Pascal Bornet is an AI and automation expert, best-selling author, keynote speaker, and CDO at Aera Technology.
Businesses make decisions every minute. Typically, these decisions range from designing strategy and market selection to recruiting talent or authorizing bill payments. The success or failure of a business depends on the effectiveness of the decisions made by its employees.
However, people often have to make decisions under time pressure and with limited information or with too much information, causing information overload. Even the most important decisions are often made without thinking or influenced by prejudices, emotions and subjectivity. Most decisions are made by isolated individuals based on limited experience and siled data, depriving companies of the benefits of collective wisdom and shared information. According to a survey by Signal, the top barriers to decision-making are an overwhelming amount of data (44%), lack of time (21%), stakeholder alignment (16%), and implementation of the solution (19%).
Finally, there is a lack of systematic evaluation and feedback on the success of decisions, which prevents people from learning and improving the way they make decisions. All of this results in poor decision-making processes.
These poorly made decisions have a significant financial impact. A McKinsey survey found that ineffective decision-making costs the average S&P 500 company $250 million a year, and harvard business review noted that a company’s financial performance is 95% correlated with the effectiveness of its decisions.
The growing pace
Increasingly, human decision-making is unable to meet the demands of a modern, technology-driven environment. According to Gartner, Inc., 65% of executives believe the decisions they make are more complex than two years ago.
Current technologies and approaches to try to support human decision-making are not enough to close the gap. On the one hand, machine learning technologies focus on producing accurate insights through predictions or optimizations. However, they lack the link to the specific expected results and the execution of the decisions they help to inform. On the other hand, process automation applications (e.g., robotic process automation) focus on process tasks, partly improving efficiency but not impacting the quality of decisions.
Signal found that 85% of executives believe that $4.26 trillion in additional revenue could be brought to the US economy (nearly twice the UK’s GDP) by applying technology to support decision-making. .
The need for decision intelligence
Decision Intelligence (DI) is a new field that uses technology to support, augment and automate business decisions. DI was popularized in Lorien Pratt’s 2019 book, Link, and identified by Gartner as one of the most impactful technology trends for 2022. DI technology combines the data-driven insight and power of artificial intelligence and machine learning with the task-driven capabilities of process automation tools, connecting them in a decision-oriented process. streamlined pipeline. DI connects data, decisions, actions and results. It also incorporates the ability to monitor and improve one’s decision-making abilities based on human feedback and comparing actual decision outcomes with predicted ones.
DI can help with decision-making based on three approaches.
• Help with the decision. Machines provide tools such as data analysis to support human decision making. Generally, this is what we call business intelligence technologies. The decision is made by people with the support of the machine.
• Increase in decision. The machines generate recommendations for decisions, including an expected business outcome such as: “Buy X units from supplier Y, then you’ll save Z million dollars.” The machine proposes the decision, but people take it. The user accepts, rejects or modifies the decision recommendation.
• Automation of decisions. Machines make autonomous decisions and implement them without human intervention. People are focused on monitoring the system and reviewing the results.
The right approach will depend on the complexity and frequency of decisions. The simpler and more repetitive a decision is, the easier it can be automated and the more value can be unlocked by freeing humans from it.
As they gain more confidence in the technology, users can choose to evolve their decision-making approach from support to augmentation and, finally, to automation. They may also choose to switch from automation to manual in the face of unpredictable events such as the Covid-19 pandemic.
How to Succeed with Business Intelligence
Start by defining your business goals and the key decisions that are being made or could be made in service of those goals. Calculate the financial impact of each decision so you can prioritize them and define how you will measure their success.
Then, evaluate the decisions along two axes: complexity and frequency. This will allow you to categorize them into groups suitable for decision automation (the simplest and most frequent decisions), decision support (the most complex and less frequent decisions) and augmentation. decisions (the intermediate ones).
Build your implementation roadmap, starting with the most impactful decisions, and use a phased approach to roll out the rest of the implementation.
Put the right governance in place: a center of excellence to own the DI deployment and an automation operations center to maintain it. Ask analysts to specialize by type of decision rather than by business function. Practice continuous learning and improvement by comparing decision outcomes to planned outcomes.
Business intelligence is a new lever for business success in our modern era, and the ways to implement it are clear. Has your company already started its journey towards business intelligence?
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs, and technology executives. Am I eligible?