The Gap
There is a costly — and growing — insights gap between the few highly technical data experts who can analyze big data — and the masses of business users and analysts seeking insights to make better data driven decisions. On one side of the gap are data experts like data scientists using Python, Tensorflow, and SQL to decode signals from the noise of cloud-native applications, streaming social media, and other sources. On the other side are domain experts using BI platforms and spreadsheets — tools that were never designed to handle big data workloads. These users can perform basic analysis on aggregated subsets of the data which has been indexed, modeled, and or parameterized to reveal static relationships and trends in BI dashboards — but to go beyond “what happened” to “why” and “how”-type insights — requires help from data analysts who perform labor-intensive SQL slicing and dicing or data scientists performing advanced analytics. Insights are still largely handcrafted and the machine learning models necessary to get at these are difficult to create, explore, and interpret by the business.
Business leaders experience the insights gap in the form of:
- slow analytic request turnarounds by overloaded data experts
- inability to go beyond dashboards and get second-order answers that inform best actions
- lack of accessibility and interpretability of ML/AI to business teams
Analytics leaders experience the insights gap too, in the form of:
- growing backlog of requests from the business
- inability to explore, discover, visualize and model data in one place, leading to disjointed tools for different activities and unnecessary friction in the analytics process
- lack of a guided experience in current data science tool with limited “pushing” of insights, it is all “pull”
This gap causes analytic bottlenecks and costly business decision delays at a time when stakes are already incredibly high for firms to identify and act on insights or risk losing their competitive advantage. Some of these high level costs include:
- Inefficiencies: organizations are forced to spend a significant amount of time, resources, and advanced data science expertise to get actionable insights — resources they could have otherwise spent taking action.
- Missed opportunities: Delayed business decisions can mean missed opportunities for growth and innovation.
- Increased business risk: A lack of robust insight for intelligent decision making could drive risk up.
Finally, the insights gap grows with each additional investment in AI, ML, and alternative data models and sources. Business and Analytics Leaders alike agree that progressing from data to decisions is harder than it looks, and requires answers to what happened, why metrics changed, and how to improve outcomes, which unfortunately today reside in a hodgepodge of point solutions. Is there a better way to obtain insights from your data? Is there an easy way to bridge this great divide?
Introducing Tellius: the Synthesis of ML and BI
Tellius uses artificial intelligence (AI) to empower business users to obtain answers from their data directly—without a complex data science workflow. As a decision intelligence platform, it enables business users to obtain relevant insights from their data via an intuitive user interface that features natural language search and conversational queries. Natural Language Processing (NLP) technology allows users to ask questions in business terms. Natural Language Generation (NLG) technology presents the results in narrative form.
Tellius surfaces the most important findings without the complex feature engineering that typifies the traditional data science workflow—from discovering root causes to identifying key drivers, to making cohort comparisons, to identifying segments and clusters. This allows people of all skill levels to easily obtain insights from their data simply by typing a search string or asking a question, as shown in Figure 1.
Figure 1, Tellius democratizes analytics by making it easy to access data, gain relevant insights, and collaborate with co-workers.
While traditional business intelligence and data visualization tools focus primarily on aggregating data and curating subsets of data for manual analysis, Tellius automatically curates the insights that users need. Not only is it much more intuitive and proactive than traditional ML tools, but it gets smarter over time. The system gradually learns what metrics and data you are interested in, and then it monitors those data points and proactively pushes out new results.
Tellius selects the pertinent models. It knows which statistical algorithms to invoke based on the context of each query and the data that you wish to explore. It automatically finds patterns, trends, and anomalies in your data. It surfaces these artifacts as you issue queries and drill into the results.
Under the covers, Tellius utilizes a columnar database engine for processing BI queries and a high-throughput, in-memory compute engine for executing machine learning tasks. Bringing these two engines together significantly broadens the scope of your analytic initiatives and enables a much simpler workflow:
- AutoML technology shapes the way you query the data and how the system presents the insights
- A “Google-like” interface allows you to ask questions and obtain answers dynamically, either by typing or speaking
- Tellius automatically discovers relevant insights and explains the results in business terms
For example, you can click on hot points in a chart to interactively explore connections, correlations, and anomalies as you dig deeper and deeper into the data. You don’t have to employ advanced data science techniques such as Python, TensorFlow, or even SQL to discover anomalies, correlations, trends, and change-drivers in the data. Users access the intelligence directly, and obtain answers fast.
Tellius also automates data-preparation tasks by combining disparate data sets into cohesive business views. It examines the data to suggest new dimensions, joins, and KPIs based on logical connections and inferences within the data. For example, if your database includes columns for order time and fulfillment time, Tellius might create a third column that records the lag time between the two, so you can monitor how quickly you are fulfilling orders.
Figure 2: Tellius automates the entire analytics workflow, from forging connections to popular data sources to preparing data and making it consumable to business users.
Tellius is the natural synthesis of BI and data science. No other company unifies both sides of the equation. To learn more about this unique decision intelligence tool, download our new white paper, Bridging Insights Gaps with AI-Driven Decision Intelligence or give the platform a spin yourself.
Written by: Ajay Khanna,
Founder and CEO at Tellius
