Published on January 27, 2017 by Boris Kraft
Many have speculated if the advent of artificial intelligence (AI) signals the beginning of the end of content management as we know it.
But AI can give your web CMS new life and make it much richer. After all, the best ideas often come from our darkest dreams.
2017 could well be the year of AI, with intelligent machines invading our lives on a daily basis. Think Amazon’s Echo speakers and Alexa assistant, Google’s Allo messenger or IBM’s Watson supercomputer. You already know this brave new world of self-driving cars, self-flying drones, virtual butlers running households and sensors to monitor hospital patients.
Yet content management systems, a logical area for AI, have yet to tap its potential.
Early inroads are being made. AI is being used in content management to analyze past and current consumer behavior, to help determine priorities, to predict which customers are likely to buy particular products or services and the best way to reach and engage them.
Google directs a large part of its millions of search queries to RankBrain, an AI system that interprets and processes the search terms or questions. RankBrain’s mission is to learn more about the semantics of your search and to teach itself how to better give you the answers it thinks you want.
The Associated Press (AP) uses the Wordsmith platform from Automated Insights for most of its corporate earnings reports.
Short stories are now produced in less time than it takes reporters. The number of stories increased from 300 human-created to over 4,400 automatically generated each quarter. Gartner estimates that 20 percent of business content will be authored by machines by 2018.1
Imagine an AI assistant that asks you a few questions about your business and the kind of website you want. Et voilà — you get a fully functional website, without any coding or project management. Platforms like The Grid or Wix already have algorithms and rules in place for creating websites.
Twitter bots can make this process faster and more streamlined, enabling marketers to publish and push content automatically across the platforms they want. Automated tweets can even be set to match user moods and emoticons.
EasyJet uses AI to predict customer choices from anything like flight destinations to food and drink items served on flights. The North Face and 1-800-Flowers.com use AI tools as shopping assistants: to quickly and correctly answer product questions and to make spot-on recommendations.
Airbnb uses an in-house AI tool that lets Airbnb hosts see how they should set their property prices each day to make it most likely to be rented. If a host has priced it right, those dates show up green on the calendar. If the price is too high, the dates show up red. The host can adjust the price on a slider and hit that sweet spot to cover costs and not lose customers.
A lot of AI seems to be directed at automating tasks in the content management chain, such as content creation, posting, tweeting and advertising. But is this AI? Is it “artificial” and “intelligent”?
In an ideal CMS world, we can dream of some AI applications.
An AI tool that identifies gaps in content and generates tasks such as “create new tree,” “populate sub-page” or “rename file resources” could potentially drive any on-the-ball web editor wild.
The AI site doctor identifies dead pages that don’t bring in traffic, flags up broken links and sends alerts when images and videos don’t display correctly.
These are currently rule-based. True AI would be bold enough to give recommendations such as “Your image and video sizes are not optimized for mobile devices” or “Your page doesn’t convert on Galaxy 5.”
AI tools can get faster results than conventional A/B testing. Imagine if you could simultaneously test different permutations of page and component variants. Instead of running one test at a time, monitoring the traffic and doing the analysis, you could mix several tests at once and let AI get the results for you.
AI-powered chatbots abound across sectors to provide customer support or get specific tasks done. Why shouldn’t there be one for your CMS?
So you could say, “Hey Webcrawler, I could resize this image in that content app, but why can’t I do it in this field?” And the all-knowing Webcrawler would find the answer for you, or even generate a ticket to your IT or devops support team.
Right now, AIs tend to be set up within relatively closed environments. What if they could talk to one another? “I’m trying to pull a feed from your website, but the format is not compatible and it’s not working. Can you fix on your end?” Could we be seeing the next escalation of spyware and business intelligence?
AI is based on predictable behaviors. But human behaviors are not always predictable or rational, and the subtleties vary across cultures.
One of the basics of content management is a consistent user interface or author experience. AI cannot create that. It cannot convey personal experiences or spin engaging stories. However, once the tone is set, it could possibly mimic and replicate.
And talking to robots can be frustrating. My recent encounter with a chat service saw me going in rounds because the mechanized responses just couldn’t understand my human impatience to solve the problem out of the box, out of the rules that they have learnt.
I’ve always been an advocate of the best-of-breed approach in selecting a CMS for your enterprise. So it’s no surprise that I believe the ideal solution lies in the combination of human and artificial intelligence.
Let science and technology be the backbone of making your content and engagement robust. Robots can reduce or take over repeated tasks that stretch over lots of data, tasks that can be automated or follow a logical roadmap, recurring tasks that are activated by specific actions and are time-consuming.
But let the emotional storytelling and social insights come from humans with all their complexities and imperfections. A marketer cannot plough through millions of blog posts daily. Use AI tools to track good keywords and popular topics. Then use the human touch to spin the tale.
AI won’t kill content management, at least, not in the foreseeable future. But it will challenge content management to become better at its core competence: asking the right questions of developers, authors and marketers, building compelling user stories, connecting AI applications, and knowing how to dock onto the raw data to create a strategy that resonates.
Boris Kraft has been creating and selling software since the age of 16. He is the Co-founder and Chief Visionary Officer of Magnolia. Boris is a member of the European CMS Expert Group and an experienced speaker and panelist. Boris is also a prolific writer: He likes to blog about all things Magnolia and regularly publishes articles in online and print magazines as well. He is a regular contributor for CMSWire. When’s he’s not thinking about the future of content management in Magnolia’s Basel headquarters, he loves to go sailing on Lake Lucerne, skiing in the Swiss Alps or admiring art around the world.
See all posts on Boris Kraft