Content automation explained: Wordsmith and the pros of an auto-prose hose

A photo of interviewee Robbie Allen on content automationTFM interviews Founder and CEO of Automated Insights, Robbie Allen, on Wordsmith, content personalisation at scale and ways to wring more out of marketing data.

There are two fairly wide scale problems that can seemingly be solved with a single technology. First, meaningful content takes time and resources to produce; second, there are mountains of data being collected but not necessarily utilised effectively.

That technology is content automation, or at least so says Automated Insights. The US firm, with its Wordsmith tool, has a plan to capitalise on both of these issues. We spoke to Founder and CEO Robbie Allen about the potential for the software.

At least, we’re mostly sure the responses we received were written by a flesh-and-blood human…

What, by Jove, is content automation?

The Internet is awash with it. A lot of marketers are obsessed with it. But does “content” mean in the context of content automation? Says Allen: “Content, in our case, means text.

“It could be a report, a social media post, a news article, an email, even spoken text from a device like the Amazon Echo [Amazon’s streaming audio smart speaker]. Just about any type of narrative can be produced using a content automation tool like Wordsmith.

“Automation simply means that the content is generated by the platform—often in very high volumes—rather than written out manually, piece by piece. Users tie data to a narrative structure, and Wordsmith automatically generates the content at scale.”

Who uses content automation and for what purposes?

Allen explains: “Our clients automate content to save time and money on content they currently generate, to produce new kinds of content that they never could otherwise (like customised stories for readers) or to make their data more understandable and actionable—or some combination of the three.

Allen gives TFM a few examples: “One of our best known use cases is with [newswire company] Associated Press (AP). If you see an AP story about corporate earnings, you’re probably seeing a story that Wordsmith generated or contributed to.” For example, the technology was used to create a story about Apple’s Q1 profits:
An example of content automation on Yahoo Finance

Another example of Wordsmith in action comes from Yahoo Fantasy [American] Football, where it is employed to generate personalised recaps and drafts reports that sound like they come from a snarky reporter:
An example of content automation from yahoo fantasy football

Other uses include Bodybuilding.com sending automated workout recaps to users of its BodySpace app and IoT electricity automation firm digitalSTROM which uses data from household devices to generate written and spoken reports for homeowners, so they can use resources more wisely.

Allen points out that a particular beneficiary of content automation might be the ecommerce sector: “Online shoppers are seeing an increasing number of product descriptions that are automated, engaging and designed to convert.”

Could content automation lead to content over-saturation?

In response to TFM’s suggestion that a world of businesses each churning out massive quantities of automated content could hasten over-saturation, Allen pivots: “We see a world of increasingly personalised content, each piece highly relevant to a single individual.

“One danger [for companies] is to not embrace tools to automate as much content as possible. Companies that do so can free valuable time for people to work on more complex tasks and projects that require human ingenuity—things that can’t be solved by artificial intelligence…yet.

“Firms that are late to adopt this type of automation may find themselves losing a competitive advantage.”

A fountain pen used by a wordsmith

A key tool for the personalisation generation

Allen agrees that content automation fits hand-in-glove with personalisation efforts of companies: “Personalisation is a big part of the reason we delivered 1.5 billion narratives last year. Many of those were made for an audience of just one or two people.

“The old media model was to write one story and hope a million people would read it. With Wordsmith, you can generate a million articles, each of them relevant to an audience of one.”

He points out that content automation is a way to put customer data to work, possibly even generating value along the way: “NBA basketball team Orlando Magic sends Wordsmith-produced emails to season ticket holders that offer personalised messages encouraging them to trade tickets they may not be able to use or resell for perks, helping fans get more from their investment and helping the team keep fans loyal.

“These messages, which simply weren’t possible before this technology became available, can transform a customer’s data into plain language and make it easily actionable.”

Could a slapdash content rehash lead to a backlash?

TFM asked Allen if consumers might have a problem with an automated content engine adding adjectives into previously subjective write-ups. For instance, we expect estate agents to get creative with property descriptions (e.g. “this charming two-bedroom maisonette, nestled in an agreeable countryside nook”), but Automated Insights includes similarly florid examples of real estate descriptions on its website.

Examples of real estate content automation

He places responsibility on the client: “Wordsmith empowers [the tool’s] users to make their own decisions about what kind of automated content they generate. Clients configure narratives in Wordsmith to employ whatever tone and style they want, or we can do it for them.

“The platform is built to be highly flexible, allowing for a lot of variation among narratives. For example, if a real estate information company is using Wordsmith to create property descriptions, they can apply logic to their template that says a property should be called ‘lovely’ only if it is located in certain areas of town. What’s more, ‘lovely’ can have any number of synonyms applied to it, so every property in those areas doesn’t come out sounding the same.”

“The content automation platform makes it possible to produce variable content on a massive scale.”

Automated Insights wouldn’t be the first technology company attempting to remain agnostic (within reason) as to how appropriately businesses use its tools—don’t shoot the messenger.

Content automation seems like it could be slightly murkier however, in that the tool is also crafting the message.

If an automated property description mistakenly includes the phrase “stunning view” in reference to the only house in the area that overlooks a landfill, it could lead to an uncomfortable conversation between agent, prospective buyer and seller. This is probably an unlikely scenario, but it raises the question: Who polices the resulting content? And what happens in cases where product descriptions are subject to regulation, such as UK advertising standards?

That said, one assumes there isn’t nearly as much scope for AI shenanigans as with Microsoft’s controversial Twitter chat robot, Tay, which went “full Nazi” within 24 hours after pranksters gamed its machine learning capabilities.

Bringing storytelling to “left-brained” endeavours

TFM’s recent interview with Stephen Ingledew, marketing head at financial services firm Standard Life, raised the idea of financial services being a bit “left-brain”, requiring some storytelling nouse to make sense of certain marketing and customer data.

Allen sees that there is something in the suggestion that content automation could benefit such companies and help them convey the stories found within otherwise dense spreadsheets and dry data.

“It’s certainly true that there are a lot of Wordsmith use cases around ‘left-brained’ endeavours such as business intelligence and finance. And it’s also true that we’re pretty fond of spreadsheets ourselves.” Indeed, Automated Insights recently released Wordsmith integrations with Microsoft Excel and Google Sheets.

“One reason that companies might work with us is to do with business intelligence. Charts and graphs often don’t tell the full story of what’s going on in a data set. They can be difficult to interpret and hard to take action on.

“You can imagine a manager at a financial services firm looking at a lot of charts, turning to an analyst, and asking for a few sentences about what the data means and what should be done about it. That’s what Wordsmith can automate—clear, actionable insights in plain language everyone understands.

“Firms are turning to artificial intelligence because there is too much data in the world for human specialists to parse, analyse, and explain. Next time you read through your investment portfolio summary, try to figure out if it was written by a person or by really smart software.”

Robbie Allen is Founder and CEO of Automated Insights – www.automatedinsights.com

TFM 2017 Keynotes and Seminar Sessions 

Learn from the achievements of Google, Shazam and BT, plus many more. Be inspired by thought leaders and explore what you can learn from the industry peers at TFM 2017.

Register now for FREE to be at the front line of martech revolution.
Thomas Quillfeldt

Thomas Quillfeldt

Content Contributor, UBM

Thomas Quillfeldt is a writer for UBM covering customer technology across Technology for Marketing, eCommerce Insights and Callcentre.co.uk.

In another universe, he enthuses about music and video games. Oh, and it’s “Tom”.

June 20, 2016

Get the TFM Insights newsletter

The latest marketing technology Insights

Download

The Future of Chatbots Report [Download]