Why Your BI and Analytics are Incomplete without NLG

Natural language generation adds a new level of intelligent automation to BI and robotics process automation, freeing up knowledge workers’ time and enabling higher levels of productivity.


As we move forward into the digital age, we’re now creating 2.5 quintillion bytes of data every day. Smart companies and institutions rely on data to make informed decisions and create competitive advantages. However, with so much data to sift through, knowledge workers are taxed with extracting the essential nuggets and presenting their findings in an easily digestible format.

Technologies like BI (business intelligence) and robotics process automation (RPA) help alleviate some of the burdens, but they still require mining a vast pool of information in search of insights, which can be extremely time-consuming. As businesses try to get to the nirvana of real-time data analyses, it still takes several days for a person to convert charts, graphs and other data points into a meaningful summary.

To achieve positive outcomes, decision-makers must know the reasons behind key business metrics conveyed in corporate and financial reports. Natural language generation (NLG), an advanced form of artificial intelligence, converts structured data into highly valuable, easy-to-understand written content. This both demystifies complex analytics and extends the reach of data intelligence across all levels of an organization.

Put simply: with access to actionable intelligence from data, managers can make better-informed decisions, faster.

NLG is often embedded into business intelligence (BI) and automation platforms to communicate the full story from all underlying data. Gartner, recognizing the many benefits and value of NLG, predicts that By 2022, 25% of enterprises will use some form of natural language generation technology.

In many instances, the lag time between acquiring the data, interpreting it and taking action is where competitive advantages are gained and lost. And In some cases, the consequences don’t just affect profit margins and market share, but the ability to save lives.

In pharma, for example, NLG can help analyze large sets of data for clinical trials, in addition to pre-clinical studies, health records and genetic profiles, freeing up researchers to devote their time and energies to other essential tasks. It also allows researchers to more easily recognize trends and patterns, thereby allowing them to develop hypotheses and make decisions at a faster rate.

An international study published by the Center for Information and Study on Clinical Research Participation (CISCRP) last year shows how profound the need is for an easier way to provide clinical trial reports. The study of 12,451 representatives from pharmaceutical, biotechnology, contract research organizations, and investigative sites, found that an overwhelming majority (85 percent) of them would like to receive a plain-language summary of the clinical trial.

Yet, 61 percent said they didn’t receive any information on the findings from the trial in which they had an interest. Some experts noted that one of the reasons for this might be the difficulty in creating reports that can be understood by the layperson.

NLG can improve the integrity of the data collected during clinical trials and encourage trust in the trials and processes. Pharmaceutical companies that make use of NLG can streamline clinical safety reporting processes without compromising the integrity of their data.

Why we need Intelligent Automation and Augmented Analytics

BI and automation technologies like RPA are vital components to solving the challenges described above. NLG enhances and augments BI and RPA, creating solutions referred to as augmented analytics and intelligent automation, respectively.

The current COVID-19 pandemic gives us several examples of how NLG creates a whole new level of analytics that enables knowledge workers to populate reports complete with contextual narratives in a matter of minutes instead of days.

COVID-19 Research and Best Practices

During the 2020 coronavirus pandemic, the constant mantra from leading infectious disease specialists is that we need to flatten the curve to avoid overwhelming the healthcare system and to find alternative treatments until a vaccine can be created. We also must figure out when and how to reopen the country while minimizing a second spike in infections.

Think about how challenging it’s been trying to keep up with news reports describing how other countries are dealing with the virus, anecdotal research talking about antiviral medicines that may lessen the length and severity of the infection, and differing opinions about reopening non-essential businesses.

Arria NLG, a recognized industry leader, is working with Microsoft, USAFacts.org, TIBCO and Johns Hopkins University on two critical initiatives to turn COVID-19 data into fact-based, easy-to-understand summaries that enable individuals and media to serve their local communities better. Here’s an overview of how these initiatives work.

COVID-19 US and Global Tracking Reports

The first initiative leverages the Microsoft Power BI kit with the latest data from USAFacts to augment data with written summaries in real-time. Users go to arria.com/covid19 and select their state and a date range via an interactive dashboard to generate localized comprehensive narrative updates. For instance, a user from Pennsylvania who visits the site on Sunday, April 19 sees a state-by-state summary in the right column, including this:

PA is currently reporting 31,069 cases and 836 lives lost. During the past seven days, new confirmed cases have increased by 9,414. The seven days rolling average is 1,345 cases.”

In the left column, there’s a map of the United States that includes an interactive drop-down menu with three criteria the user can select, such as “Last,” “7,” and “Days.” By clicking on the state of Pennsylvania, for example, the user is immediately presented with a summary of every county within the state, sorted by active cases. The first county listed on April 19, for instance, says:

“Philadelphia County, PA with 8,502 cases and 136 lives lost. On Apr 18, 2020 there were 364 new cases and no fatalities reported. During the past seven days, cases have increased by 2,480, which means the seven days rolling average for cases is 354.”

Earlier this week, Arria NLG announced a second COVID-19 initiative, featuring a partnership with Johns Hopkins University and infrastructure software and analytics provider, TIBCO. Adding natural language generation to TIBCO’s freely available COVID-19 Live Report dashboard — which incorporates the latest data from Johns Hopkins University — provides a data-driven narrative that explains the graphical insights as you drill into the dashboard.

“Adding the natural language capabilities of Arria enables users to obtain dynamic interpretations of data science and visual analyses as they explore the applications,” says Michael O’Connell, Chief Data Analytics Officer, TIBCO. “The combination of visual analysis, data science, and epidemiology models on the effects of non-pharmaceutical interventions has led to significant uptake of these tools by our partners in the healthcare and life sciences sectors. This shows that, when properly applied, technology can be used as a force for good.”

TIBCO and Arria are together committed to helping the world navigate this time by empowering people with shared, accurate information. This dashboard is a free resource that delivers easy to read localized and global coverage for individuals to reference directly and for media organizations to better serve and inform their local communities.

“This is critical at a time when the sheer flood of data and information we see daily is nearly impossible to process without the help of technology,” says Sharon Daniels, CEO, Arria NLG. “This dashboard is a tool that cuts through the information clutter, gets straight to the facts and keeps the public informed.”

The Need for Advanced Analytics is Everywhere

Although COVID-19 related topics are dominating all the headlines at present, there will come a time when things go back to some kind of normalcy, and we can focus on other issues. One thing we’ll notice that has not waned before, during and post-pandemic is the need for reliable data that’s summarized for easy consumption.

Whether we’re looking to financial analysts for retirement investment advice, life sciences experts for clinical trials summaries, retailers for personalized shopping experiences or a host of other sectors for their expertise, there are three things we always need: accurate information, timely information, and information that’s easily understood.