How AI is Changing CMS Systems as We Know Them
The 'Amazon Effect' has been redefining the retail customer experience and blazing a trail to create a new standard of customization. Retailers need to pull out all of the stops to ensure the success of their eCommerce experience. With so many Amazon-type companies dominating the online retail landscape, personalization plays a significant role in a company's success.
Current commerce platforms use deep learning and analytics—machine learning (ML)—to help drive customer personalization. Today's customers are presented with individualized product recommendations, along with offers and deals that have been specially selected for them. The 'Amazon Effect' has opened up the eCommerce playing field, but many retailers still struggle to keep pace with the speed and breadth of eCommerce's technology evolution.
While effective content management has always been fundamental to personalization, the integration of artifical intelligence (AI) into content management systems (CMS) is playing a significant role in augmenting its capabilities. Gartner predicts that AI will become ubiquitous within new software products over the next three years. Retailers have become increasingly aware that their CMSs are vital for enhancing the customer experience and are gaining an appreciation for how AI is greatly impacting the CMS landscape.
CMSC Media connected with Rasmus Skjoldan, Chief Marketing Officer and Head of Product Management at Magnolia, to discuss the implications of incorporating AI and ML into CMS systems and how the developments affect retailers.
How are new AI and ML features inside modern CMS systems changing the way retailers sell online?
“The big push happening around digital retail experiences is AI matching audience to offers. AI is being put to good use in analyzing the behavior of the visitor, understanding her or his preferences and context — and then displaying more relevant offers. At Magnolia, we’re seeing increased interest in realizing the potential of this kind of hyper-personalization with the goal to put an end to all of the dishearteningly irrelevant experiences where a customer gets blasted with promotions that are unconnected from what they want.
We also expect to see AI being used to optimize the shopping cart flow in real-time. There are many visitor signals to listen to and to understand during the flow towards conversion, which can be used to decide whether a consumer is actually susceptible to upsell options, or if a more straightforward route to the final conversion has a higher likelihood of success.”
Could you provide some examples of how online readers could use AI and ML to automate their approach to content creation and personalization?
“AI-powered content creation is still in its infancy, but it’s clear AI/ML can be used to optimize, [as an example], combinations of content. Say a potential luggage buyer exhibits a preference for two-wheeled carry-on suitcases, and in her user journey, has also signaled interest in messenger bags, an automated combination of suitcases that fit with messenger bags can be assembled on the fly. This is not content creation but curation or real-time assembly of combined offers. When it comes to the actual creation of content, areas like sports results or weather are still at the forefront.
In retail, the risks of automated creation of content being off-brand or ending up in awkward experiences are still too high. We expect to see more AI-assisted content creation where the editor will get increasingly on-point suggestions for how to end sentences, what points to remember or what recommendations to give — will be in focus.”
What are some of the new ways that online retailers can employ to create customized experiences for each site visitor?
“Tricks, like changing the experience and product line-up based on contexts, like local weather and season, are surprisingly underutilized. Retailers have plenty of options when it comes to personalizing experiences, even without using personal information, based on context clues and user journey. Lead scoring is also still an important method when optimizing experiences on the fly. As soon as you start listening to the user’s preferences through analyzing the session, you get many hints for what to display next. Timing of interactions, scroll-depth, topical preferences, heat-mapping, and so many other signals give you brilliant clues about how to tailor the real-time experience.”
Why is the concept of hyper-personalization making headway in the world of online shopping experiences?
“If you turn hyper-personalization on its head, it’s called irrelevance. If you take a moment to do a meta-view on your own browsing behavior and analyze what is actually presented to you, you quickly realize just how massively irrelevant most of what is displayed to you really is.
Hyper-personalization is the promise to improve the poor status quo. Just think of hotel booking sites as an example. Every single time you look at a hotel option, you’re told if they accommodate dogs even if you genuinely dislike dogs. Or every single time you browse, you get the option to filter by swimming pools even though you’ve never searched for it or even if you can’t swim.
The current state of personalization is most often appalling — which is understandable because it’s so hard to scale. One hope I personally have is that more experiences would allow you to choose 'never see this again' — to tell the brand what you’re certain the product will remain irrelevant to you. Over time, that in itself would dramatically increase the relevance of the experience, based on your very personal preferences.”
Below, Magnolia highlighted prime areas where AI will provide a positive impact on content management in the near future:
"Image tagging: Currently, industry-leading CMS providers are already using well-performing application program interfaces (APIs) to help eliminate the need for cumbersome image tagging. However, as AI for CMS evolves, companies will be able to deploy it to automate and augment increasingly sophisticated tasks, such as mining content tags to drive perception analysis. Among leading CMS providers, work is already underway to use AI to provide content creators with an idea of how their content will be perceived."
"Audience segmentation: AI’s role in this area has the ability to be transformative. Currently, e-commerce audience segmentation is a time-intensive, manual process that involves defining distinct user personas and creating journey maps based on those personas. AI will help bring not only automation but also a higher level of granularity to the segmentation process. By deploying an AI-driven CMS, retailers will be able to create sophisticated segmentation that accounts for common customer behaviors like audience hopping — e.g., a young man buys a gift online for his great aunt, thereby deviating from his persona. The reality is that people are more than their segment, and AI will be able to process and account for the fluidity of audience hopping."
"Improved testing: When it comes to designing a personalized experience for the end user, you won’t always hit the mark on the first try. But AI can play a significant role in improving continuous testing that creates a better experience. By integrating AI, CMS providers will be able to pinpoint areas of suboptimal engagement and use analytics to automatically adjust the experience."
AI has become a game-changer for CMS systems, and while AI is making inroads to transforming and advancing CMS's design and productiveness, we still need to approach its adoption strategically. For retailers, there are a plethora of options out there, but when it comes to finding and integrating a solution, flexibility should be put at the top of the requirement list. CMS providers that offer an intuitive user interface along with the ability to easily integrate with other applications will, in turn, provide intelligent management along with personalized and engaging customer experiences.