This content is part of the Essential Guide: Augmented analytics tools: Business uses, benefits and barriers

AI at the core of next-generation BI

AI is part of BI platforms, but this is only the start of a new analytics generation. In a Q&A, Constellation Research analyst Doug Henschen discusses AI tools in BI platforms.

Next-generation BI is upon us, and has been for a few years now.

The first generation of business intelligence, beginning in the 1980s and extending through the turn of the 21st century, relied entirely on information technology experts. It was about business reporting, and was inaccessible to all but a very few with specialized skills.

The second introduced self-service analytics, and lasted until just a few years ago. The technology was accessible to data analysts, and defined by data visualization, data preparation and data discovery.

Next-generation BI -- the third generation -- is characterized by augmented intelligence, machine learning and natural language processing. It's open everyday business users, and trust and transparency are important aspects. It's also changing the direction data looks, becoming more predictive.

In September, Constellation Research released "Augmented Analytics: How Smart Features Are Changing Business Intelligence." The report, authored by analyst Doug Henschen, took a deep look at next-generation BI.

Henschen reflected on some of his findings about the third generation of business analytics for a two-part Q&A.

In Part I, Henschen 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 next-generation business intelligence is headed next.

In your report you peg 2015 as the beginning of next generation BI -- what features were you seeing in analytics platforms at that time that signaled a new era?

Doug HenschenDoug Henschen

Doug Henschen: There was a lot percolating at the time, but I don't think that it's about a specific technology coming out in 2015. That's an approximation of when augmented analytics really became something that was ensconced as a buying criteria. That's I think approximately when we shifted -- the previous decade was really when self-service became really important and the majority of deployments were driving toward it, and I pegged 2015 as the approximate time at which augmented started getting on everyone's radar.

Beyond the technology itself, what were some things that happened in the market around the time of 2015 that showed things were changing?

Henschen: There were lots of technology things that led up to that -- Watson playing Jeopardy was in 2011, SAP acquired KXEN in 2013, IBM introduced Watson Analytics in 2014. Some startups like ThoughtSpot and BeyondCore came in during the middle of the decade, Salesforce introduced Einstein in 2016 and ended up acquiring BeyondCore in 2016. A lot of stuff was percolating in the decade, and 2015 is about when it became about, 'OK, we want augmented analytics on our list. We want to see these features coming up on roadmaps.'

What are you seeing now that has advanced next-generation BI beyond what was available in 2015?

Anything that is proactive, that provides recommendations, that helps automate work that was tedious, that surfaces insights that humans would have a tough time recognizing but that machines can recognize -- that's helpful to everybody.
Doug HenschenAnalyst, Constellation Research

Henschen: In the report I dive into four areas -- data preparation, data discovery and analysis, natural language interfaces and interaction, and forecasting and prediction -- and in every category you've seen certain capabilities become commonplace, while other capabilities have been emerging and are on the bleeding edge. In data prep, everyone can pretty much do auto data profiling, but recommended or suggested data sources and joins are a little bit less common. Guided approaches that walk you through how to cleanse this, how to format this, where and how to join -- that's a little bit more advanced and not everybody does it.

Similarly, in the other categories, recommended data visualization is pretty common in discover and analysis, but intent-driven recommendations that track what individuals are doing and make recommendations based on patterns among people are more on the bleeding edge. It applies in every category. There's stuff that is now widely done by most products, and stuff that is more bleeding edge where some companies are innovating and leading.

Who benefits from next-generation BI that didn't benefit in previous generations -- what types of users?

Henschen: I think these features will benefit all. Anything that is proactive, that provides recommendations, that helps automate work that was tedious, that surfaces insights that humans would have a tough time recognizing but that machines can recognize -- that's helpful to everybody. It has long been an ambition in BI and analytics to spread this capability to the many, to the business users, as well as the analysts who have long served the business users, and this extends the trend of self-service to more users, but it also saves time and supports even the more sophisticated users.

Obviously, larger companies have teams of data analysts and data engineers and have more people of that sort -- they have data scientists. Midsize companies don't have as many of those assets, so I think [augmented capabilities] stand to be more beneficial to midsize companies. Things like recommended visualizations and starting points for data exploration, those are very helpful when you don't have an expert on hand and a team at your disposal to develop a dashboard to address a problem or look at the impact of something on sales. I think [augmented capabilities] are going to benefit all, but midsize companies and those with fewer people and resources stand to benefit more.  

You referred to medium-sized businesses, but what about small businesses?

Henschen: In the BI and analytics world there are products that are geared to reporting and helping companies at scale. The desktop products are more popular with small companies -- Tableau, Microsoft Power BI, Tibco Spotfire are some that have desktop options, and small companies are turning also to SaaS options. We focus on enterprise analytics -- midsize companies and up -- and I think enterprise software vendors are focused that way, but there are definitely cloud services, SaaS vendors and desktop options. Salesforce has some good small business options. Augmented capabilities are coming into those tools as well.

Editor's note: This interview has been edited for clarity and conciseness.

Dig Deeper on Data science and analytics

Data Management
SearchAWS
Content Management
SearchOracle
SearchSAP
Close