How artificial intelligence is changing web content management

Profile picture for user barb.mosher By Barb Mosher Zinck December 10, 2018
Summary:
We've all heard about how AI is changing marketing software and analytics. But what about AI's impact on web content management? Yes, it's happening. Barb Mosher Zinck gets an inside look from Jan Haderka of Magnolia.

ai-workplace-machines
Some call it the core of the martech stack, others simply an important component. Where ever it sits in your stack, your web content management system will go through some major changes in the next few years. And most of them are related to artificial intelligence (AI).

It’s a favorite topic among marketers but mainly because they still aren’t sure how AI will improve the way they interact and convert customers.

Figuring out where AI is needed most

Ready or not, AI is here, and it’s slowly working its way into web content management. I spoke with Jan Haderka, CTO of Magnolia CMS, about the ways his team is leveraging AI to help Magnolia customers create better digital experiences.

Haderka said he’s been watching the hype curve for AI for a while and at Magnolia, they are always trying to understand where new trends (like AI and IoT) fit in the CMS. He talked about the time they built an IoT game with a partner and a client. When using public transportation, you would search for beacons. A global scoreboard let you see how you were doing and people could win prizes and have fun.

They look at AI similarly, trying to find ways to incorporate it and improve experiences. Haderka acknowledged that as a smaller vendor in the CMS space they can’t do everything, so they have to be smart about the changes they make, where things can lead and what it means for their audience.

Magnolia CMS is an open CMS. Serving as a hub, it takes data in from many sources, transforms it, combines it with other data and pushes it out to many channels. Haderka had to figure out how AI would help with this model.

Oh, the things we can do with AI

Image tagging was one way. You can upload an image and get suggestions for tags. Take it a step further and then use AI to relate that image with content residing in Magnolia and provide “related” or “recommended” content to the content author/editor. He talked about a demo they did where they were mining content for tags and sentiment as it was being written and providing insights on how that content would be perceived. It’s about helping Magnolia customers build a better experience for their customers.

Personalization is another key reason to leverage AI in the CMS. The manual approach of defining audiences and creating separate content and journeys for each audience in the web experience is often either too simplistic to make a difference or too complex to manage and scale. Plus, a person doesn’t necessarily fit into only one audience or segment, or they move in and out of a segment depending on different attributes (such as time of year). AI resolves this problem by automating much of the work.

Haderka said you can use AI and machine learning to automatically analyze how audiences are interacting with content and provide feedback on the performance of those interactions. It can help you understand if your audiences are working or if there are different segments or smaller groups you should focus on to deliver improved personalization and recommendations. It’s here that Haderka believes CMS vendors like Magnolia need to focus their efforts going forward.

It’s not just about one AI - but multiple AIs purpose-trained for specific tasks. This is why content management systems need to support an open architecture that enables them to integrate and plug in different types of AI at different places in the CMS to analyze content and provide the needed functionality.

Another area where AI benefits the CMS is with testing. You can feed analytics back into the CMS, and the AI can analyze the data to adjust experiences automatically. So, for example, if a certain audience is not reacting as expected to a certain personalization approach, the CMS can automatically change what that audience is seeing. In this case, the AI is continually testing and observing and improving the experience.

Integration, iteration key to successful AI

Your Web CMS may not have all these capabilities built in. Like Magnolia, you may need to integrate with many tools, connect them and feed the parameters to the AI, training it to work a certain way. It does take work. As Haderka said, the mechanisms are there to improve the system, but it’s still more work than training humans to analyze and make changes. In both cases, however, it takes time.

A key benefit of AI is that you can have multiple versions; one running live while you are training another one. So you always have some level of personalization working at any time. As Haderka pointed out, the frameworks are continually improving.

As for the CMS industry overall, Haderka said it’s a great time. He said AI is forcing customers to innovate. The biggest problem with innovation is that you will fail a lot before you succeed. But by allowing you to fail fast, you can cover the cost of the trials and get to success faster.

My take

Every web content management system will have some AI built-in in the next few years. It will revolve around improving the experience either by refining audiences or improving personalization.

But should everyone be leveraging AI and personalizing experiences? Is it more a B2C thing? Not so much said Haderka. It’s about the different types and sizes of your audiences. The more segmented your audience is, the more differences there are in terms of needs and demands. The bigger your audience is the more work required to personalize the experience.