In a recent interview with Eckerson Group, Joe Caserta, president of Caserta, a New York-based innovation consulting and technology services firm, said, “The next leap [in technology] is natural language processing for querying data” and predicted, “Keyboards are going to become something that we talk about in the history books that we don’t use anymore, in our life time.”

Natural language technology (NLT) is the building blocks of Caserta’s prediction and will play a major role in simplifying how humans interact with computers and do daily tasks.

According to a report from Tractica, a market intelligence firm, worldwide revenue of natural language processing (NLP) software will grow from $136 million in 2016 to 5.4 billion by 2025, a whopping 4000% increase. And if professional services and hardware is included, that number is $22.3 billion.

As NLT matures and integrates into more commercial products, everyone, including business intelligence (BI) users, will use their keyboards and mouse pads less. Eventually, voice commands will be the only actions needed to do most computer-based tasks.

NLT in a Nutshell

NLT constitutes the subfield of artificial intelligence involving natural language. These technologies include natural language processing (NLP), natural language understanding (NLU), natural language generation (NLG), and speech synthesis. All are used in commercial products and have reached different levels of maturity.

• NLP. Various definitions of NLP exist, but in practice many consider it to be the ability to transcribe audio into text and semantically understand it or abstract data from it. The part of NLP that turns audio into text is a mature technology. 
• NLU. The technology that understands the meaning of text and human intent, known as NLU, has a long way to go before it reaches maturity. Today, Siri, Alexa, and other virtual assistants use basic NLU to determine responses and actions. Although NLU is a distinct technology, most think of it as a subfield of NLP.
• NLG. Natural language generation software uses data input and a combination of templates, rules, and algorithms to generate text, known as narratives, that read as if a human wrote them. NLG tools enable users to automate writing tasks such as reports, financial news, and product descriptions, summarize and explain datasets and BI dashboards, and write mass personalized communications. To find out more about NLG, check out my recent report, The Ultimate Guide to Natural Language. NLG is considered a developing technology. 
• Speech synthesis. This is the ability for a computer to generate spoken language from text. This technology has been around for some time, and today, any smart speaker or virtual assistant has the capability to speak back responses.

Together, these four fields make it possible to interact with computers without keyboards. Once all four reach full maturity, a computer will be able to listen and transcribe audio into text (NLP), analyze the text to understand it (NLU), generate a written response based on the NLU analysis (NLG), and speak the written response (speech synthesis).

Everyday Context

Caserta’s prediction is easy to envision because it’s already becoming true. According to the Consumer Technology Association, smart speakers sales grew 279% in 2017 and predicts a 60% increase in 2018. People use these NLT-powered smart speakers and virtual assistants such as Google Home, Amazon’s Echo, Microsoft’s Cortana, Apple’s Siri, and others instead of keyboards to do simple tasks, from getting directions and listening to the news to buying items on Amazon and playing songs. And in the future, people will check Twitter and Facebook, watch YouTube videos and movies on Netflix, and much more solely through voice commands.

Soon, experts believe business professionals will find it easier to use voice prompts than typing on a keyboard to check expenses, send calendar invites, and write documents. Smart speakers will also benefit people that need a hands-free interface at work, such as package handlers and drivers. For example, store operations leaders at Dickey’s Barbecue Restaurants, a large barbecue chain with 550 locations, use iOLAP’s Enterprise Voice connected to a wireless headsets to check sales metrics, customer feedback, daily briefings, inventory, etc.

Obstacles. People in certain vocations, such as hardcore developers and professional writers, will still need or want keyboards. It is much easier to communicate accurately and succinctly via typing than speaking. Anyone who has ever transcribed audio will understand that when people speak, they often repeat themselves, do not finish thoughts, and jump from idea to idea. Also, using keyboard-less devices in crowded or noisy places, such as the airport, cafes, or open offices, does not lend itself to privacy and can interfere with transcription accuracy.

NLT for BI 

NLT already augments individuals’ daily lives, but it also augments BI users’ jobs. BI vendors integrate NLT products into their products to simplify the tools and so users can realize insights faster and with more consistency. BI vendors use NLP to drive search-based queries and NLG to generate summaries and explanations of visualizations and dashboards.

NLP in BI. In the last few years, vendors have integrated NLP capabilities into BI tools and other business software, primarily for natural language search, so users can type words into a search bar to generate an ad hoc query or find and pull up existing dashboards and visualizations. In some cases, BI vendors integrate smart speakers or virtual assistants, so users can speak their queries, rather than type them.

Here are some examples:

• Microsoft Power BI users can tell Cortana, Microsoft’s virtual assistant, via speaking or typing text to pull up existing dashboards and report pages.
• Sisense, a BI and analytics platform, offers an API so users can leverage Amazon’s Echo and use voice commands to widgets and dashboards to retrieve and hear numeric data as well as dashboard summaries. 
• iOLAP offers Enterprise Voice that connects to any backend data source (CRM, EDW, ERP) and voice platform (Google Home, Apple HomePod, Amazon Echo), enabling users to speak queries and receive answers back.
• Tellius offers a business analytics platform with a natural language search bar that users can use to query data.

In reference to using natural language search for Sisense, John Loury, president of Cause + Effect Strategy and Marketing, said,

“Even though the insights pulled through NLP are not earth breaking, they’re still fantastic to non-analysts. If that person can interact with Alexa or a chatbot…to access insights, we should nurture that. With NLP, it’s not just analytics on demand, it’s analytics the way you want them… and the ways that are easiest to consume.”

Today, natural language search may not unearth hidden insights or simplify data analysts’ jobs. However, nontechnical users gain benefits that typically require a data analysts’ skillset. This is indicative of the future where anyone will be able to ascertain data driven insights without knowing where the underlying data is or even that they’re using a BI tool.

NLG in BI. NLG customers can build solutions that analyze dashboards, visualizations, and underlying data and write summaries and explanations. For example, solutions built with Yseop’s Compose analyze dashboards and underlying data to summarize key insights, explain changes and trends, and prescribe next steps, all in natural language.

NLG effectively automates analysis so users don’t have to decipher charts, graphs, dashboards, and visualizations. This makes it easier for new users to use a BI tool and can increase adoption. Automated analysis also ensures that users interpret data in the same way and don’t miss significant insights.

Some NLG vendors, Automated Insights, Arria, Narrative Science, and Yseop, have partnered with BI vendors and offer plug-ins, public APIs, or prebuilt solutions for BI tools. For example, Narrative Science partners with MicroStrategy, Tableau, and Qlik and offers a plug-in that analyzes visualizations and underlying data to automatically write narratives summarizing key points. Users can customize the format and level of detail of the narratives. And narratives automatically adjust as users filter and drill down.

The Future 

NLTs have the potential to make BI tools pervasive and invisible. In the future, BI users will not recognize they’re using a BI tool. Analysts will type their queries into a search bar or speak them aloud and receive all the answers they need in whatever format they want, whether it’s a verbal response or a visual one, such as a chart, graph, map, or text. This will enable people to make faster data-driven decisions without having to learn to use a BI tool.

NLTs are going to play a major role in BI tools. New hires will learn tools in a matter of minutes and more people will make quick, data driven decisions. If you’re looking for practical steps forward, find out if your BI vendor partners with an NLG tool or has a way to do natural language search. Research NLG vendors and find out which is best for you and which ones offer a BI API or prebuilt solution. If you are looking for a new BI tool or are buying one for the first time, ask about NLP and NLG capabilities.