Overview of the fundamental attributes of AI and BI to give you a clearer sense of the goals of each, so you can move forward with greater clarity about how to progress on your overall Analytics (or Product or Operational) strategy.
Recently we've seen headlines talking about how Artificial Intelligence is making Business Intelligence obsolete. While AI is becoming more approachable than ever, it is still - and we posit will remain - a complement to BI, not a true replacement. Advances in AI and BI, specifically in the Healthcare industry, have been particularly impactful. As the Provider industry moves toward value-based care, having an efficient analytics platform in place is key. AI is able to complement BI by providing targeted insights to enable analysts and decision-makers to make better choices, faster.
Let's dive into the fundamental attributes of AI and BI to give you a clearer sense of the goals of each, so you can move forward with greater clarity about how to progress on your overall Analytics (or Product or Operational) strategy.
Questions vs. Answers
It's best to think of Business Intelligence (i.e. reports and dashboards using past metrics and KPI data) as a "question framework." Meaning, the focus of BI is to set up a flexible environment allowing users to ask various questions of their data and get an answer efficiently. For example, a regional manager for a hospital system may use a performance dashboard in Power BI that displays trends in key metrics like re-admissions, ER wait times, and outcomes for specific diagnoses. Undoubtedly, this dashboard will be parametrized and interactive so that the manager may ask more nuanced questions of the data to get an answer. While the solution is flexible, it is still incumbent upon the user to ask the question in order to get the answer he or she is driving toward.
Contrast this with Machine Learning (the most common application of Artificial Intelligence today) which can be thought of as an "answer framework." Advancements are being made in the areas of generalized AI, but for the most part, an organization is going to leverage the more narrow application of AI: Machine Learning. Because Machine Learning models produce an answer (i.e. a prediction) for a very specific question, they are best leveraged as a framework for directly providing answers. One single model cannot provide an answer to any arbitrary question. As it was once so eloquently stated,: "If you ask Deep Blue2 for a recipe for apple pie, it will fall on its face." To continue the example above, the Regional Manager may leverage Machine Learning to answer the question, "What is the capacity of my hospitals forecasted to be next quarter?" or, "What characteristics of our patients are driving re-admission rates?" They use a Machine Learning model that was trained to answer a specific question accurately
AI-augmented BI: Where the Synergies are
Integrating BI into AI is straight-forward: a data scientist must perform a substantial amount of descriptive analysis of the data he or she is attempting to model. The idea of leveraging BI tools and concepts in this process is given. The process of infusing AI into BI is where the real transformation is happening. To unlock the predictive power of historical data and enhance the BI user experience, BI platform-providers are making strides at integrating AI features into their BI tools. For the strides that are being made, however, the functionality isn't quite as seamless as their marketing departments might wish.
The primary AI-driven feature in BI tools currently is leveraging AI to make the interaction with the BI tool more natural and fluid. An AI tool called Natural Language Generation (NLG) is used to bridge the gap between a user’s spoken or written language and the technology that houses the data. Gartner estimates within the next 1-2 years conversational analytics and natural-language processing will boost BI analytics adoption to over 50% as new swaths of roles within a company are able to tap into data-driven insight1. NLG helps remove any technical requirements whatsoever from the user and simply requires the user to "speak" his or her question - amazing!
The utmost goal of any data organization must be getting the right data into the right hands at the right time so better business decisions can be made. As data democratization increases and users throughout the organization are self-service enabled, it is critical to success that the users interact with data in the way that is most efficient for them3. How might that look for the regional manager we considered previously? Imagine she is touring a facility and, while making her way around the ER department, walked up to a computer station and spoke, “How has COVID-19 affected ER wait times over the last three months?” With the ease of expressing her query in a natural way, she can get that insight readily from the analytics platform, thanks to Artificial Intelligence turbocharging Business Intelligence.
The confluence of AI and BI is driving efficiency more than ever and the result is increased data-driven decision making throughout the organization, regardless of the user's technical proficiency. This is especially applicable in the Healthcare Provider sector as data consumers with backgrounds focused on patient care instead of technology can take advantage of the benefits of analytics as easily as they can speak or type their queries.