Curating, creating, and delivering fresh content pushes your marketing game to the next level. Your goal: Get the right content to the right people, when and where they want to see it. To learn about how to growthhack content creation, don’t miss this VB Live event.
The Associated Press, or AP, is an independent news gathering company that supplies content to 1,400 daily U.S. newspapers and thousands of television and radio broadcasters. A huge chunk of that content is quarterly earnings, essential to providing fair and accurate business reporting and research. But during earnings season, reporters are overloaded, turnaround time is tight, and the monotony and stress is always in danger of undermining both accuracy and speed.
So when a natural language generation startup approached AP in 2013, they leapt on board to experiment with content automation — could that system offer the speed, volume, and style AP needed to feel confident enough to turn over a huge chunk of their reporting?
With structured data about earnings releases and report templates, they launched their first major automation project.
“We went from our human journalists on the business news desk writing about 300 earnings stories a quarter to automating about 3,700 stories a quarter,” says Lisa Gibbs, director of news partnerships for the Associated Press.
In this case, there’s no artificial intelligence or machine learning, Gibbs explains. It’s simply about using tools to be able to generate a volume of stories far more efficiently than humans could do it.
“We considered that project enormously successful because we understood the power of freeing our journalists from what we would call routine, low-value news tasks,” Gibbs says. “We freed them up to work on more high-impact journalism.”
Instead of the entire department of business news reporters having to drop whatever interesting story they were working on four times a year to pump up the volume, now they’re able to zero in on the corporate stories that were truly interesting, and dramatically increase the amount of investigative work and the amount of coverage in other areas.
“Automation allowed us to shift our resources into higher level work that we want our journalists doing,” Gibbs says. “So the success of that project really signaled to AP that automation, and by extension AI, was an area of strategic importance for us. We really wanted to put resources around examining how these technologies could be used as tools in the newsroom to make us more efficient or help us gather insights in our news gathering, or help us produce the news faster — all sorts of things.”
The company now has a cross-company automation and AI strategy committee that regularly brings together representatives from news, product, technology, and strategy to identify projects that they think would be worth exploring.
“We look to identify areas as possible targets for automating,” she explains. “Is this something our customers, our media customers would want? Would this volume of content be useful for them? Is there staff-time savings involved if we could automate? Those are the kinds of things we take into consideration when deciding whether to automate.”
Today, they have projects underway in the area of natural language processing, summarization, image recognition, and tagging. They’re using machine learning based transcription tools to create transcripts of video much faster than ever. And they’re even doing some experimenting with machine learning and analyzation of big data sets to see if they can produce insights for journalists.
“We do have a matrix for evaluating automation and AI projects that gets at factors like staff time saved, increased content, and other revenue opportunities for that content,” Gibbs says. “We weigh that against the cost of the tools and the maintenance of the automation system. We have hard strategic criteria around these projects.”
Gibbs admits that the barrier to entry for smaller companies can be high.
“The investment is not insignificant,” she says. “There’s a certain level of data proficiency and expertise necessary if you want to really dig into these technologies. Or there’s the money that you need to spend to get the right tools to make it happen.”
But she suggests starting small, as with any new technology investment.
“Experiment with using some automation to help you test the waters without a big investment and gain some insights that will lead you further,” she says. Look at low-value routine tasks from blog posts to compiling email newsletters — and keep experimenting as the cost of entry continues to drop.
“The technology really is changing incredibly quickly,” she adds. “There are all sorts of interesting startups entering this field. What we think we can do with the technology today — who knows what we’ll be doing even a year from now? It’s exciting.”
To learn more about the tools you need to make content production more efficient, the techniques you can use to magnify your message, and the data you need to optimize, optimize, optimize, don’t miss this VB Live event!
By attending this VB Live event, you’ll learn how optimizing content can:
- Increase relevance for consumers across channels
- Leverage data to increase consumer engagement with content
- Increase the volume and scope of content, netting stronger conversions
- Seize customer journey opportunities when they appear
- Lisa Gibbs, Director of News Partnerships, Associated Press
- Ari S. Goldberg, Founder, Barber Surgeon’s Guild
- Peggy Anne Salz, Principal, Mobile Groove Consulting
- Rachael Brownell, Moderator, VentureBeat