Decision Automation: A Simple Puzzle That Can Teach Us A Lot

In this article, we want to illustrate the importance of augmenting automated decisioning and business intelligence systems with human intelligence.

Here’s a simple logic puzzle for starters. Please look at the picture at the page bottom. There are four cards lying on a table in the way pictured below. Identify which card(s) you would flip in order to verify the validity of the rule: “If a vowel is written on one side of the card, then there’s an even number written on the other side.”

The majority of people answer that checking the other side of the “A” card is sufficient to verify the rule. Some people would flip two cards: the ones with “A” and “2” on them.

“A” is a vowel, and we don’t know what’s on the other side of the “A” card, so we definitely need to check this one. What about the “2” card? This is an even number; and our rule says that if a vowel is written on one side of the card, then an even number is written on the other side…But it doesn’t say anything on whether the even numbers should be written exclusively on the cards with a vowel on another side. Therefore, there’s no need to examine the other side of the “2” card.

Does this mean that it is enough to flip the “A” card? No, it doesn’t. We need to check if the other side of the “7” card has a vowel written on it. This would contradict our rule. Therefore, the correct answer is: we should flip the “A” and the “7” cards.

This task is formulated by cognitive psychologist Peter Wason and is called “Wason selection task”. Wason’s research has shown that less than one out of five respondents can give the correct answer.

People tend to make conclusions based on certain facts they know – in this case these facts are represented by symbols on the cards that are visible to us. People are leery of speculating on what is unknown or has a high level of uncertainty.

But exactly the information "hidden" in the data is essential to make the right decision. This is why it is necessary to embed artificial intelligence in every human decision. As James Taylor, a leading expert and independent consultant in Decision Management, has said in our recent conversation, automated decisioning systems should not simply pass the initiative to humans to handle uncertain or ambiguous situations; they should optimize human workload, provide human participants with all that is necessary for taking informed decisions, and be ready to 'take back' transactions if automation becomes possible again due to a change.

Let us provide a few examples to demonstrate that automated decisioning systems can and should flawlessly interact with humans.

New business origination (lending, contract origination, customer acquisition): decision automation technologies allow instantly rendering the right decisions for pricing parameters, which are optimal for the client and the business. This decreases customer acquisition costs, improves operational efficiency and overall profitability. Companies that implement decision automation for new business origination gain an information advantage over competitors by utilizing more holistic information about the potential customer. This increases staff productivity and helps drive customer loyalty through fast and accurate service.

Customer lifecycle management and marketing activities: with automated decisioning, companies can better understand changing customer demands, gain insight into sales opportunities, and keep their focus on the most productive products and services, identify the causes of churn & improve retention, capture cross-sell opportunities, predict product profitability across different target groups. Automated tracking of customer behavior and decision processes allows companies to detect and anticipate unmet customer needs and establish deeper relationships with their customers.

No one disputes anymore that today’s technologies allow making effective, efficient decisions and integrating information into everyday operations. Companies that can set up and maintain robust interaction between human mind and the artificial intelligence will go even beyond accessing new, higher-quality data. They will automate the path from information to action, to actually leverage the precious insights into their daily operations.