Personalization in the AI era

February 28, 2018 | 3 min read

jocelyn-callier

Jocelyn Callier

Director, Data Intelligence

Thought Leadership

Digital interactions leave traces that organizations can leverage. These interactions can help us better understand ourselves and provide better service.

This is the very basis of the personalization concept.

A Salesforce1 report shows that 72% of respondents expect a personalized experience and not receiving it can make them abandon a brand.

However, besides increased customer retention, what are the benefits of personalization? According to the Boston Consulting Group (BCG), personalized marketing programs generate revenues that are 6 to 10% higher than mass communication programs2. At Relation1, we’ve seen similar numbers with our clients.

Is everyone personalizing? No, they aren’t. Because, as simple as the concept itself may be, its implementation is more complex. On a small scale, if you look at the example of a small neighborhood shop owner, it's easy. A human being can easily recognize their clients, communicate with them in a distinctive way and even anticipate their needs. But how is it possible to industrialize this process on a scale suitable for a large organization? One of the answers lies in the other hot topic of the day, artificial intelligence. While it’s not yet possible to program artificial intelligence with a "Personalize" command, it is possible to break the concept into smaller tasks and assign them to multiple intelligence agents that will work together to achieve this end.

Returning to the example of the shop owner, it’s clear some essential elements must be kept in mind. For each of these elements, current technologies and artificial intelligence can accelerate the deployment of personalization.

Memory and recognition

Artificial intelligence and machine learning allow for improved customer recognition and highlight their distinctive characteristics. Given the volume of data to be processed, algorithms make possible groupings no human could foresee. They are also able to recognize behavioural traits useful for communication and anticipation strategies.

In another vein, significant improvements in facial recognition provide new customer recognition capabilities that go beyond the digital world.

Communication

For organizations, personalized communications are becoming more complex to manage. Various marketing tactics collide based on different stakeholders’ needs and the customer life cycle. Acting like an orchestra conductor, artificial intelligence makes thousands of micro-decisions human marketers can’t make.

In a global system that includes human-defined strategies and the automation of machine-based artificial intelligence processes, large-scale personalization becomes a reality.

Anticipation

The ability to create models that recommend the best products, services and content for each client and even predict their next actions is the purview of artificial intelligence. These models can be integrated at various levels, whether for customer communications or to anticipate inventory replenishment. Current models integrate more and more internal or external data (weather, events, news) and are that much more efficient.

Artificial intelligence acts as a powerful accelerator when it comes to personalization, but it alone is not enough. No matter how smart or savvy the elements may be, a roadmap must be established to ensure they work together for a common purpose. According to the BCG, the absence of a roadmap is the strongest barrier to the adoption of personalization for 60% of companies.

A roadmap must be drawn up based on business objectives, customers, data, technological and operational components, a performance measurement plan and a prioritized project sequence. The exercise is a demanding, in-depth one and must reconcile technological issues with the organization’s human capabilities. The clients we have coached through this process were able to execute impactful strategies in a relatively short time, with results that often exceeded expectations.

1 see reference
2 see reference

© Relation1. All Rights Reserved

Privacy Policy | Contact Us | Follow Us