Doing more with less is a key business strategy driver across many resource-intensive industries. Enterprises are looking to get more out of their artificial intelligence (AI) and machine learning (ML) infrastructure investments than mere insights provide.
Instead, they need practical, data-based recommendations to simplify complex decisions around mission-critical areas, such as resource allocation, task management, supply chain, and profitability.
That is the role decision intelligence (DI) is designed to play.
Our Ultimate Decision Intelligence Guide defines what decision intelligence is, describes the benefits of its use, how it works, and why decision intelligence provides a superior solution for using big data to meet business needs and improve business practices.
Decision intelligence is the enterprise’s ability to make smarter decisions by analyzing large amounts of data to determine how actions lead to outcomes. It takes the intent of business intelligence—to facilitate more informed business decisions—and makes it accessible throughout the enterprise.
By analyzing cause and effect, decision intelligence transforms the massive amount of data businesses collect today into comprehensive decision support platforms to provide answers when and where stakeholders need them most.
While many companies still make business decisions based on anecdotal evidence alone—a 2020 Gartner survey shows that only half of the decisions businesses make today are influenced by data—entire reservoirs of actionable insight reside within data warehouses that cannot be uncovered apart from deep AI and ML analysis.
These three industry use cases illustrate how decision intelligence works:
Financial services providers can use decision intelligence to process credit applications for mortgages and car loans. Driven by artificial intelligence, DI can use a customer’s credit score, income, and other information to determine eligibility for certain services.
By integrating decision intelligence, retailers can easily decide how much inventory to purchase to meet optimal stock levels. They can also make decisions about when to place orders to ensure just-in-time warehouse management. For instance, a home improvement retailer could make inventory purchasing decisions based on predicting a particularly hot summer.
Employing decision intelligence can positively impact the COVID-19 crisis, according to Dr. Loren Pratt, chief scientist at Quantellia, an AI and ML technology company, and author of “Link: How Decision Intelligence Connects Data, Actions, and Outcomes for a Better World”.
“Although decision intelligence and [intelligent] orchestration are both young fields (just over ten years old), they are the right fit for the COVID-19 challenge,” she said. “What better time to address them to move to the next level of maturity in how we make evidence-based decisions, from data-driven decisions to decision-driven data?”
Cassie Kozyrkov, chief decision scientist at Google and noted decision sciences expert, believes decision intelligence is crucial to the AI era:
Exploring, analyzing, and fully understanding these data-driven facts helps organizational stakeholders make more informed decisions, enact better policies, and implement more effective practices that serve to move the needle toward greater business success effectively.
The disciplines that make up the qualitative side of data (as opposed to quantitative) have traditionally been referred to as the decision sciences. Decision science and decision intelligence share a common goal: to help organizations make better decisions using data.
Think of “decision science” as the overarching term while “decision intelligence” is its underlying operational component. Decision sciences working alongside decision intelligence-driven support systems ensure unbiased decisions are made effectively.
Strategic intelligence is a type of business intelligence that deals with implementing data-based insights to drive and inform strategy. It can also assist businesses in understanding current industry trends and make sense of consumer behavior patterns to help enterprises decide on and plan a future course of action.
As good as that may sound, according to Kozyrkov, strategies based on pure mathematical rationality without a qualitative understanding of decision-making and human behavior can be “pretty naive and tend to underperform relative to those based on joint mastery of the quantitative and qualitative sides.”
Not all outputs or suggestions are decisions, Kozyrkov says. In decision analysis terminology, a decision is only made once an irrevocable allocation of resources takes place. If you can change your mind for free, no decision has yet been made.
Decision intelligence offers several benefits for organizations willing to implement its use. However, four predominate:
IDC, a global market intelligence firm, predicts there will be 175 zettabytes of data worldwide by 2025, much of it unstructured. That’s 175 billion terabytes—imagine if all 7 billion people on Earth had a 1-terabyte hard drive in their computer, which is common. 175 zettabytes of data would fill the computers of 25 Earths.
Humans will be unable to process this massive volume of data manually. Instead, organizations will need to use decision intelligence combined with advanced machine learning algorithms to augment and improve human decision-making capabilities in these increasingly complex systems.
A second benefit is that decision intelligence helps reduce decision latency, the time it takes to make a decision.
The larger the organization, the more complex the decision-making process is. Not only do slow decisions impede progress, but profitability as well. Machines process information quickly, helping reduce the risk of unforeseen outcomes without slowing the process itself.
“It can take days—and in some cases weeks—for even the most sophisticated analysts to understand what data is most relevant to a given business question, generate the hundreds of queries required to sufficiently answer the question, and to present the answer in an interpretable, actionable format to the business,” says Peter Bailis, Sisu founder and CEO. “Today, relative to our data architectures, it’s the people who are overwhelmingly expensive and slow.”
McKinsey senior partner Kate Smaje wrote that organizations are now accomplishing in 10 days what used to take 10 months. With data powering better and faster decisions, affording more informed decision making, business leaders have time to make thoughtful assessments about how to respond.
Machines can also help reduce biases humans run into when considering the types of decisions that need to be made. For example, a person’s preference for doing things a certain way could lead an enterprise to favor one approach over another despite growing evidence to the contrary. Decision intelligence can help overcome these pitfalls.
