With software already writing routine stories for The Associated Press (my former employer) it’s only natural to wonder when such apps might start writing press releases, cases studies, Tweets, blog posts and other marketing content.
The AP claims its robo-writing of sports and quarterly-earning pieces hasn’t cost any jobs, but freed staffers to create more nuanced, in-depth stories. But as writing software gets better, it will be able to tackle more complex stories and (by extension) more complex marketing content such as white papers and responding to (just not creating) social media content.
For the sake of (choose your age bracket) our mortgages, our kids’ college funds and our 401(k)s we need to keep moving up the value chain and away from anything that is too formulaic and hum drum. Here are five content creation “fortresses” I see as safe for humans for a while yet.
Thought leadership: In a story about a game or a company’s quarterly earnings, the inputs (runs, hits, errors, revenue, extraordinary charges, etc.) are all well known. So is the format of the finished story. (“Gregg Jones led a fourth quarter rushing blitz that led the Panthers to a last-minute 33-29 win over the Cougars.”)
But I haven’t seen software that can suggest an idea or concept you can’t describe beforehand. That’s at the core of much of the work I do with clients developing thought leadership pieces. For one client, for example, I’ve spent weeks and dozens of hours reviewing background briefings and sitting in on discussions of their technology “vision.” It’s hard and necessary work to tease out and creatively package the unique content, and nothing I can see a set of business rules or algorithms tackling.
Cleaning Dirty Data: In the enterprise application world, dirty data might be three customer records for the same person with different combinations of first and last names, which can lead the company to think it has three different customers when there’s only one. Garbage in, garbage out.
In content creation, “dirty data” is raw material that is incomplete, inconsistent, unclear, loaded with jargon, too long or too short. How do you train an application to sift through a 60-slide PowerPoint full of buzzwords and distill the new, compelling message? How would you write a business rule defining “transform” when it can mean anything from cutting the cost of on-premise software to moving it to the cloud?
Asking the right questions: The more time I spend in marketing writing, the more value I find I provide by asking seemingly obvious questions. They might be as straightforward as “How do you define the cloud?” (Ask three experts and you’ll get four different answers.) Does any robot know to even ask the question, much less keep asking if the answer isn’t good enough to use in marketing collateral? Or creatively take four seemingly different answers and combine them into a new, compelling marketing message?
Coping with chaos: Software works great in defined, predictable environments where inputs and outputs can be predicted and rules created to respond to them. Ever seen a product manager or marketing campaign that works by set predictable rules with everyone following the workflow? How would you design software that can automatically reconcile multiple dueling agendas as new “cooks” in the form of product managers and outside agencies dip their spoon into the content stew?
Applying street smarts: I’m currently moderating a LinkedIn group for IBM on sales performance management software. When one study indicated some sales people are more motivated by “the thrill of the chase” than the size of their bonuses, I reacted with a very human “Really?” based on personal experience. How would you build such real-world knowledge into an application? Or, for example, the knowledge that sales and marketing staffs never seem to get along, or that security and operations staffs are always at odds because one is paid to assure safety, the other uptime? Applying such real-world perspective is essential to showing your prospects you know their business and can meet their needs.
Don’t get me wrong. Robo-writing software will get better at all these things, and maybe more quickly than we expect. That’s why we human marketing and PR types need to keep finding the “fortresses” of expertise that can’t be quantified in algorithms and business rules. Which ones did I miss — and are you worried about content creation robots?
Filed under: Uncategorized
Like this post? Subscribe to my RSS feed and get loads more!