Personalization.AI: A New Paradigm in Customer Engagement
Posted on: May 17, 2018 | Posted by: Debjani Deb
In the last six months we have seen significant emergence and focus on personalization projects within the large enterprise. Today most of retail and banking industries, as well as other B-to-C industries such as hospitality and travel, are either deploying or planning to deploy funds towards initiatives on personalization. However, personalization as a concept is not new and has been batted around in different formats for the last decade. Website personalization, email personalization, personalization within digital commerce, have each seen significant focus. personalization as a concept is not new and has been batted around in different formats for the last decade.
- Method of Outreach: In the past, personalization has largely been defined by the outreach method e.g. email personalization, or by channel e.g. website. In today’s enterprise, these siloed personalization efforts are no longer able to scale to the demands of a multi-channel/multi-touch world, where users’ interaction is ranging from email, to push, to chatbot, to POS, to kiosk, to social. By definition this demands a new personalization layer that is channel agnostic, however sits across all of them. This leads to the definition of a new class of products in this category.
- Data: Today, enterprises have had to embrace the notion of a unified Customer Data Platform (CDP) that enables the bringing together of data from all systems of record. This is a necessary move due to the proliferation and pace of data and analytics being generated to harness insights and to take action on them. Personalization in the past used a subset of these sources and usually had hard coded connectors to the systems of record. In today’s world, the variety and volume of data demands that personalization platforms be able to access a multitude of systems of record, in most cases through an API or microservices layer that enables the most efficient contextual use of that data. Due to this, today’s personalization efforts are fairly strongly tied to the adoption of CDP and/or an API layer.
- Real-Time Triggers: The third dimension of difference between past personalization efforts and today, is our ability to detect signals or triggers within data that is in-flight or in motion. This is made possible with the leaps made in compute and memory in the past 5 years. This form of in-memory detection of triggers allows us to personalize experiences in flight based on data that is unfolding now, vs. crunching and analyzing past data to influence the experiences of the future. This ability gives the enterprise an entirely new dimension for personalization to deploy and deliver highly contextual experiences in real-time, in-session.
- Content versus Context: In the past personalization was highly focused on what the user sees, i.e. the content they see in the email or the content they see on the website. One could go so far as to say that personalization was also significantly focused on offers and the ability to understand the users propensity towards certain content and offers. In the new avatar of personalization, the definition has been broadened to context, which not only includes content, but also has added dimensions of time and place that defines the full experience of the user. In this avatar, the time, the location, the channel, all become as important as what the user sees. Therefore the focus is on user experience rather than purely marketing.
- Machine Learning: Finally, it goes without saying that the fifth and most fundamental distinction is that all the above defined complexity i.e, new channels, data, time, space, context, all lead to the need for reducing human defined rules to orchestrate experiences, and increasing the machines ability to understand the individual and personalize their experience along all these dimensions. I think of it as an influence zone for each individual, that is detected by the AI powered orchestration layer, based on the user’s location, the stage of the customer journey, as well as the past and present context that define the user’s need for personalization. The new personalization platforms should be able span data at rest and the data in flight specific to each individual, through a combination of supervised and unsupervised learning algorithms, to pinpoint the influence zone and appropriately personalize their experience.
There is a lot happening along each of the dimensions described above. New concepts and technologies are being introduced each day and I hope to discuss each in more detail in the weeks and months to come! Exciting times as we get to define a new paradigm in customer engagement that is highly personalized and powered by AI.