If you’re starting off with a new automation project, you need to understand the process you’re about to kick start so that you can get the best results out of your investment, right? If you chart them out, monitor and analyze them you can find out exactly where you need to tighten the screws for better results.
Task Mining allows you to gain an end-to-end process insight that combines business data with desktop data to create tremendous levels of automation and productivity.
A task mining solution for your business is an excellent way to monitor and improve the performance and productivity of your team no matter where they are working from.
You can oversee how every moving element (or each member) in a team is performing down to a specific task level. Here is a beginner’s guide for those of you who are new to task mining.
A2K Partners is a certified Celonis consulting and implementation partner.
Our practice leaders and business analysts have intimate understanding of what business leaders need to achieve with their systems in terms of productivity, performance, and financial objectives.
Our digital work automation services include all aspects of management consulting and technology implementation featuring process mining, intelligent automation and execution management via Celonis certified analysts, architects and engineers.
Watch the webinar on how to achieve your business objectives with Celonis here.
What Is Task Mining?
Task mining is essentially an automated collection process where user interaction data on a computing device is captured by software to be discovered, monitored and analyzed.
Task Mining collects user data by capturing things like user clicks, application actions, and screenshots. It also includes enterprise-ready user privacy features that enable allow listing of applications, hashing and obfuscation of sensitive data like passwords and personal information.
Typically, task mining is a huge part of businesses that have elaborate operational systems for all projects that involve several calculated steps to achieve specific goals. These operational systems tell you what steps have been taken at any given point in time.
What the captured user interaction data tells you is how each member of the activity, process or project has proceeded to finish steps and achieve their work. So, when you build a bridge between the business data that you have and the user interaction data, you get the whole informational story. It helps you find more opportunities to improve productivity and chart out a future roadmap to make changes. But, more importantly, it helps you answer specific questions like:
- How are your employees able to achieve the goals you set for them or they set for themselves?
- How much time they are spending on the tasks given to them?
- What are they doing when they have to wait for someone else to finish their part and hand it over?
- How many of your team members are being unproductive (even if it’s for no fault of their own) and what is causing the roadblock?
In fact, you can also use the concept of task mining even for processes that are not part of huge operational systems. That means, you can analyze the productivity of a certain action even if it is not a step in a particular process.
While process mining works at a macro level, task mining has the capacity to analyze the microcosm of the business altogether. All you need to do is install a local agent on every desktop that you want to monitor.
The software analyzes minute details like keystrokes, mouse clicks and switching between applications. That’s how you get information about idle time and productivity. This information is put into a specific context so that you can understand how the productivity of any specific team member on a certain desktop is being influenced. You can use this data to make changes to the process or the project.
There is an element of controversy here vis-a-vis data privacy since it deals with user interactions at a very minute level. It is one of the challenges with this technology and must be discussed at length with each team before the software is deployed. But more on that later.
How Is Task Mining Different from Process Mining?
Task mining is often looked at in congruence with process mining. Even though they are two separate areas of interest, when you combine the two, the benefits are amplified.
There is a better understanding of how to allocate resources and maximize potential. Many task-mining software come with features and tools like chatbots and robotic process automation or RPA. This kind of digital intelligence helps organizations identify talent that suits the specific projects coming their way. It can also be used to plug loopholes and remove bottlenecks to improve everyday efficiency.
In terms of operation, task mining in concept is similar to process mining. The idea with both concepts is to improve productivity but at different levels. And that is precisely where the differences lie—parameters.
Process mining calculates metrics that are specific to the steps that are involved in a process. These software analyze log files on the computer and look for problems with the key performance indicators or KPIs. Task mining looks at the performance of team members on a project at an individual level and analyze their productivity for the overall betterment of the project.
The Importance of Understanding Data
Before we delve deeper into the advantages of task mining, let’s take a look at why understanding data is an important step for every business.
Now, when you are entering the process of data mining, it is important to remember that the ability to understand the data at hand is just as important (if not more) as mining the data itself. For that to happen, you must employ the right techniques and make sure that the personnel in charge of the analysis of the data and drawing conclusions from that analysis are well versed with all the variables in the equation.
They must also know how to modify those variables and all the possible outcomes of such modifications. Having all that information, they must have the ability to pick the best plan of action, whether it is for a project or the entire business model.
