organizations use a combination of both approaches for comprehensive results.
Why is process discovery important?
Process discovery reveals how work actually happens versus how it's supposed to happen. This evidence-based understanding helps organizations identify bottlenecks, eliminate waste, and make improvements that address real problems rather than assumed ones.
What data is needed for automated process discovery?
Automated process discovery requires event logs containing three key elements: a case ID (to group related activities), activity names (describing what happened), and timestamps (showing when activities occurred). This data typically comes from ERP, CRM, or workflow systems.
How long does process discovery take?
Manual discovery can take weeks or months depending on process complexity and stakeholder availability. Automated process discovery using mining tools can generate initial results in days once event logs are accessible, though analysis and validation still require time.
What are common process discovery tools?
Common tools include process mining platforms like Celonis, UiPath Process Mining, and ABBYY Timeline. For task mining, tools like UiPath Task Mining and Automation Anywhere Discovery Bot capture desktop-level activities. Many organizations also use simpler observation and interview techniques.
When should you use process discovery?
Use process discovery before any major process improvement initiative, when documentation is outdated or missing, when preparing for automation, during digital transformation projects, or when you suspect significant gaps between documented procedures and actual practice.
What is task mining in process discovery?
Task mining captures user interactions on desktops—clicks, keystrokes, application switches—to understand work that happens outside core business systems. It fills gaps in event log data by revealing the detailed steps employees take within applications.
How does process discovery support automation?
Process discovery identifies which processes are good candidates for automation by revealing their actual complexity, variations, and exceptions. Understanding the true process state helps organizations avoid automating broken processes and focus on high-impact opportunities.
Turn any process into a step-by-step guide