It is left to humans, not machines, to make most critical business decisions. However, unlike machines, humans are not optimization engines.
“We’re satisficers,” Kozyrkov says, “which is a fancy word for corner cutters who are satisfied with good enough as opposed to perfect.”
By relying on intuition, humans arrive at decisions without conscious reasoning. Research by the University of Minnesota found that 89 percent of managers admitted to sometimes using intuition to make decisions, and 59 percent said they do so often.
Decision intelligence can adapt to the advantages of intuition and other human judgments while at the same time eliminating errors that result from such behaviors.
This section provides an overview of the different types of decision models and how they relate to decision intelligence.
Four types of decision models exist:
Intuitive decision-making is much less structured than its rational counterparts and opts for subjective opinions, although not “gut” feelings. It takes into consideration such factors as:
Industrial barons, such as Henry Ford, founder of Ford Motor Company, and Bill Allen, the CEO of Boeing in the 1950s, often relied on their intuition when making crucial business decisions. For example, Allen bet $16 million to achieve civilian air travel as we know it today.
Researchers have found that an intuitive decision-making model yields good results when dealing with areas where someone has a lot of expertise or experience. However, going with your gut is less effective and efficient when faced with an unfamiliar circumstance because you don’t have enough experience to recognize patterns quickly.
The rational model is the most used decision-making method. Typically, it consists of the following:
Larger innovation companies in Sweden, such as Volvo and Ericsson, adhere to the rational model, and use structured processes to manage their operations, often collaborating with a large number of people, all with differing areas of expertise.
Bounded rationality is a human decision-making process in which we attempt to “satisfice” rather than optimize, to quote Kozyrkov. In other words, we seek a decision that is “good enough” instead of the best possible action.
This model is based on three main limitations that result in suboptimal decision making: cognitive limitations, imperfect information, and time constraint.
Cognitive limitation refers to our inability to consider all available factors in our decision-making. Second, information imperfection refers to the lack of information a consumer has. The third, time constraints, often occur when making a purchase.
Let’s say you stop at a store during a lunch break. You have places to be and decisions to make. You cannot spend half an hour in the store deciding on the most optimal option to buy. Constrained by time, you grab the first item that appeals to you.
The creative decision model applies when the decision-maker arrives at an original and unique decision to solve a problem. After gathering information and insights about the problem, the decision-maker undergoes a period of incubation. He does not actively think about the solutions but lets his unconscious mind take over. After the period, the answer surfaces in a “eureka” moment.
The downside of this model is that the success depends mainly on the decision maker’s personal traits, such as his creativity and the contextual situation.
As you can see, each model comes with its inherent limitations. For example, intuitive decision-making is subjective, relying heavily on human experience, judgment, and perception. Much the same holds with the creative model. The bounded model is constrained by time while the rational model can be constrained by insufficient information, which creates a problem when there is a need to evaluate alternative courses of action.
While these models impact decision-making, their inability to capture all the uncertainty factors can lead to unpredictable outcomes. By introducing AI and ML algorithms into the decision-making process, decision intelligence can consider all aspects of selecting between options to turn information into better actions at scale.
Decision intelligence works through five sequential steps:
The model begins with collecting all relevant information from a variety of sources: historical and transactional, behavioral and attitudinal, structured and unstructured, transient and persistent, and external or internal. Whatever information you collect will assist in recreating the outcome, informing other stakeholders and departments, or improving processes.
Once collected, analyze the data to gain a clear understanding of the situation and investigate possible actions.
Consider existing business capabilities to produce alternative actions. Because organizations cannot always see the complete picture, these actions must explain causalities that lead to other scenarios.
The model might run out of options due to the high complexity of the situation and time constraints. Data analysts should provide decision-makers with a range of actions they can perform quickly.
Ultimately, the decision model will choose and implement a particular action. This step also includes measuring the action’s impact to improve the model.
However, current business intelligence tools and platforms lack the power to provide the answers business users need to drive meaningful outcomes. The result is that business analysts default to intuition and reliance on “gut” feel rather than trusting what the data reveals.
That is where decision intelligence fills the gap.
Even though it is still an emerging field, by 2023, more than a third of large organizations will have analysts practicing decision intelligence, according to Gartner. This shift is the next step for organizations looking to establish a unified decision strategy across people and departments. The return on investment will depend on the particular use case, implementation, and how transformative DI becomes for the enterprise.
By improving the decision-making process, organizations can derive more value from their data and machine learning model investments and stay ahead of their competitors by arriving at decisions faster. As enterprises gain access to more data, this technology will only grow in value. In a few more years, we may wonder how we ever made decisions without it.
Sisu is the industry’s first decision intelligence engine using augmented analytics.
With its ability to analyze millions of possibilities in your data, surface key facts behind your customers’ behavior, and remove the analytics bottleneck so you can spend more time informing strategic decisions, Sisu accelerates your analysis faster than any other tool on the market.
With Sisu, you can explore cloud-scale data rapidly, go deeper than your dashboard and study customer behavior from every angle, and define and manage behaviors centrally to drive confident business decisions in real-time.
With more and more companies (including your competitors) relying on data for critical decision-making, isn’t it time your company does too?
Schedule a free demo today. One of our data experts will show you Sisu in action and how quickly you can find the facts buried in your data.