IT-Based Decision-Making – Automated vs. Workflow-Supported Decisions

What is Decision Management? The topic is on the rise and keeps growing due to digital transformation. However, people are still confused due to the lack of a clear definition. We want to show and explain the two basic types of IT-based decision-making processes using a real-life example.

Share article:

Reading time:

Decision Management in Anti Money Laundering

The online banking of a financial institution executes millions of transactions a day. The bank checks all these transactions for money laundering. It uses predefined decision models or machine-learning algorithms to detect suspicious behavioral patterns, words or risky countries. The decision as to whether a transaction is suspicious is fully automated (Decision Automation, scenario 1).

Among the multitude of online banking transactions, a suspected case has now been identified. The bank now wants to shed more light on this case. This is where interactive decision-making comes into play (scenario 2). The transaction is handed over with all the relevant data to an employee who examines the individual case once again. The employee can then approve the transaction if it was okay or, if the suspicion is confirmed, take further action.

IT-based decision-making: Automated and workflow-supported decisions

Fully Automated IT-Based Decision-Making

The first type of IT-based decision-making is called decision automation. Complex decisions such as credit risk calculationsfraud detection or pricing can be carried out fully automatically as a stand-alone process or within a higher-level business process. These fully automated decisions are the precondition for digital transformation, because they allow interruption-free, “truly digital” interactions.

Decision automation is usually realized using web services (Decision Services) that are used consistently and reusably in the various processes, systems, applications, and channels. In this way, companies create uninterrupted, fully digital processes that enable end-to-end customer journeys.

More and more companies are combining expert knowledge and data knowledge to make better automated decisions. What does that mean? Expert knowledge refers to expertise that is in the minds of individuals or in (e.g. regulatory) requirements. Examples include lists of embargoed countries or internal lending guidelines – in other words: clearly pre-defined knowledge. Data knowledge, on the other hand, includes knowledge that is gained through modern analytic technologies. These are often complex relationships extracted from historical data.

For example, machine learning algorithms conclude that an earlier transaction made within 10 minutes in two locations 1,000 kilometers away, was fraudulently motivated. In this way, a company can significantly improve the hit rate of decisions (in the case of fraud detection).

The DMN Decision Requirements Diagram connects domain knowledge (left) with data knowledge (right) to detect fraud faster and more accurately.

Workflow-Supported IT-Based Decision-Making

Not all decisions can be completely automated. Especially case-by-case decisions may require human judgment. These semi-automated, workflow-supported decisions can also be optimized through IT – mostly using a workflow-supported, interactive decision-management application. Typical examples for this type of IT-based decision-making are the processing of credit applications, which often require the inclusion of different roles.

In the past, case-by-case decisions involved paper forms, data tapping and much coordination. Today, IT-based workflows provide case workers with the necessary data and documents at the right time. This allows caseworkers to focus on case clearance instead of dealing with manual (and erroneous) data collection. The decision management application documents the individual workflow steps in the system. The advantages of the IT-supported processing lie in the high efficiency and the audit-proof traceability of the decision. Sometimes, interactive decision management applications are also refered to as ‘decision support systems’. Although this is true, the term can be misleading, because the definition is more universal (see Wikipedia).

An example of business solutions that integrate both decision management scenarios is the Credit Risk Management Platform, which is based on the ACTICO Platform for Decision Management.