Cloud Hosting

Machines, rather than people, will make data-intensive decisions and repetitive, manual tasks will be automated. However, the goal is not to replace human managers and the workforce. The aim is to improve their efficiency and productivity. Even when systems are operating at a high maturity, humans will still be necessary for decisions that arent made frequently, or those that require a high degree of human engagement. For other, more menial tasks, we can free up time for humans to focus on more value-based, creative or strategic thinking activities. IT processes, operational planning, marketing, sales, and accounting are all examples of areas that can be fully automated and AI:fied.

To identify the areas that automation and AI can make the largest contribution to human work in an organization, we start with what we call a discovery phase. During this phase, the technical and business requirements of a specific use case are identified. Examples of use cases include those that need encryption capabilities, for instance:

The traditional ways of working in a discovery phase is very manual and consists to a large extent of detailed interviews or workshops with subject matter experts (SMEs). One of the most significant challenges with manual process mapping is effectively extracting information from employees involved in a process. Knowledge gaps, employee validation, and human subjectivity will unavoidably arise when piecing together perceived reality to form a process map. Doing this manual work is also time consuming and takes away the SMEs time from more productive work. By automating the discovery phase, we reduce costs, remove any bias (which is more likely to be introduced through SMEs being interviewed by humans), and obtain a complete picture of processes that are otherwise hard to discover.

The discovery of use cases is done automatically with the help of technologies such as task mining, process mining and semantic search. These technologies leverage a data-driven and automated approach to a discovery, allowing the users to remain objective in their decision-making and drive improvements based on facts. These solutions take advantage of AI and machine learning (also Auto-ML) capabilities to understand more than a human ever could. Figure 1 shows the flow of events in a discovery process.

Figure 1: Automated discovery flow.

Task mining works by monitoring the actions users take. A recorder is installed on an employees computer to capture their interactions in the different applications they use, and it records data like keystrokes, clicks, data entry, copying and pasting, and other routinely performed actions to uncover how tasks are completed. Task mining focuses on tasks. Tasks are smaller components of a process or subprocess containing several steps, usually performed manually by employees at their workstations. Task mining enables organizations to discover, understand, analyze, optimize, improve, and even automate the tasks employees perform as they relate to completing larger processes. The main objective with task mining is to find ways to improve how tasks are carried out, or automate them to increase operational efficiency, reduce errors, and improve employee engagement.

Another technology that enables automated discovery is process mining. Unlike task mining, process mining focuses on processes. Process mining turns event logs from various IT systems into an instant visual model of a given flow. It depicts the as-is process based on facts with all its cases, variants, and paths. Process mining revolves around discovering, modeling, analyzing, monitoring, and optimizing end-to-end processes and their subprocesses. It allows us to identify any deviations and understand the reason for variation so we can make the necessary improvement or alignment. It can also be used to create key performance indicators (KPIs), identify root causes of variations and support the identification of process improvements.

Finally, a third technology that enables automated discovery is semantic search. In this case, a search engine is able to guess the semantic meaning of the input text using natural language processing (NLP) and other AI algorithms.

To evaluate the benefits of activity mining, we engaged on a proof of concept with an internal Ericsson team. The Ericsson Supply ECP Order and Delivery Management team used task mining to find 18 percent effort savings through automating processes and optimizing application usage. The team wanted to discover process automation and optimization opportunities and document selected processes to transform the as-is to to-be for automation. They also needed a way to understand usage patterns of the application portfolio.

The proof of concept was limited to a team consisting of 20 selected employees who worked across 10 markets. The business processes for this team are complex and are performed using a variety of applications, the majority of which do not produce application logs (for example, in Outlook, Excel, Adobe). Their requirements towards automation were:

In this manner, a task mining tool for process discovery across eight processes was deployed and the results were used to understand how the team was performing each process with minimal input required from the end users. In addition, the tool was used to generate process documentation. In a collaborative environment, SMEs and process experts can now transform the process from their discovered as-is state to the to-be state for faster and more accurate automation development.

After analyzing the interactions of the 20 users, the following insights were found: The tool identified significant savings opportunities of 8 percent through transformation, from eight business processes in scope. An additional 10 percent saving opportunity was identified associated with task effort, representing added potential automation opportunity to save efforts in tasks performed in Excel and SAP.

Processes and interactions are basics in the execution and scaling of digital transformation, new AI capabilities and new forms of automation. Activity mining enables our organization to have a complete picture of the as-is process and help identify the areas where automation and AI can make the largest contribution.

To understand the as-is process is critical to knowing whether its worth investing in improvements, where performance problems exist, and how much variation there is in the process across the organization. Activity mining allows for automation beyond a single technology, and is opening the door to the next automation and AI operating systems with hyperautomation at the core.

Read more about AI in networks

Read out blog post Democratizing AI with automated machine learning technology.

Heres how blockchain technology will pave the way to connected industries.

Excerpt from:

What are task and process mining in AI? - Ericsson