Cut down on customization time and costs with Artificial Intelligence!

Interview with Antonio Comelli, Marketing Officer @ THRON

Relying on data rather than on your instinct can make the difference between a meaningful interaction with the visitor and a missed opportunity. Today, battles are won or lost on the basis of how well a company knows the users that crowd its channels and is able to address them as people via a personalized communication. In this process, Artificial Intelligence is taking on a role of fundamental importance.

Here at the Content Intelligence Network, we have involved Antonio Comelli, Marketing Officer at THRON, to tell us about how AI algorithms contribute to improving marketing team productivity and increasing user engagement.


D: Why has the hyper-personalization of the customer experience become imperative for brands? In what way can the integration of machine learning algorithms within your own systems contribute to the large-scale provision of customized experiences?

R: Nowadays, the digital landscape offers a wide range of content that can be used on a variety of channels, allowing anyone access to a myriad of information in a very short space of time. It’s something that the industry experts define as “content shock”: supply and demand do not go hand in hand and the quantity of content available exceeds the ability to consume it.

The huge crowding of content that characterizes the web today (and that continues to increase at an exponential rate) does not agree with the amount of content that can be “consumed”, or rather, read by users. This scenario means that it is fundamental to succeed in intercepting the user’s interest when, traveling along their customer journey, they pass through one of the brand’s touchpoints. Only with personalized content will we be able to capture their attention and ensure that they remain on our channels with the desire to further their journey: until they find the answer they are looking for to fulfill their needs, perhaps thanks to one of our products/services.

The effort involved in intercepting a user’s interest, which, by its very nature, may change over time, and the selection of the right content to offer them are activities that, nowadays, cannot possibly be expected entirely of the marketing team. Artificial Intelligence tools are required, specifically machine learning. We are talking about that ability that machines have to learn independently and, therefore, “map” people’s interests on the basis of the topics of the content used. Subsequently, thanks to an analysis of the data collected (number of views, behavior of similar users, browsing journey, etc.), they can identify what the best content is for each visitor.


D: AI enables intelligent automation, automatically adapting the content of pages in accordance with those who visit them. What do you think the concrete advantages are on a business level?

R: First of all, to ensure that this type of automation is truly effective, the choice of tool is fundamental. Going to integrate complex systems, such as image recognition or semantic analysis tools to classify content, view collection tools for each front-end channel, content recommendation engines and, especially, making sure that all these work in a synergic and integrated manner, would involve exorbitant costs for the company.

This is why it is necessary to start out by choosing a technological solution that allows you to get the most value out of your content in a transversal way and without any integration costs. Of course, once the right tool has been identified, the benefits certainly concern an increase in user engagement and retention on brand channels. As we were saying, if a user comes across content that captures their attention and deals with topics that truly relate to their interests, it is much more likely that they will continue in their browsing of these channels. And here, they will find the solution to their problem, converting and purchasing one of our products or services.


D: Could you quickly explain how THRON uses Artificial Intelligence tools to recommend the right content to the audience?

R: Within THRON, there are essentially two layers of Artificial Intelligence: the first is the one that intervenes when the content is uploaded to the DAM system. Here, image recognition, speech to text and semantic analysis algorithms process the content, automatically recognize its main characteristics (topics dealt with, subjects involved, editorial characteristics, etc.) and, on the basis of the information collected, classify it with tags that can then be altered or enhanced manually depending on the brand’s objectives. 

Instead, the second layer of Artificial Intelligence takes care of tracking interactions between the users who access the channels and the content viewed. Essentially, on the basis of a simple aphorism, “We are what we read”, it associates the content topics and characteristics with the profiles of the users who take advantage of it, identifying their preferences and interests in this way.

At this point, THRON’s neural network is able to aggregate all such data and, on the basis of algorithms that “weigh” views, content classifications, user profiles and their behavior, it is able to provide each visitor with the most relevant content.