Task Mining & Process Mining: Why to Employ Them in Tandem?

article thumbnail

Share on social media

  
close

Download article as *pdf

Just give us your email where we will
send you pdf file article.

pdf

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. 

Table of Contents

    The Confusion Between Process and Task Mining

    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.

    What Is Process Mining

    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

    Minit Dashboards visualizing process mining results

    End-to-End Processes

    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. 

    Process Map

    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. 

    Minit AI-powered root-cause analysis feature

    AI-powered root-cause analysis feature

    Event Logs

    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.

    Free webinar: Discover new possibilities with the power of Process and Task Mining

    What Is Task Mining

    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

    Hierarchical mining capability enabling drill-down to the task level

    Understanding User Actions

    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.”

    Task Mining Works With 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 vs Process Mining Compared
    Task Mining vs Process Mining Compared

    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.

    Business Process Disassembled to Tasks and Processes-1
    Business process granularity: tasks and processes

    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

    These are the most common cases when process mining is used:

    • Order-to-Cash
    • Procure-to-Pay
    • Robotic Process Automation
    • Audit and Compliance Checking
    • Customer Journey
    • Process Optimization
    • Process Management

    Typical Task Mining Use Cases

    These are the most common cases when task mining is employed:

    • Discovering Automation Opportunities
    • Unifying Process Variants for Higher Efficiency
    • Improving User Experience
    • Discarding Redundant Steps or Actions

    3 Most Frequently Asked Questions About Process vs Task Mining

    How do you define a task and a process? What's the difference between the two?

    Processes and subprocesses are groupings of logically related tasks and decisions that have a starting and ending point. Task consists of steps performed by a machine or human with the goal for a case to move from one phase of the process to the next. The difference is in granularity of data, sources of data, ability to capture and analyze the data, outcomes, number of resources involved, and use cases.

    Can Task Mining be used to describe end-to-end processes?

    No because even though process mining can be used to process task mining data, other technologies such as OCR, text mining and specific ML algorithms must be equipped in order to analyze that form of data. Unsupervised (unfocus) user interaction recordings include an extremely high level of noise and very high granularity, which makes end-to-end process discovery inefficient and in most real life cases impossible.

    Does Minit provide task mining capabilities?

    Yes. To address this need, we have partnered with various technology partners and provide full scope of process intelligence capabilities.

    Cover photo by Adolfo Félix on Unsplash

    Picture of Michal Tomek

    Written by Michal Tomek Writer & Content Manager at Minit