Decision Automation enables organisations to automate the decision-making process in many different areas (i.e., within a process, application, department, etc.). There are many benefits in automating decisions. For instance, automated decisions will increase productivity and reduce risks and error rates in a decision-making process. It also increases the consistency of decisions as these will be less reliant on individual discretion.

Decision Automation tends be applicable to routine, repetitive decisions that occur in an organisation’s day-to-day activities. These kinds of decisions can be found in various functional areas and departments of companies. Perhaps you would be surprised at how many decisions can occur in the course of a business day! These types of decisions are called ‘operational’, which means that they drive the day-to-day operations of a particular business.

Operational Decisions can be driven by business rules knowledge (i.e., rule-based driven decisions) or can be based on data and information knowledge (i.e., data driven decisions using statistical, analytic and predictive algorithms). In advanced cases, both of these techniques should work alongside each other in order to automate a decision.

Rule-based Decision Automation

Decision Automation can be fully based on business rules. In many industries, such as insurance, finance, banking, and so on, routine and reparative decisions are automated in order to ensure that quality and consistency are not compromised. For example, in insurance many of the decisions that are made in underwriting, claims processing, pricing, discount, etc., can be automated using business rules automation.
Business rules automation becomes very critical when the decisions are in a regulated environment. When the ‘explainability’ and reasoning behind certain decisions are important for the customer, regulator, or other third parties, decision automation using automated business rules is particularly effective.

Data driven Decision Automation

There are many scenarios where the decision is not based on some rules, but instead is based on how a case in a certain situation is developing, and if there is uncertainty surrounding how the case may behave. For example, in a car insurance policy scenario, there is an uncertainty about how risky a driver may be. Or in cross-sell offers there is an uncertainty about how the party may respond to the offer.

These are the cases where a predictive model can work well in decision automation, especially when it is combined with rule-based decision automation. This will help in addressing any ambiguity in the decision-making process.

Human Intervention

Not all decisions in organisations can be fully automated, and some of these will require human intervention. Decision Automation should allow scenarios in which fully automated decisions are not possible because of ambiguities, uncertainty, and so on regarding the decisions. Instead, these require domain experts’ inputs and intervention.

Although it is tempting to build intervention handling as part of the applications logic or process automation, the caveat to these approaches is that handling intervention will introduce a disconnected decisioning experience, which should always be avoided.

Multistep and Long Running Decisions

Many decisions are instantaneous. For example, calculating a discount based on quantity. You provide the quantity and you receive a discount percentage. However, what if the discount is based not only on quantity, but also on customer loyalty? What if loyalty is determined by the total value of customer purchases of last year or month?

Decision Automation-Calculate discount with modelling dependencies between multiple decisions

Calculate discount with modelling dependencies between multiple decisions

This is a consumer-based example, but it shows the general idea in which a single decision about Discount requires the results of multiple other decisions. This is called multistep decisioning.
In multistep decisioning, there might be some cases where the next decision requires some external information, or some intervention from an expert. These may not be immediately available when the previous decision is being executed. So, transition from one decision to another may take time (e.g., hours, day, weeks etc.) and are not immediately ready to determine the results for any reason. These special types of multistep decisions are called ‘long running decisions’.

In these cases, the multistep and long running decisions can be implemented as part of the application or process automation, but will introduce the disconnected decisioning experience.

Disconnected Decisioning Experience

Building and deploying end-to-end decision automation is challenging, because it is easy to lose the end-to-end view of the decision and the results are lost or cannot be associated with a single valid business-oriented context.

The end-to-end view gives organisations the ability to ensure everything related to the decision is in one place. Therefore it can easily be measured, reconfigured, and monitored in order to ensure the semi-automated decisions deliver value to an organisation and these don’t become just external tasks requiring human intervention. Once the end-to-end view is available, after a while semi-automated decisions can be fully automated by removing the referral (intervention) part.

Book Your FREE 30 Minutes
Consultation Now

Your Name (required):

Your Email (required):

Your Industry:






Part of ensuring that the decision automation end-to-end view is not lost is to make sure that the methodology can find the right decisions, and the means to automate them is also in place – the Decision Centric Approach. By using the Decision-Centric Approach as an automation methodology, organisations can identify decision types, their maturity level and how to automate them.

Why Should You Care about Decision Automation?

We talked about decision automation benefits earlier, but in short, decision automation will help organisations to make better decisions. The “better decision” depends on different scenarios, but it consistently delivers accurate results for customer-centric decisions in a particular situation (situation-aware) that may involve calculation, data, and domain expertise and knowledge. It also removes the risk of decisions that rely on an individual’s discretion.

Decision automation always works and delivers great results in repetitive operational decisions. It increases the quality of decisions as well as the productivity of workers, as they will not need to worry about making those decisions. It is also an opportunity for organisations to focus workers time and effort in areas that cannot be automated.

Last updated January 10th, 2020 at 07:28 am, Published January 6th, 2020 at 07:28 am