With just one month left in the quarter, corporate earnings season is right around the corner and chances are the next quarterly earnings article you read will not be written by a business reporter, but by automation software.
In 2014, The Associated Press (AP) began using a platform called Wordsmith to automate the writing of earnings reports. Wordsmith, which was developed by Automated Insights, uses natural language generation to turn raw earnings data from Zacks Investment Research into a story, in AP style no less.
Before Wordsmith, the AP estimated that it was producing quarterly earnings coverage for approximately 300 companies. For context, at the end of 2013 there were approximately 5,000 public companies traded on major U.S. stock exchanges, each required by the Securities and Exchange Commission (SEC) to file their quarterly and annual corporate earnings reports.
With the new automated system, the AP has been able to produce 3,750 earnings reports every quarter – more than 10 times its usual capacity. Of those, only a fraction need a human touch, either by updating the original story or by writing a separate follow-up article.
Recently, NPR put Wordsmith to the test, going head to head against the algorithm to recap Denny’s Corporation’s first quarter earnings report. Following below are the two stories – one written by veteran NPR reporter Scott Horsley in seven minutes, the other written by Wordsmith in two minutes:
Which Denny’s earnings story did you like better? The differences are pretty obvious, but spoiler alert: the first story was written by Wordsmith and the second by Horsley.
Earlier this year, the AP announced it would further streamline its reporting by using Wordsmith to write sports stories as well, mainly college baseball and some college basketball and football stories.
So will platforms like Wordsmith mean an end to journalism? Not so fast.
On the surface, it may seem like Wordsmith could easily do the job of a business or sports reporter. For data-driven, just-the-facts reporting, automating stories like earnings reports and game summaries would seem like a logical progression and Wordsmith clearly does the job.
However, articles requiring any kind of “color” still require a journalist’s expertise, intuition and uncanny ability to tell a story. According to the AP, automation has “freed up valuable reporting time and reduced the amount of data-processing type work.” In turn, automation could allow reporters to focus on writing bigger stories.
Where automation could have greater impact is in the general arena of content creation. As a result of the escalating need for content and the increasing appetite to consume it, Wordsmith now generates millions of content pieces every week for partners that include Allstate, Comcast and Yahoo!
A quote from Automated Insights CEO Robbie Allen in a March 2015 WIRED article sums up best where the future of content creation could be headed:
“We sort of flip the traditional content creation model on its head. Instead of one story with a million page view[s], we’ll have a million stories with one page view each.”
How do you think technologies like Wordsmith will impact journalists and/or content creators?
Tags: Journalism, NPR, story telling, The Associated Press, Wordsmith, writing Filed under: INDUSTRY, Media, PUBLIC RELATIONS, Strategy, Tech Industry
1 Comment
Should you be pitching to algorithms rather than humans? - The Social Olive |
[…] Seattle-based Communiqué Public Relations. A version of this article originally appeared on the Communiqué blog. (adsbygoogle = window.adsbygoogle || […]