Creating new problems while trying to fix existing ones is the last thing you want. In this case, we are talking about both task and process mining.
This means that the personnel require a good understanding not just of the way the software works and the goals they are meant to achieve but also have basic business knowledge. A good combination of both keeps the investment in these technologies sustainable and promotes growth.
How Task Mining Works
When it comes to task mining, there are several steps to take care of. Don’t worry. They are all quite simple to adapt to. Most software operate through a centralized admin portal which is used to capture and compile the user interaction data. That’s just the beginning. You take it from there through the many phases without causing any disruptions to the team while they are working.
Task Recording
The first step, as mentioned above, is to collect the user interaction data. In the olden days, this was done manually. Experts or bosses would sit down with each team member and do the analysis.
Traditionally, this was done by monitoring the progress of their work over time, asking questions and mapping the results. The end result would be compiled into a binder or a PDF. This a specific skill set and, of course, the most valuable resource of all—time.
But when you employ task mining, this data is collected by a software by recording them from each team member’s desktop. Now, capturing the data from individual desktops is a sensitive task. And thankfully, unlike the traditional way where human resources were deployed for the same, the way the software works is that it runs in the background while the team member is at work.
It has active access to all the applications that are being used at any given point in time. This means the software has the ability to collect data user actions like scrolls and number of clicks. All of this is collected along with timestamps. This data is enriched by adding screenshots and desktop data files.
It takes several screenshots and collects pieces of information at different time periods while the employee is at work. Some software also gets information about how users are interacting with specific applications and also monitoring their clicks.
The specifics of this data depend on the software you install. If you get something like Task Mining Client, you will be able to get this data by enabling the enterprise-ready privacy settings. This way, you get the data that will help you analyze performance and productivity without dipping into sensitive information like passwords.
Context Recognition
The second step after collecting the data is to place it in a context. The user interaction data is a collection of records that have been captured at arbitrary periods of time while they were working.
This is done so that the data is unbiased to make sure the analysis is fair to the said team member. But now it is time to look at it through a specific lens to work towards the desired outcome.
At this stage, you tap into the optical character recognition features that come with whatever software you are using. This is used to translate the user interaction data that has been captured from different tasks into information.
If you are using software like Celonis, this textual data is grouped into what are categorized as activities using clustering algorithms. It is a machine learning technique that is employed to group certain data points (like updating an order or doing research on a customer, in this case) and is regularly used by data scientists.
Clustering algorithms are used to analyze certain behaviors and understand them in a specific context. All the desktop data is collected, matched with the operational data and organized based on timestamps, IDs and such.
Advanced Custom Recording
Depending on the specific project, task mining software and the requirements of a process, sometimes you might need to do a detailed analysis. This might even lead to a need to record more data. Other times, it is just another step before you go ahead and capture the next set of data points to repeat the process all over again.
At this stage, since you have a basic level of data, you know what specific data points you need to capture. That is called advanced custom recording. This gives you data on a specific task and its execution by one or more members of a team.
Using this data, you can gather information on each executed or completed task. The parameters in terms of the steps in the process, the time spent on each of them and the amount of time that was lost due to roadblocks can be modified.
Analysis
Once you have the base level user interaction data (or advanced custom recordings), you can use it to analyze the level of productivity. It gives you an understanding of each of the steps in the process and what you need to do, if any, to improve efficiency and hence, the productivity of each team member.
When you use software like Celonis, you can integrate this data into their Intelligent Business Cloud too. This helps in improving process automation too. And you can integrate it with operational apps.
Where task mining automation ends, human analysis begins. While the software can do most of the work for you, it is not entirely free of human involvement. The modifications that need to be made to a certain project or tasks given to specific teams or even their individual members is ultimately the decision for a human. But right until that point, the task mining software will take care of it for you.
Benefits of Task Mining
We have briefly touched upon the advantages of automation in the overall setup of a business operation. Let’s look at how task mining is specifically helpful to projects of all sizes.
In today’s world, task mining is a huge asset because it is a great way to spot the loopholes in a specific work model. These inefficiencies are outside the realm of operational systems and the data can be used to optimize the levels of productivity. When it is used in tandem with process mining, the circle of monitoring and loopholes is complete.
