Share on social media
Learn the difference between tasks and processes, what techniques and technologies are used to analyze them, and why you should consider both to increase operational efficiency.
Over the course of developing and spreading our process mining solution among businesses and partners, we've witnessed a big deal of confusion and misinterpretations about how task mining and process mining correlate.
Are those interchangeable terms? Are they complementary? Or do they have nothing in common?
Simple answer is that process mining and task mining are both the right solutions for a business, yet each has a different purpose and application.
If used complementary, they help process analysts and business stakeholders analyze and optimize their operations on multiple levels.
To explain what processes and tasks are, a simple analogy from the world of book publishing will do.
Book publishing process starts with writing a book. It ends once the book was distributed to the shelves of bookstores. Looking at it from a high level point of view, this is what we can call an end-to-end process: from the first point which is writing (“starting end”) to the last point, distribution (“finishing end”).
In-between, you can identify dozens of subprocesses, and most importantly, tasks. But on that later.
Minit Dashboards visualizing process mining results
Process mining revolves around discovering, analyzing, and monitoring end-to-end processes and their subprocesses.
If we go with the book publishing analogy, process mining would focus on how the process is running among the author, publisher, editor, designer, proofreader, printing company, and distribution channels.
In enterprise terms, for example, a specific invoice is a process mining instance which runs through multiple departments, being handled by numerous employees and entering various enterprise systems.
The full journey of every invoice – or any other process instance for that matter – is discovered and displayed in a process map. Process map depicts the real process flow as it happens with all its cases, variants, and paths.
You can further analyze and monitor the process. As a Minit user, you can use AI powered Root-Cause Analysis, Business Rules, Simulations, and other Analyst features for deeper and more thorough process diagnostics and improvement suggestions.
AI-powered root-cause analysis feature
As a data source, process mining uses so-called event logs, which is data containing information about activity performed (eg. Purchase Order Creation), case (Purchase Order No 24395), and timestamp.
Process mining solutions take these logs from various IT systems such as ERP, CRM, Supply Chain Management, and others. Minit provides connectors to Salesforce, SAP, Oracle, Google, Amazon Web Services, and many others.
As opposed to process mining, task mining technique focuses on tasks – smaller components of a process or subprocess containing a number of steps, usually performed manually by employees on their workstations.
In terms of books publishing analogy, the activity of printing a book would include tasks such as creating the print layout, print layout review, print machine configuration, loading paper into printing machine, and so on.
Hierarchical mining capability enabling drill-down to the task level
Task mining records and analyzes the user's actions with a goal to understand the activity in more detail, optimize, improve or even automate these tasks or parts of it.
Typical tasks in a business environment include:
- Inserting and copying data,
- Downloading and uploading files,
- Logging into and navigating through business systems.
For example, within the accounts payable process, an invoice needs to be downloaded from an email, information copied to a table, and pdf invoice uploaded to the accounting system. Therefore, the tasks an employee performs include opening the email, downloading the attachment, entering information into the form, uploading the documents, saving the changes, forming the activity “Invoice Submit.”
User Interface Logs
Task mining data sources differ from process mining's event logs. The backbone of task mining initiatives is the UI log based on the recording of the user's interaction with their workstation interfaces.
In practice, this means recording the mouse clicks, keystrokes, copying and pasting, and other routinely performed actions. The technologies used include data mining, pattern recognition, natural language processing (NLP), and optical character recognition (OCR).
The result of task mining analysis is a depiction of a sequence of steps and their variants performed by the user, which can later be used as a skeleton for Robotic Process Automation initiatives.
Task Mining and Process Mining As Complementary Techniques
Both task mining and process mining techniques have their place within enterprises looking to optimize and automate their processes.
The right approach and combination of the two will provide better visibility and transparency into company operations on many levels. Why?
Because process mining bridges the gap between data from task analyses and actionable intelligence through insights built on a broader view of processes in relation to enterprise strategy.
To conclude, here's what Minit CPO Michal Rosik and EdgeVerve's Global Product Head Sateesh Seetharamaiah wrote in the The New Paradigm of Process Excellence whitepaper which focuses on the complementary aspect of task mining and process mining:
Typical enterprise processes have varying levels of granularity (L1-L5), and process mining focuses on L1-L3, analyzing event commits and application logs to power discovery, monitoring, and process improvement based on current organizational information. It then allows enterprises to review organization-wide process maps supported by a comprehensive understanding of process structure.
By tracking human-system interactions at the keystroke level, task mining adds on-ground intelligence, eliminating any subjectivity from the automation planning process.
So as a combined process intelligence stack, process and task mining techniques are exponentially more effective in ensuring automation success, helping enterprises:
- Discover, Analyze, and Automate processes that will deliver the most significant results
- Measure the impact of deviations with extreme granularity
- Identify imperfect process execution with a thorough understanding of the end-to-end journey
Typical Process Mining Use Cases
- Robotic Process Automation
- Audit and Compliance Checking
- Customer Journey
- Process Optimization
- Process Management
Typical Task Mining Use Cases
- Discovering Automation Opportunities
- Unifying Process Variants for Higher Efficiency
- Improving User Experience
- Discarding Redundant Steps or Actions
Cover photo by Adolfo Félix on Unsplash