Natural language querying includes a set techniques for improving the user interface for BI. This allows business people to generate queries, explore data and receive and act on insights in natural language using voice or text. Early implementations are focused on improving to meet existing metrics via voice, text queries and embedded chat channels for a larger number employees.
As the technology matures, natural language querying could be enhanced with AI guidance for improved insight, as well as natural language processing (NLP) techniques applied to back-end data analysis like conversational analytics and sentiment analysis. Conversational analytics evaluates customer service or employee-to-employee interactions, while sentiment analysis can help summarize consumer opinion from social media.
Some advantages natural language querying tools include the following capabilities:
Simplifying employee access
Corporación Hijos de Rivera S.L., producer Estrella Galicia and other leading beverage brands, is one early adopter natural language querying to improve access to BI for front-line workers, on top its MicroStrategy BI implementation. JJ Delgado, chief digital officer at Estrella Galicia, has been working on a natural language generation interface that simplifies the ability to add new types of queries.
"We believe augmented analytics changes the game for us," Rivera said. "Not only can our users connect to the MicroStrategy platform to surface business information and get the insights they're looking for by asking Alexa a question, but they can also, through our NLG technology implementation, read intelligent narratives that describe the analyses they are seeing. We will continue to explore these direct pathways to information that give us unmatched speed and make us a lot more competitive."
Marge Breya, chief marketing officer at MicroStrategy, said "Our experience shows that while there are large swathes of corporate employees that need immediate insights from their analytics solution, many are less comfortable with or lack the skills to interpret complex graphs and visualizations." Many users don't have the time to uncover the answers themselves.
Voice interfaces like Alexa will make it easier for these users to interpret insights and leverage voice-enabled technology to ask questions and get answers in a natural manner. For example, MicroStrategy has created NLP capabilities that can formulate a visualization from a sentence of text entered by a user, and it has added chatbot support and Alexa connectivity.
Driving deeper insights
Some experts believe that natural language querying could help drive deeper insights by lowering the expertise required to interact with the tools. Instead of being limited to data scientists or engineers, tools become directly accessible to business experts, which Gartner calls democratizing analysis.
Micha Breakstone, co-founder and head of R&D at Chorus.ai, a conversational analytics service, said "Many deep insights come about through careful iterative processes where queries lead to noisy results with a very subtle signal hidden within, and through careful clean-up, cross-analysis, projections, the noise is cleaned, the signal becomes clearer and deep insights emerge." The ability to use NLP for querying could vastly simplify such iterations and allow rapid progress as data specialists and non-technical business experts collaborate with systems that can speak both their languages, and ultimately lead to deeper analytics insights.
Natural language querying also hides some of the complexity of locating related data through multiple systems. For example, if a user might be challenged in searching for a data element relating to credit card numbers across several different data sources that use different names for the field, natural language querying could help locate and identify the different names across the systems, said Gal Ziton, CTO and co-founder of Octopai, a metadata management platform.
Natural language generation (NLG) enables the BI tool to create narratives from data so that trends, variances and exceptions can be described as well as visualized. The adage, "A picture is worth a thousand words" is often true, but for many people, there are many ways to interpret that picture or visualization.
"Narration describes a visualization so there is no ambiguity [about] what it means," said John Hagerty, vice president of product management for business analytics at Oracle. Additionally, for many business analyst roles, creating a narration of results takes up huge amounts of time. NLG accelerates that activity in a profound way.
Structuring the unstructured
The flip side of natural language querying on the front end lies in applying natural language processing techniques to make sense of unstructured data. "NLP makes sense of that unstructured data, making it organized, queryable and searchable," said Stephen Blum, founder and CTO of PubNub, a data management API provider.
An example of unstructured data would be social media data around a brand. The executives behind the brand want to make sense of what people are saying and how they feel about it. NLP can both categorize social media mentions by topic and analyze things like sentiment to better understand how people feel about those topics. This gives end users a new way to understand all the unstructured data out there, which an often-cited statistic claims could be up to 88% of enterprise -- yet some experts say could be an exaggeration.