What is Machine Learning?

Machine learning (ML) is a branch of Artificial Intelligence (AI) that employs various technologies and algorithms. ML relies on historical or precise data to make decisions or take action.

A practical example of machine learning is Amazon’s Alexa, a virtual assistant. Alexa uses voice recognition technology to associate sounds with specific actions. For instance, Alexa recognizes different sounds and responds accordingly through algorithms. While this appears intelligent, it involves minimal learning on the part of the technology.

What is Artificial Intelligence (AI)?

Artificial Intelligence (AI) traditionally refers to the theory and development of computer systems capable of performing tasks that typically require human intelligence, such as speech recognition, decision-making, and pattern identification. Today, “AI” is an umbrella term encompassing various technologies, including machine learning, deep learning, and natural language processing (NLP).

Although these AI technologies have enhanced user experiences, there is debate over whether they qualify as true artificial intelligence, as they do not possess independent thought. True AI should be able to learn and adapt using vast amounts of data to surpass human intelligence. Currently, AI is still in the research and development phase, and what we call “AI” today is closer to advanced machine learning.

Combining Machine Learning & Process Mining

When combined with Process Mining, machine learning adds intelligence to the reading, visualization, and analysis of data and processes in everyday operations.

Intelligent Process Mining Opportunities

Both Machine Learning and Process Mining depend on data. Machine Learning enhances Process Mining into what is known as Intelligent Process Mining. This combination offers several opportunities:

Descriptive Process Mining

The initial step in process mining is to provide visibility into how business processes operate. By retrieving logs from various information systems (ERP, CRM, etc.), process mining reconstructs and timestamps the flows and main stages of each process unit (e.g., an order, a customer, an invoice). Known as “process discovery,” this step is crucial for visualizing processes. Users can conduct preliminary analyses by examining specific deviations or anomalies observed in the process representations. However, due to the volume of data and the complexity of some processes, this analysis can be tedious and complex.

Clustering Process Mining

Machine Learning and AI add significant value at this stage by automatically detecting and highlighting major anomalies, deviations, and non-conformities in the process. For example, these technologies can compare the “as-is” process with the originally designed process, categorize the types of deviations, and evaluate their impact, such as “rework,” bottlenecks, and backtracking. This allows users to analyze the number of affected process units, the frequency of occurrences, and the temporal and financial impacts.

Diagnostic Process Mining

Once deviations and anomalies are identified, AI and ML use statistics and algorithms to pinpoint the primary causes. For instance, the tool might determine why a particular category of products is consistently delivered late, identifying issues such as a specific supplier, a particular region, or an upstream bottleneck. Users can visualize the main causes based on their importance and impact on the process.

Predictive Process Mining

Process mining collects and stores all data and logs from information systems. Machine learning uses predictive algorithms to analyze this historical data and forecast future anomalies. This enables precise visualization of future order flows, including each stage of an order and the forecasted delivery time within a supply chain process.

Intelligent Process Mining Benefits

• Anticipate friction points and process anomalies: Predict future problems in processes based on historical data to prevent them proactively.
• Predict the next steps in the process: Forecast upcoming bottlenecks in a supply chain process, allowing adjustments in stock levels or workforce to ensure smooth operations.
• Adjust the available resources during the process: anticipate future issues in processes based on historical data to improve resource management.
• Recommend process evolutions to gain in performance and productivity: In processes like procure-to-pay, identify and streamline time-consuming steps, especially when dealing with known suppliers.