Solid business processes are the bedrock of a well-functioning operations team. They help new employees onboard quickly, give structure and predictability to tasks that drive revenue, and make it much easier to ensure the execution of the processes is efficient. But the reality is that creating and maintaining good processes is challenging because business requirements and employee behavior naturally drift over time. As a business grows and changes and the people involved in the operations change, there is increasing variation in both the processes and how employees execute those processes, resulting in increased opacity of how work actually gets done.
Both Zeitworks and Process Mining platforms (e.g., from companies like Celonis, UIPath, SAP, IBM, and others) offer technology platforms that help operations teams understand how processes are executed. The fundamental goal of both platforms is to help operations teams understand the inefficiencies of their process design and execution behaviors, so their efficiency can be improved.
Desktop processes typically span the usage of multiple applications (e.g., web applications, office productivity tools, communications, native Windows applications, etc.). As a result, there is no process-centric or application data to mine for process execution patterns or behavior. Instead, a set of technologies and techniques called “Task Mining” automatically collects low-level events generated by employees as they go about their daily information processing work. (See the “Task Mining” section below for more info). Once a stream of data is captured, additional data mining and analysis capabilities interpret the data, which can be used to analyze both process structure and team behavior and productivity.
Zeitworks can identify, measure and track unique instances of processes and not just the steps that make up a process. Note that this capability is similar to traditional process mining, but the Zeitworks platform can do this with desktop processes using data collected with task mining technologies.
Importantly, the process executions are confined to these systems and do not span desktop application usage. Process mining algorithms consume trace log files from enterprise systems that result from the execution of business processes such as accounts payable/receivable. These log files are typically well structured and contain a time-stamp and a case ID representing the specific unit of work. Historical log files can be retrieved and analyzed by process mining algorithms which will produce process maps of individual or typical processes. This information can then be used by a business process analyst to assess the current process structure and execution paths, providing a baseline for understanding and improvement.
Because the input data source for process mining technologies is the transactional logs of centralized enterprise systems, the process map output does not include insights or details into employee-specific productivity or behaviors.
Input device data (mouse, keyboard), application UX, and web/browser events.
Modern machine learning, incl. data denoising/ smoothing, statistical inference, sequence/ pattern mining, NLP, and others.
Continuous measurement and improvement of operations teams, including insights into people, processes, and productivity.
Identify/uncover process structure and execution variance and costs, application usage patterns, and employee behavioral inefficiencies.
Process improvement, simplification, redesign, team training, optimal work assignments, outlier identification, etc.
Data is voluminous, noisy, and has high variance. Machine learning technologies produce probabilistic results.
Enterprise System log files (ERP, CRMs, etc.).
Network graph mining algorithms, process modeling techniques.
Analysis of processes operations that use a centralized enterprise system.
Visualize past process structure. Help identify bottlenecks and the costs of each process section/step.
Identify where, when and why the process deviates from its ideal version.
Provides a baseline for redesign or engineering.
Silo’d Analysis: Process mining is limited to enterprise (e.g., ERP) systems and has no awareness of out-of-system activity.
Back-end integrations: Requires licensing from ERP and CRM companies to mine data.
Zeitworks and some RPA (Robotic Process Automation) platforms utilize a set of technologies and techniques commonly referred to as “task mining.” Task mining software collects information about the execution of work performed at a desktop computer by observing and analyzing behavior and event data generated by users performing their job. Importantly, task mining captures repetitive business process work typically involving several desktop applications, including web and communication applications and productivity tools. This data may be from individual users or a cohort of individuals (as, for example, in a call center) and can take the form of data from input devices (keystrokes, mouse clicks), data and events from applications (application windows and input controls), and images (screenshots). After capturing data, additional mining, machine learning and analysis capabilities interpret the data by applying sequential pattern recognition, graph/network analysis, and optical character recognition to extract meaningful business insights. Task mining can help enterprises identify process and team inefficiencies, automation potential, and opportunities to lower operational costs and enhance the employee experience.
Note that several RPA platforms offer features that use task mining technologies to help users record and retrieve the optimal steps of a process so that those steps can be more easily converted into an RPA bot. As can be seen below, Zeitworks uses task mining technologies for a much broader set of applications and utilities.
Zeitworks AI-powered Business Process Intelligence Platform observes and measures teams, processes, and productivity continuously, empowering leaders to keep their fingers on the pulse of their operations.
Zeitworks provides operational data and insights continuously and longitudinally across the whole team. This always-on and comprehensive approach provides insight into time, costs, and efficiency of unique process executions, application usage, and employee work patterns and behaviors. The continuous nature of Zeitworks data collection and analysis enables operations leaders and their teams to make iterative improvements and optimizations. Several key capabilities distinguish Zeitworks from other products and platforms in the market:
Zeitworks runs continuously and not just for limited or discrete periods. Because the Zeitworks desktop sensor software runs in the background, passively measuring teams and operations, no work is ever stopped or interrupted. Managers can keep their fingers on the pulse of their teams as their work and business change.
Zeitworks can identify, measure and track unique instances of processes and not just the steps that make up a process. Note that this capability is similar to traditional process mining, but the Zeitworks platform can do this with desktop processes using data collected with task mining technologies. This means that Zeitworks can identify, measure, and track individual cases, claims, customers, applications, orders, or whatever unit of work the processes produce. This analysis provides a new level of detail into an organization's business metrics.
Zeitworks reveals insights about the unique instances or executions of your processes (e.g., person A’s loan vs. person B’s loan) and not just the generic steps of a process
Zeitworks measures not just your processes but also the whole team. Discover who are the best performers and the behaviors that differentiate them. Learn who needs further training and coaching.
The Zeitworks desktop sensor software is easy to install, and the Analytics Portal is intuitive to use and requires no special training, learning curve, or operation of arcane tools. Just log in and start learning about how work really gets done!
The Zeitworks Analytics Portal provides a wide depth and breadth of data and insights about an organization’s people processes and productivity.
Process mining is a technique used to evaluate business processes that are executed on and within centralized enterprise systems. Process mining technologies analyze system log files which contain digital “footprints” of executed process activity steps. Logs file entries include case IDs, time stamps, and activity categories. The output from process mining tools is typically a process map (see image).
Through analysis of this process map, an analyst is able to Identify where, when and why the process deviates from its ideal version. They can also understand time lapses between each step of a process and, therefore, identify areas ripe for enhancement or redesign. Because the input data source for process mining technologies is the transactional logs of centralized enterprise systems, the process map output does not include details or insights into employee-specific productivity or behaviors.
An example process map extracted from log files of an enterprise system and analyzed by process mining algorithms.