AI for BI at the heart of third-generation analytics
Doug Henschen, analyst at Constellation Research, details in a Q&A some of the findings of his recent report on augmented capabilities in BI platforms.
AI for BI is a key tenet of the third generation of analytics.
Sometime in the middle of the current decade, features such as augmented intelligence, machine learning and natural language processing started to become key parts of business intelligence platforms.
In the years since, although analytics platforms have progressed, AI for BI still hasn't matured to the point where analytics tools can truly free up humans from the mundane tasks associated with data analysis, to the point where data analysis is part of everyday applications rather than a stand-alone application unto itself, or to the point at which BI platforms can predict for humans a likely outcome before they even request it.
And it hasn't gotten to the point where it's accessible to everyone.
In September, Constellation Research released a report entitled "Augmented Analytics: How Smart Features Are Changing Business Intelligence."
Authored by analyst Doug Henschen, the report took a deep look at the third generation of business intelligence, which Henschen approximates began in 2015. The report homes in on the data preparation, data discovery and analysis, natural language interfaces and interaction, and forecasting and prediction capabilities of the BI platforms offered by leading vendors.
Henschen discussed some of his findings about AI for BI for a two-part Q&A. In Part I, he addressed what marked the beginning of this new era and who stands to benefit most from augmented BI capabilities. In Part II, he looked at which vendors are positioned to succeed, and where the third generation of BI is headed next.
Which vendors are in the best position to succeed in this new generation of AI for BI, and why?
Doug Henschen: I think there have been companies that have been more aggressive about augmented analytics capabilities that have led the way, been first movers. IBM came out with Watson Analytics in 2014, SAP acquired KXEN in 2013, ThoughtSpot and BeyondCore got started in the middle of the decade. They were first movers, and then after some of the new capabilities emerged the fast followers in the market responded. Tableau was early with data visualization recommendations. But I don't think any one company has augmented analytics locked up.
If not excelling across the board, who is doing well in certain areas of third-generation business intelligence?
Henschen: Across the four areas I've been seeing leaders -- ThoughtSpot on search and natural language query, Oracle has stepped up a lot on natural language query and has a really good mobile app for natural language query support. Salesforce has done a lot with Einstein Analytics in trending and forecasting and prediction to focus on outcomes ... three steps removed from the actual business action. In BI and analytics there's this desire now, particularly as you move toward the business community, to say, 'Don't show me a dashboard, don't show me a report that I have to interpret; tell me what to do.' In that area Salesforce has been very aggressive with Einstein and Einstein Analytics. My next report is going to be on embedded analytics, which is about bringing analytics into applications, and I think that's where we're really going to see this idea of democratization realized.
Which vendors are in a precarious position, not adapting to this era of AI for BI quickly enough?
Henschen: Generally the market is responding. For vendors that have been more focused on reporting, augmented intelligence is not as much of a factor, so they've been less motivated to move into that area. That point of 2015 is when augmented analytics became a criteria that got added to the list, but it was by no means the only criteria, or the key criteria. I think we're going to gradually see, as this stuff gets more and more powerful, it will become more and more important, and as it matures we will also see more capabilities. I graded vendors on four categories, but when some reports first started looking at augmented capabilities there was one grade. I'm sure in the future there will be more aspects of augmented capabilities to look at. It's just the cycle of maturation. As capabilities become proven it will become more and more commonplace for every vendor to have to have that.
Beyond the four capabilities you discussed in your report, what's something we can expect to see soon in the evolution of AI for BI?
Doug HenschenAnalyst, Constellation Research
Henschen: BI has always been about decision support, so you have these broad horizontal platforms and tools that let you analyze anything, but where the rubber meets the road is where these insights actually help customers make a decision. That's my next report on embedded analytics -- ways we're going to be doing a better and better job of embedding analytics right into the context of decisions and transactions and applications. There are some technologies that are helping to make that happen. Microservices will let us make delivery insight more granular, so instead of a full page report or a full page dashboard we can have a micro-chart or just a KPI or imbed just that nugget of information -- that nugget of insight that helps us make the decision -- into the app.
There's also this trend of no-code and low-code application development, and that will enable analysts and even business users to start developing these analytical applications or develop applications with analytics embedded within them. Those are good trends, but this is just the start of these things. The vision is always well ahead of the reality of what's actually happening in the market.
What's something we're seeing the beginnings of now that you expect to be improved?
Henschen: We're seeing a lot of natural language query, for example, but I think that's still really immature. One of the takeaways from my report is that we're in the opening innings, and this stuff needs to get a lot better. The customers I talked to are getting a lot more out of natural language generation than they are out of natural language query, and it's because query is very context-specific. As we see more sophisticated query capabilities, and we see the blending of query capabilities with understanding of intent -- looking at the patterns of what people are asking, what groups of people are asking for, and learning of that -- that query capability is going to become more powerful than it is today.
Editor's note: This interview has been edited for clarity and conciseness.