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. 


4 Steps to Building Narratives with Quill

Now more than ever it’s imperative to leverage your data to its fullest while maintaining efficiency on your team. Quill, our intent-driven Natural Language Generation (NLG) system allows you to do just that. Let’s dive in to see what it looks like to work with Quill.

To ensure your team gets the most out of our platform, we go through a few core steps to get Quill prepped: data design, configuration, iteration, and quality assurance.

Data Design

Behind any good data lies a narrative. So in order to produce the best narratives, good data is fundamental. This means the data needs to be structured, standardized, robust, and clean. In this first step of bringing Quill into your organization, we work with your team to ensure we can regularly collect and process the data to the standard needed for your report.


At Narrative Science, we talk a lot about NLG and intent-driven systems. For your team, this means we need to understand the narrative you want written. This is where your subject matter expert comes into play to help us answer questions like: What is it you are trying to say and why? What are the rules around these statements?

There are patterns in how we come to understand, write about, and explain the state of the world as reflected in data. Once we have defined those patterns, we can begin to teach Quill.

Think of it this way - when you compose a report or data-driven narrative, you determine the purpose of each paragraph, sentence, and clause. You also question the intent of the author. Sometimes the intent is to compare a metric across two timeframes. Sometimes it’s to perform a distribution analysis across a group based on a particular metric or two. Sometimes it’s to call out key drivers of a metric or to explain the drivers of a performance metric over time. And so on. Giving Quill this knowledge allows it to configure stories that meet your needs.


Within the iteration phase, we begin the process of reviewing the outputs produced by Quill. Like any new analyst on your team, Quill goes through a learning process as it begins to create your narratives. It’s during this phase that we perfect and polish the outputs produced by Quill to create something that meets the needs of your business.

Are you writing a risk analysis, financial, or regulatory report? If so, you may want a more structured narrative with rigid language and less colorful writing. Quill can do that. Or are you writing product descriptions, a review of sales, or the effectiveness of a marketing campaign? In these instances, you may want more explanatory, varied, verbose, or colorful writing. Quill can do that, too. Within this iteration phase, Quill perfects its writing per your requirements and creates reports that meet your business needs.

Quality Assurance

In this last stage, we give the narrative a final review. Our team runs through the narrative with a fine-tooth comb, finding any errors in writing, logic, sentence structure, grammar, or anything else. All reports go through this phase to get the final stamp of approval. And with that, your new team member, Quill, is ready to go! 

choosing an ai solution


Privacy Policy | © 2018 Narrative Science