By using website you agree to our use of cookies as described in our cookie policy. Learn More

TDWI Upside - Where Data Means Business

AI for BI: Better or Faster Decisions?

AI was big news in 2018, but we overlooked the two distinct aspects of AI -- how it can augment human decision making or automate decision making. Here's why the distinction is important.

The past year has seen virtually every BI and analytics vendor claiming enhancement of their tools with various flavors of artificial intelligence (AI). Some claims are genuine. Others are little more than a thin veneer of AI terminology applied to planned or existing features. I have two questions regarding the emergence of AI for BI in 2018: have there been some real breakthroughs and what happens next?

For Further Reading:

Will AI Kill the Data Scientist?

Machine Learning that Automates Data Management Tasks and Processes

How to Sell AI to Your Enterprise

I wrote a series of three articles in March 2016 posing the question: "Where is Cognitive Decision Making in BI?" I had already been asking similar questions as early as 2011 but was bold enough to predict that cognitive would replace analytics within another year or two. In reality, AI blended into analytics rather than replaced it. The similarities in aim and approach were increasingly recognized -- both AI and analytics try to predict the future based on large quantities of data.

Predicting future outcomes is only one aspect of decision support, as I've argued recently. However, business needs support in two more fundamental areas: distilling more insight from huge volumes of data and moving to more near-real-time decision making. In terms of AI, the first area involves augmenting current human decision making, leading -- in theory -- to better-informed decisions. In the second area, automating decision making in all or in parts of a process accelerates decision making by reducing human involvement.

The business rationale differs substantially between the two approaches. Automation directly reduces labor costs; augmentation promises better business insights and decisions. Financial considerations often favor the former but miss the possible growth opportunities of the latter. Social considerations often favor augmentation over automation. Although the two approaches are not mutually exclusive, most implementations emphasize one over the other.

AI's Use in BI

Two BI product announcements in late 2018 addressed the augmentation and automation of decision-making support. I'll use them as examples to illustrate what AI for BI has most recently achieved and to form the basis for thinking about what's next in 2019.

IBM Cognos Analytics uses AI to augment the process of data discovery and exploration, as well as the creation of dashboards and reports. An "Augmented Intelligence Architecture" embeds machine learning and natural language processing tools to provide a conversational user interface with an extensive recommendation engine.

The system analyzes the contents of data sources and ontologies (where available) to guide users toward a more comprehensive and integrated view of the data available and the relationships within it, as well as offering smart alternatives for visualizing it. By engaging with data in new ways, business users can gain and share previously hidden patterns and insights in the data, pursue more open-minded lines of investigation, and ultimately make better decisions.

Yellowfin Signals, a new addition to Yellowfin's product suite, uses AI in a very different way by focusing on automation of the overall decision-making process. Here, the AI function continuously monitors data streams for exceptional changes, correlates changes across data streams, and generates personalized alerts and suggestions to decision makers about possible decisions or actions.

While providing typical dashboarding and analytical functionality to users, the emphasis is on the overall decision-making process. Yellowfin's philosophy is to drive transparency in information availability across the organization and allow business people to focus only on data that signifies important aspects of the business, thereby accelerating decision making at all levels.

These two examples illustrate genuine and very different applications of AI in the BI and analytics world today. Under the covers, however, both use comparable underlying AI functionality to improve some of the fundamental processes in decision making. IBM Cognos Analytics addresses the process of data discovery and reporting; Yellowfin Signal focuses on process monitoring and reacting to change in the business.

Forward-Looking Statements

Augmenting users' skills for both business and more technical people is the most obvious and most easily addressed area for applying current AI functionality to BI and analytics. In addition, these improvements apply to the core function of BI tools and are thus more easily implemented by vendors. The messaging is also familiar: improved personal productivity leads to overall decision-making success. Further expansion and improvement of BI tools based on this approach are likely to become table stakes in this market for vendors in 2019.

Automation of decision-making (as opposed to data discovery and reporting) processes is potentially a game-changing approach. A focus on exception handling rather than continuous human monitoring for business processes offers potentially significant productivity improvements at both personal and organizational levels. However, the challenge is that this approach to AI for BI demands a greater level of organizational change to achieve its full benefit. If this organizational challenge can be shown to be addressed, this use of AI in BI will be a significant development this year.

Marketplace Hype

There is an additional consideration to the outlook for AI in BI this year. Broadly speaking, AI appears to be at or near maximum hype in the marketplace. It is increasingly hard to find any IT tool or household gadget that isn't being "enhanced" with AI. We are thus primed for a crash in confidence in AI if some major calamity emerges that can be blamed on AI. If that happens, the application of AI to BI may well continue, but we'll hear a lot less about it.

About the Author

Dr. Barry Devlin is among the foremost authorities on business insight and one of the founders of data warehousing in 1988. With over 40 years of IT experience, including 20 years with IBM as a Distinguished Engineer, he is a widely respected analyst, consultant, lecturer, and author of “Data Warehouse -- from Architecture to Implementation" and "Business unIntelligence--Insight and Innovation beyond Analytics and Big Data" as well as numerous white papers. As founder and principal of 9sight Consulting, Devlin develops new architectural models and provides international, strategic thought leadership from Cornwall. His latest book, "Cloud Data Warehousing, Volume I: Architecting Data Warehouse, Lakehouse, Mesh, and Fabric," is now available.

TDWI Membership

Accelerate Your Projects,
and Your Career

TDWI Members have access to exclusive research reports, publications, communities and training.

Individual, Student, and Team memberships available.