The New Narrative 

Want to know what makes us different? This is a great place to start. See how we’re innovating and experimenting, constantly. And get our take on the ideas of the moment— and where we think tomorrow will take us—across the industry and around the world. 


Why Your Next Performance Report Will be Written by a Machine

Nearly 50 years ago the world was introduced to the HAL 9000 in the movie 2001: A Space Odyssey. HAL 9000 was a sentient, all-knowing computer that captured our imaginations and painted a bright (yet sometimes scary) picture of a future where humans and machines work harmoniously together to accomplish great feats.

Today, 16 years after the movie was set, the world still hasn’t been introduced to a machine like HAL that is both informed and communicative enough to state, “I've just picked up a fault in the AE35 unit. It's going to go 100% failure in 72 hours.”, as HAL did in the film.

However, recent advancements in artificial intelligence (AI) and Natural Language Generation (NLG) are bringing the dream of a machine analyst within grasp of every executive, manager, and employee. The benefits of a machine analyst are vast. With the help of AI technologies, individuals can get the answers to tough performance reporting questions, that require the analysis of large amounts of data, in near real time.

To understand why a machine powered by Narrative Science’s NLG engine, Quill, will be your next analyst, consider the three main tasks an analyst is responsible for and how machines are assisting in this work.

Step 1: Processing performance reporting requests in the form of questions

An analyst’s work often begins by receiving a request for information from an internal stakeholder, such as a team member or supervisor. Questions such as “How many units were sold online yesterday?” and “What drove the increase in expenses last month?” Humans have an amazing ability to quickly parse requests and translate them into actions. On the other hand, machines have historically required extremely well structured inputs in order to complete the same task efficiently.

But, with advancements in Natural Language Understanding (NLU), machines are beginning to understand simple requests in natural language. Common examples of NLU are sentiment and topic identification algorithms that can take written text and extract meaning, such as understanding that a user wrote a positive review of a film.

An enterprise example would be the NLU software understanding an analysis’ request for the top three highest selling products by volume in California in the last seven days. Fast progression in the field of NLU will unlock the first step required for a machine to be your next analyst as it enables a computer to understand the analysis you need by comprehending your written or spoken requests.

Step 2: Aggregating and analyzing data to fulfill requests for information

Once the request has been understood by the computer, the machine needs to know what information is required to fulfill the request and what analysis should be done.

Quill is able to learn and analyze data from any domain once it is taught the entities of that domain, the relationship between them, and the common values associated with them. With this understanding of the domain, Quill can execute the necessary analytics to uncover the insights required to fulfill the user’s request for information.

In the case of our example, Quill would understand where in the data to find unit sales by volume, as well as having an understanding of how to filter and group the data by dimensions such as, region and time.

Step 3: Distilling and communicating the findings in natural language

The final step in the three-part journey to insights is for the results of the analysis to be communicated back to the individual that made the request. Simply displaying a list of field names, values, and calculations does not include the necessary context to efficiently fulfill the request for insights.

The most effective way to communicate the results of the analysis to the user is through the same way the user made their request, through natural language. With Quill’s NLG capabilities, the platform is able to communicate insights clearly in natural language.

In our example, Quill would express the insights by writing:

“The three highest selling products in California in the last week were The Model D (265k units), The Model J (244k units), and the Model Z (212k units)."

When all three steps are combined it’s easy to get excited about having your own machine analyst. When combined with NLU, Quill can understand requests for information in natural language, identify and analyze the relevant data to answer the request, and communicate the resulting insights in written or spoken English, all within a matter of moments.

Learn more about the value of automating your company's performance reporting and analysis by reading our white paper linked below:

the automated analyst


Privacy Policy | Cookies Policy | © 2018 Narrative Science