which was vital to both understanding the disease and testing the vaccine.
Further, scientists, healthcare workers, and public health organizations around the world shared data and collaborated on finding insights and answers. The global response to the pandemic was a stellar display of how effective humankind can be in tamping down any threat when countries, health entities, and experts cooperate. The effort should be celebrated and commemorated for time eternal.
But all decision-makers can also learn from the shortcomings as well as the successes in this huge undertaking to end a dangerous pandemic. Chief among the shortcomings is that there is still uncertainty, after many months, about the specific actions that should be taken despite massive global data sets and ongoing analysis.
Businesses and other organizations find themselves in a similar predicament even in the absence of urgency, alarm, and dire consequences. In other words, even with the luxury of time and calmer heads, you can glean insights from data and still not know what to do about or with them. To put this in proper context, you should always remember this:
Even so, you can and should improve how you make decisions and judge them by their real world impacts. That requires the combined applications of several disciplines and more human input — a more than fitting definition of decision intelligence.
Going where humans fear to tread on data
Though the processes used under the big umbrella known as decision intelligence vary from one entity to the next, they’re likely to be more warmly embraced by people who were previously concerned that data analytics, and particularly those associated with AI, would eliminate their jobs.
AI, more often than traditional automation, is perceived by some as a direct competitor by managers and executives by virtue of science fiction depictions where AI is smarter than humans and capable of doing even high level jobs. That’s also partly because of the frequent and often wrong assumption that automation is limited to replacing jobs on the lower rungs of the career ladder. By comparison, AI cuts directly from the top. That point was first driven home when Deep Knowledge Ventures, a Hong Kong-based venture capitalist fund, added an algorithm named VITAL as a member of the board of directors in 2017. After that, it appeared that no job was safe from a machine takeover.
OK, some did note that appointing an algorithm to the board was likely a publicity stunt, since most board of directors use data to inform their votes, but the scare that AI may replace business leaders nevertheless lingers.
Deep Knowledge Ventures credits its algorithm with saving the company from bankruptcy caused by “overinvesting in overhyped projects.” Known as Vital (short for Validating Investment Tool for Advancing Life Science), the algorithm established itself as a seemingly crucial member of the board. It’s interesting that the rest of the board — or perhaps it was the stockholders? — apparently had little regard for the directors’ ability to keep the company financially strong.
Executives, whether at the head of business lines or at the top of the company pinnacle, typically fear data fueled algorithms. On the one hand, they’re expected to toe the data-driven company line. On the other hand, data-driven decisions may make their own talents obsolete.
In doesn’t help that executive pay, benefits, and perks are large line items in the biggest of company expenses: payroll costs. You can easily see where the same cost cutting logic that executives use every day could eliminate them as well.
Decision intelligence rebalances the scale by adding more weight to human roles in making key business decisions. That alone makes the concept welcome to leadership. However, decision intelligence is not a license nor the means to return to gut instinct, seat-of-the-pants, ego-driven, or agenda-loaded decision manipulations. The value in decision intelligence is that it is a far more effective way to make business decisions and savvy leaders will instantly grasp its importance to their organizations and careers.
In short, it is a rebalancing of how data is used and viewed. The evolution is in step with maturation patterns in other disciplines and a payback of sorts for data science’s contributions to those developments. One example speaks for many: computing and data science spurred the emergence of Digital Humanities as a new field in the 1950s and has enabled its steady improvement ever since. Now a similar development process is flowing in the other direction.
Decision intelligence is a recipe wherein data, automation, AI and human decision-making capabilities are blended to bake better outcomes into the processes. Further, it is a renewed focus beyond mechanical and digital efficiencies to make the outcomes more meaningful in human applications and impacts.
For many experts and observers, including many executives who have always highly valued business acumen in themselves and other people, decision intelligence’s acknowledgment and inclusion of the same is a natural progression in business applications.
It is a result of the formal recognition of another truth too: no matter how far AI/ML has advanced, combining it with in-house business knowledge always makes for better business outcomes.
Ushering in The Great Revival: Institutional knowledge and human expertise
Two of the biggest casualties in traditional data mining are institutional knowledge and human expertise. Institutional knowledge is defined as the knowledge within an organization about its own business and customers that’s passed on from older workers and leaders to newer ones in informal and usually verbal person-to-person exchanges. Because much of it is stored in the minds of workers and executives, it’s supremely difficult to identify, retrieve, and digitalize for inclusion in a data set. Therefore, it’s often lost when a person with some of this knowledge retires, dies, changes jobs, or otherwise stops being an active part of the company. Without this key information, business decisions can be made in the wrong context for the situation and result in failure or undesirable consequences.
Human expertise works similarly: It’s the knowledge gained by an individual by way of education, intuitive intelligence, talent, accumulated skill sets, experience, exposure, incidents of failure and success, encounters with anomalies and repetitious events, and a myriad of unique circumstances over the span of a career or lifetime. This information, too, is difficult to digitalize and add to a database. Therefore, human expertise also tends to be lost to illness, retirement, job change, or death.
The cost to any organization of the loss of either institutional knowledge or human expertise can be enormous in terms of money value, company culture, and the shape of the organization’s competitive edge. These facts are not lost to many in business leadership and data science, which is spurring a revival in both valuing and capturing these deep wells of specialized and irreplaceable data.
Some think of it as a great revival as the pendulum swings from one extreme to the other. For example, disinterest in customers from a focus on profits alone has now swung to a near-fanatical interest in personalizing every customer encounter and ensuring a great customer experience for each individual. This swing comes from a renewed appreciation for the value of human expertise (in this case in sales and marketing) and in institutional knowledge of customers and operations with regard to improving profits. In other words, once data-driven process efficiencies had mostly or completely been realized, companies learned that profits cannot be separated from customers, as the latter begats the former. Hence the resumed interest in reselling to existing customers and retaining customer knowledge beyond basic financial transaction details.
After