Analytics is expanding beyond merely providing historical insights. For decades, analyses have been largely limited to displaying information about what happened in the past. These analyses offer some insight into why things happened but very little information about what should be done in the future. The viewer of the information was expected to know what to do with the analysis. Today we can do more with our analytics. Artificial intelligence (AI) and machine learning (ML) can be used to predict outcomes and behaviors to help guide business personnel on which actions to take to accomplish their objectives. When coupled with analysis of what has happened in the past, these AI- and ML-based analyses offer a more complete picture than is otherwise available.
AI and ML analyses employ algorithms that improve their results – in effect, that learn – as more and more data is processed. Neither technology is
new, but they have only recently gained wide adoption. This is because the algorithms used for AI and ML analyses can be extremely resource intensive, requiring large amounts of data as well as significant processing resources. Fortunately, the costs of these resources have declined, making it more economically feasible to use AI and ML more frequently.
Generally, AI and ML are used to perform two types of analyses: classifying people or objects into groups that share similar characteristics and predicting the behavior of those people or objects – for example, which groups of people are most likely to buy specific products or services. The output of AI and ML analyses, often referred to as “scores,” become additional data elements that
can be used just like any other data element. So as marketing teams are selecting individuals for a new campaign, they can use the score generated from AI and ML analysis to rank the likelihood that someone will respond as one of their criteria. And as human resource managers are reviewing lists of candidates to hire, they can include predictions of which candidates are most likely to accept the offer and which candidates are most likely to succeed in the organization.
Using AI and ML analysis can provide significant benefits for organizations. Our research shows that AI and ML capabilities help organizations gain a competitive advantage, improve customer experiences and increase sales. These techniques also make it possible to use natural language processing and to personalize analyses. Natural language processing can enable line-of-business personnel to type or speak their queries in every-day language rather than query languages such as SQL. This type of conversational approach to computing makes it easier for more individuals to access and query data. And AI and ML techniques can also be used to make recommendations – for example, of which sales analyses might be most relevant – relevant to the individual based on his or her department, role and past usage patterns of the system. Both of these techniques enhance the analytical experience for users.
However, organizations can face some challenges in applying these techniques. A primary challenge identified in our research is a lack of skilled resources. One of the ways organizations seek to reduce the demand for skilled resources is to acquire AI and ML embedded into their business intelligence tools. In fact, this is the most common way in which organizations would prefer to acquire or implement AI and ML.
Organizations should be exploring ways in which to enhance their analyses using AI and ML. The predictive nature of these techniques can help them move from observations of past behaviors to recommendations of what could or should be done. These analyses will require resources with specialized skills, and that may mean looking outside the organization. But once they’re secured, the analyses can be made more accessible to others by embedding AI and ML capabilities into business intelligence and analytical applications. Taking these steps will help the organization move from insights to actions, which will help improve performance and the bottom line.