A few years ago I conducted research on content marketers’ greatest pain points. The clear winner? Creating content.
The need for more content, for better content, for enough content, is burning through marketing departments charged with feeding the beast – many beasts, in fact.
All the channels, media and platforms that demand text, video, images — all of it content.
The content problem is hardly unique to marketing. It also pervades sales, customer service and support, business intelligence, HR, news and media, and even product, as devices and things become smarter and more communicative.
Organizations and content marketers alike are locked in a continual struggle to build, implement and scale effective content strategies, never mind achieve context and meaningful personalization.
For content marketers, creating enough content — not to mention content that is specifically tailored to specific audience and/or persona types — or to content creators generating “commodity” content (e.g., sport scores, weather reports, stock market results, etc.) is onerous, time- and labor-intensive, and can be a significant cost.
The rise of automation, and most notably artificial intelligence, marks a shift of the nature of how content will be created and indeed, how we will work. Content ideation and production is no exception. From research to design, from the creative process to distribution, applying AI to content changes how teams work.
Together with my colleague Jessica Groopman I very recently published research on content automation, entitled “Automated Content: How Artificial Intelligence Impacts Content Throughout the Organization” (the full report is available for download at no cost).
This piece explores impacts to the workforce, in and well beyond marketing. It will also discuss AI content fails, illustrating what can go terribly wrong, as well as tangible steps content marketers can take to prevent it.
Information velocity and a multiplicity of sources, not to mention constant streams of information, make aggregating and curating content increasingly difficult. Algorithms can “cherry-pick” as well as appropriately distribute the relevant stories, articles, and other forms of information on a given topic with little to no human supervision.
Artificial intelligence machine learning can analyze data, make assumptions, recognize patterns, learn, and provide predictions at a scale and depth of detail impossible for individual human analysts.
Machines can “see” and analyze images and “understand” what they contain.
Rather than relying on humans to append tags and metadata to content in a digital asset management (DAM) system, for example, that content can be scanned and labeled by machines. Editors who want more of a certain type of image, or to locate related images, will be able to find that material with a few keystrokes.
Rather than create videos from scratch, a content creator such as a blogger or journalist can enter their article or post into an application (Wibbitz is an example). In less than five minutes, a video on the topic will be delivered, created from available online footage.
Newsrooms are using natural language generation software to generate articles from data sources such as public companies’ earning reports, or the seismic activity reports generated by the U.S. Geological Survey to generate articles and stories. This frees reporters from tedious, repetitive tasks and frees resources for reporting that requires more human-intensive efforts.
Bots now routinely chat with consumers online, helping them to troubleshoot and to solve problems. When issues are more complex, bots can automatically “escalate” issues to a human agent, who in turn can then pass off an almost-solved issue back to the bot. As with the example above, it relieves workplace tedium and can lead to faster resolutions and happier customers.
Creating algorithms, making data analyzable and accessible, and having staff with the skills, both technical and human, to launch and maintain automated content programs will require organizations to hire, invest in technology, to train and to constantly monitor programs. Automating content is never a set-it-and-forget-it endeavor. Human intervention is always required.
Consider, for example, a recommendation engine that tailored an ecommerce site’s suggestions of additional products for consumers to purchase. It was programmed to suggest the sites most popular items. The best-seller was a Christmas ornament which (you guessed it) made it into the company’s Easter newsletter. “Most popular” is a good start, but clearly didn’t go far enough. There are temporal and holiday considerations as well for product recommendations.
In another automated content misfire, the “Los Angeles Times” automatically published “news” of an earthquake that occurred in 1925 and had to retract the report. The misfire occurred when a staffer at the U.S. Geological Survey was updating historical data for accuracy, which caused the software to read the report as new.
As Rob Bennett, CEO of Rehab Digital put it, “Algorithms left unchecked can go off the rails. Just because something can generate its own content and makes it own decisions doesn’t mean it has awareness of ethics, cultural nuance, what resonates with a particular segment at a particular time.”
Companies must balance the tremendous efficiencies gained through content automation with resources and structures to evaluate performance, metrics and data integration.
Rebecca Lieb is Co-founder and Industry Analyst at Kaleido Insights, and she was Editor of ClickZ from 2001 to 2008.
The post How content automation will impact the way we work appeared first on ClickZ.Reblogged 1 year ago from www.clickz.com