Many of us remember the old days of enterprise business intelligence (BI) delivery where all requests for new or changed queries, reports, and dashboards had to go through a centralized IT team of BI and data professionals. Today, enterprise BI users enjoy some level of self-service from modern BI platforms with rich semantic layers, graphical user interfaces (GUI), and guided analytics.

With self-service, the percent of non-IT professionals able to fulfill their own BI requirements slowly but surely went up to about 20%. But that’s where it has stayed for the last decade. No matter what new “user friendly” and/or “intuitive” features BI vendors introduced, the 20% seems to be a current ceiling. In other words, out of all enterprise decision-makers who could and should be using analytical applications and platforms hands-on, only 20% do so today. These are your classical business analysts, data analysts, the new citizen data scientists, and citizen developers (aka power users). The other 80% still rely on the 20% for data sourcing, data discovery, data integration, building metrics and KPIs, running analytics, and delivering insights.

How do we address the needs of the other 80%? It will take three techniques that are all about bringing data to the business users vs. making business users go get data:

  • Natural language interactions with data. Using generative AI techniques (among others), one can now enable conversational ways to interact with data. Think of this as a BI chatbot. There are caveats. For example, you will have to double down on, or even increase your investments in, building up rich semantic layer. We predict this approach will empower another 10% of business users to get their own data/analytics questions answered.
  • ML-based alerting. Why will there be only a 10% adoption uplift from the previous bullet point? Asking a complex business question still requires a certain amount of data literacy, requires a user to know where (as in what app) to ask a data question, and assumes the user has the time and motivation to go look for answers. However, busy business executives should be able to simply “subscribe” to a data source and let the power of ML identify “areas of interest.” BI platforms will alert the subscriber that something is going on with the data and suggest that they should take a look. We predict this will address another 20% of users. So, we are up to a total of 50%.
  • Embedded or “ambient” BI/analytics. While the first two approaches partially or fully rely on the power of ML, this last approach is good ol’ embedded analytics — embedded in every system of work, such as enterprise resource planning (ERP), customer relationship management (CRM), and productivity and collaboration apps, where most of the last 50% of business users “live.” Ambient analytics means “it’s just there.” Keep in mind, however, that this will require professional software engineers to customize, embed, and integrate. So don’t expect you can eliminate the need for BI and software engineers.

Needless to say, there’s a cultural change necessary to educate and empower business users to become as self-sufficient in their data/analytics/BI requirements as possible. I invite you to examine Kim Herrington’s research on these important topics. Have more questions? Let’s chat.

Also, I’ll be presenting a related session at our upcoming Technology & Innovation Summit North America in Austin entitled “Democratize Enterprise Data With AI-Infused Business Intelligence” along with VP, Research Director Aaron Katz. Check out the event agenda here for details.