To get there, you must understand that people and processes don’t compete with each other but instead, complete each other. That is the secret to transforming your business from a functional one to a hugely successful one.
You need to be able to deploy systems that can implement user interaction data along with the parameters like system logs that are used by process mining systems. This kind of business data improves service, reduces friction between multiple processes and leads to transformation in the digital way.
Improved Efficiency
We have already talked about this in the above sections peripherally. But the important thing to clarify to your employees is to let them know that they are not under surveillance. Task mining systems can be extremely beneficial to them too. And that is one of the cornerstones of getting their consent to provide the desktop data that is at the heart of task mining.
Some employees have trouble completing certain tasks not because they are bad at it but because the interfaces and techniques at their disposal are poorly designed. Task mining can help identify these blockages. And in the analysis phase, effective solutions can be designed for these problems.
Task mining software are also greatly useful when it comes to identifying unnecessary actions. If an employee is going through steps that can entirely be eliminated without adversely affecting the final outcome, it can be identified using task mining.
In a traditional setup, you might be scheduling roundabout meetings to get to the bottom of this issue. But with task mining, it is all done in the background. All you need to do is modify the steps in the process and both the team lead and the member can continue working without any friction.
Since the whole system is automated, businesses gain insights into the process and identify other tasks that do not need human resources. That way, you can save your best team members to do highly skilled work that requires their input and leave the drivel to the machines. This not only saves resources and time but also reduces man-made errors in the process.
And finally, while analyzing the actions of a team, task mining can also identify blockages in the process and reduce deviations.
Better Evaluation
With remarkable insights into performance and productivity along with minute details on the frequency of interruptions, businesses can evaluate their goals and make plans well into the future.
A part of task mining also involves analyzing protocol violations. This alerts the admin to lapses in the execution process. This includes but is not limited to the order in which a certain set of tasks are performed.
Remember when we said the desktop data is collected along with timestamps? It offers insights not just into deadline lapses but also improved efficiency in the future after modifications have been made.
However, if the initial estimates were wrong and extra steps were placed in the process, the admin gets to learn about them and eliminate them to make it easier for the team members.
The opposite is also true and you can add steps into the process if the data suggests that team members are repeating a step in the process outside the parameters of the instructions given to them.
Simply put, instead of having a team meeting where the members report about bugs, the team lead gets to do it using the data without having to interrupt a highly productive workday.
When this sort of task mining is combined with process mining, businesses can gain valuable insights into:
- Segregating specific duties
- Checking how the team is conforming to set standards
- Identifying loopholes in a process
- Identifying the root cause for said violations
- Overall analysis of costs and time
Productivity Analysis
In terms of ironing out the flow of processes, there are two types of productivity analysis:
The first is task productivity which helps the business look into the details of each task in a specific process. Think of it as every step in an algorithm. This analysis helps the team lead identify which users are being the most productive and what they are working on.
The second is resource productivity. This helps the team lead to understand the level of productivity coming from each team member. The data gives you a peek into the time each of them has spent on websites and applications. You also get an idea about what they are reaching for to achieve a certain goal.
So, if all or most of the team members are accessing a specific website for information on a topic, you can use this data to build a knowledge database for them. This increases productivity and also morale.
Potential Problems
Now, automation in many areas is a new concept and hence, it is not free of problems. One of the biggest concerns with data collection, of course, is privacy. There are smaller issues like team members spending an obscene amount of time on social media websites during work hours.
Strictly speaking, it should not matter in most cases if they are meeting deadlines and submitting quality work. In such cases, they might be reluctant to provide information on their browsing time.
Since a lot of data is captured through screengrabs, the employee’s entire screen is exposed to those who will be analyzing the data. This could be an individual but depending on the size of the operation, it could also be a team.
This poses larger problems like passwords that are not hidden or information concerning addresses, finances and also private patient information. In some parts of the world, this is considered to be surveillance and there might be legal restrictions on it.
The Bottom Line
The best way to deal with these potential data exposures is to do a mock with the team before deploying the software officially. You can ask your service provider for one and if that is not an option, make a request for them to generate a simulation with mock data so that the whole team can go through the process without giving up any private data. Whichever way you go, it is always best to assume that employee consent is a must.
