How do you deliver more insights out to more people? Operationalizing BI and analytics – that is, putting the power of data in the hands of everyone across the enterprise, not just analysts and data scientists – has always been the mantra for Birst co-founder Brad Peters.
According to research from Eckerson Group, when an organization deploys a BI and analytics system, roughly 10% of employees have the skills needed to produce insights from corporate data and deliver them to decision makers. That means much of the organization depends on a few experts to leverage data to make decisions. Do you see the bottleneck?
But there are more bottlenecks. To answer even more complex business questions, or to predict valuable business opportunities, data scientists are needed to apply advanced algorithms that require highly specialized expertise. And data scientists may represent less than 1% of users in an organization.
What if we put the power of data science in the hands of every business user directly, removing these bottlenecks? With Birst Smart Analytics announced today, we do just that. A business user simply selects a KPI of interest, and machine learning algorithms run automatically across all data points that are related to generate the key reasons “why” a KPI is trending upward or downward. Key drivers affecting the KPI are automatically presented in an easy-to-understand, outcome-focused dashboard that is personalized, so the user can easily interpret results and take the best course of action.
I spoke to Brad Peters, Birst co-founder and Senior Vice President and General Manager of Analytics and BI at Infor, to provide us with Birst’s unique vision and history behind using machine learning to create a more intelligent and automated BI system.
1. When you founded Birst, you had a vision for what BI could be. With AI as the next wave of disruption in BI, how has this changed, or re-enforced your vision?
When creating Birst in 2004, the idea was to get insight out to as many people in an organization as you can, so that even non-analyst people are empowered to use data to make decisions. Our focus was correct, and we began a path of building machine learning automation into the product. We started with Automated Data Refinement (ADR), which automates the creation of analytic ready data or smart data preparation. In parallel, we were also focused in creating a robust and scalable BI and analytics service in the cloud to meet the wide range of enterprise reporting requirements.
Today, more than 14 years later, we feel we have built a solid analytics-as-a service product with many “smart” features, including ADR. Our next “smart” release addresses advanced data discovery, which aligns with what we initially set out to do: put analytics into the hands of a broader set of users.
We are particularly excited about the release of Birst Smart Analytics because this was our original vision behind the company 14 years ago when we created a patented machine learning solution to auto analyze on behalf of an individual that owns a piece of the business so that they could get very personalized, high-value insight at scale. Unfortunately, the market wasn’t ready for it at that time or wasn’t ready to trust a machine to generate customized insights for decision making.
2. Birst was at the forefront of leveraging advanced automation and machine learning. Birst even has patents for the Smart preparation and visualization of data. How does this next release of Smart Analytics in Birst leverage or extend those patents?
When we first started Birst, we came up with the idea of having the system auto-generate the right set of unique content for each individual user. We patented this initial idea, so with Birst Smart Analytics, we are leveraging this same layer of machine learning to create personalized insights at scale. Instead of an analyst creating a single report for the masses, each user has a virtual private analyst that recommends insights pertinent to that user’s slice of the business.
Fourteen years ago, it was difficult for clients to understand or were not ready to accept the analysis that was generated for them by a computer. Trusting machines back then was crazy. Fortunately, Amazon, Apple, and other vendors creating machine-generated recommendations have paved the way for people to now be more open to trusting machines. But for business decision making, it is critical to have transparency in explaining how recommendations are made. We feel the market is ready to work with machines with this type of transparency instead of a black box approach.
3. BI is key for analyzing data and presenting actionable information to help business user make more informed, better business decisions. For such an important asset, why do think BI adoption is still low, and how does Birst’s Smart Analytics change that trend?
Traditional BI takes a one-size-fits-all approach, where interactive dashboards and reports are mass produced, leaving it up to the end user to customize or visualize the information based on his/her slice of the business. When the user does not have the skill set or time to get the relevant insights extracted from a mass-produced report, the BI system is abandoned and fact-based decisions cannot be made.
With Machine Learning, analysis is customized and presented to each user automatically. This mass customization approach enables a larger of population of users to adopt the BI system because the information presented is more interesting, relevant and useful for decision making.
4. How is Birst’s approach to AI unique compared to other BI vendors?
The automated insights Birst generates are more personalized and actionable.
First, we let the individual decide what information is relevant to analyze based on a rich, user-driven semantic layer that is networked across the organization. This collective, industry-specific semantic layer is then leveraged by Birst Smart analytics to generate insights that are new and relevant.
Second, the insights are outcome focused, easy to understand, and they are presented through Birst’s Value-Based Design methodology so users can immediately take the actions needed to improve the business.
5. AI has the potential to disrupt the BI market. Which vendors will be better positioned to take advantage of this disruption? Which vendors will fall behind?
AI should be able to efficiently crawl vast sets of data on behalf of each individual user to find the unique insights for that person. We already see this unique personalization in services like Siri, but it’s not to be confused with using voice recognition or natural language generation to easily get the insights in and out. It really lies with the algorithm to isolate and find insights that are unique and relevant to each person. It is this approach that will make vendors successful and where we think BI needs to go to truly get the maximum value out of data and adoption.
Mona Patel works in Birst’s Product Strategy team. With more than 20 years of experience building analytic solutions at The Department of Water and Power, Air Touch Communications, Oracle, MicroStrategy, EMC and IBM, Mona is now growing her career at Birst. Mona received her Bachelor of Science degree in Electrical Engineering from UCLA.