Using R&D Strategically in WSJ Markets

By Alyssa Zeisler & Krista Schmidt

Traders have long used algorithms and sophisticated statistical models in financial markets. Do similar opportunities exist for reporters to apply artificial intelligence to enhance financial journalism?

How might we apply artificial intelligence to enhance financial journalism? PHOTO: iStock

This was the question posed by Charles Forelle, WSJ’s Finance Coverage Chief. In response, the R&D/Newsroom Tools Team partnered with the Strategy team and Markets editors to launch two initiatives focused on expanding and improving our audience’s experience.

Draft Assistant

Draft Assistant is a real-time article generator that uses Natural Language Processing (NLP) to create drafts based on language derived from existing WSJ articles and data APIs. First and foremost, this is about assisting our reporters to give them more time to focus on the analytical insights they bring to our journalism, and thereby serve our audience better.

The tool creates a draft for reporters and editors, capturing the tone and direction of markets and pulling in essential market data that is tedious for a reporter to write. From that starting point, a reporter can layer on analysis and insight — producing a richer story faster. With the computer focusing on the “what,” our reporters can focus on the “why.” We made the decision to follow this type of automation because we know our members value the insight our reporters bring to our journalism, and wanted to help our reporters get there faster. To reflect both contributions, these stories carry the reporters’ bylines as well as a tagline at the bottom of the piece: “An artificial-intelligence tool was used in creating this article.”

We worked with Narrativa to integrate their natural language technology with Dow Jones Market Data APIs and our own corpus, to create three different regional versions of Draft Assistant, complete with different indexes, assets, and rules for what features appear when. For now, the goal of Draft Assistant is working well to help our newsroom deliver markets news to readers more efficiently.

The Flexicle

The Flexicle is an article inset that provides additional context and explains market terms and concepts in a user-activated drop-down. Our goal was not to become an educational reference library, but to reduce the barriers to engaging with our stories. We wanted to make our stories feel less intimidating to new readers, giving us the chance to reach a broader readership. One idea is to eventually use artificial intelligence to adapt the content in the flexicle to the individual reader based on their behavioral patterns.

While we have published several articles with the Flexicle (here and here, for example), you’re unlikely to see the feature regularly used on our site for now given the complexity in implementing it across our multiple platforms. Nevertheless, prototyping and testing the Flexicle yielded important benefits:

1. We created a new feedback module to gauge audience reception (the thumbs up/thumbs down button) in experimental features, which is now being used for other novel story components.
2. We relied heavily on user testing and feedback throughout this process, which indicated an interest in this type of feature, so we are continuing to consider how we can best meet this user need.
3. We engaged with the tradeoff between fast experimentation and multiplatform support, and learned a lot about our expansive content ecosystem as a result. Our initial plan to test this on the web only had to be changed when we realized this creates complications for other downstream publishing platforms. Those lessons learned are now being applied to experimental features and our work on tailored experiences.
4. We collaborated across multiple teams: R&D, Innovation, Strategy, and News — itself a fruitful experience.

While these projects are early proofs of concept, we have received some great initial feedback from our audience and reporting teams, which provides a strong basis for future work — always led by the goal of solving problems for our newsroom and readers.

Krista Schmidt is one of the Journal’s two Lead Strategy Editors; Alyssa Zeisler is Research & Development Chief and Senior Product Manager at The Wall Street Journal.