(Over)personalization and AI

Mirror, mirror on the wall, is AI so great at all?

Artificial Intelligence is becoming a prominent part of our daily lives, both corporate and personal, garnering widespread admiration.

Renowned cosmetics company Estée Lauder used it to develop a voice-activated makeup assistant to assist visually impaired individuals in applying makeup, while Delta Airlines used AI to train their employees on how to behave before, during and after major disruptions.

Schneider Electric, a leading digital automation and energy management multinational, leveraged AI to connect its employees with opportunities for ongoing education and networking across the organization.

One of AI’s superpowers is personalization, empowering companies to gather various data about their current and potential customers and employ it strategically to drive better results. Possible applications include: 

1. Customer experience

AI can do much more than just addressing clients by their names. It can analyze customer behavior, purchase history, preferences, current weather conditions and much more to recommend relevant products and content. 

It can also create a unified customer experience across various platforms. For instance, by featuring products online that a particular person showed interest in while visiting a physical store. This is possible if employees use tablets to mark these products, and the information can be easily accessed and used by tools that influence product listings on the website. 

2. Marketing insights 

AI can analyze the activities of individual customers, identifying when they may want to place another order or when they might be at risk of switching to a competitor. Based on these insights, personalized marketing messages can be sent to encourage a specific customer to make another purchase, using techniques that have previously proven effective with that individual. 

3. Content creation 

AI-powered content creation is especially useful when descriptions, videos, pictures, and other assets need to be created for thousands of products. Additionally, it allows customers to place products on custom backgrounds. For example, placing a sofa in a picture of their own flat, so they can see exactly how it will look, thereby increasing their confidence in making a purchase. 

Overall, multiple studies demonstrate that personalization positively influences various parts of an organization. For instance, McKinsey found that it often enables companies to increase their revenue by 10–15%.

The same source also highlights that personalization is no longer just a "nice-to-have", but a necessity, since 71% of consumers expect personalized interactions from companies, and 76% become frustrated when this expectation isn’t met. 

Even though AI offers numerous benefits like content generation, data-driven conversations, analytics, and tool automation, it also brings new threats and concerns. 

Finding the right balance between personalization and over-personalization has proven challenging. The vast majority of consumers (70%, according to Salesforce) are becoming increasingly protective of their personal data, so getting the balance right becomes critical for a successful personalization strategy. 

So, mirror, mirror on the wall, how can we ensure that our personalization efforts don't cross the line? 

1. Conduct A/B testing to identify personalization strategies that resonate best with your audience and eventually drive better results.

For instance, you can compare the performance of different types of product recommendations. If you're selling bicycles online, you might test showing similar bike models based on what a customer has viewed versus recommending accessories they might need, such as helmets, lights, or locks.

Another option is to send out different cart abandonment emails, some can be more generic, such as "You left a helmet in your cart," while others can be more personalized, like "Still thinking about that helmet, Michael?" 

2. Be transparent with customers on why they’re receiving specific recommendations, as well as ensure that data you use is consented to and first-party.

Since customers are becoming more concerned about how their data is collected and used, it’s important to provide clarity to build trust and credibility.

If you are an eCommerce business, explain why a specific product or piece of content is being recommended to a particular customer. This can be done by adding phrases like "Recommended for you based on your recent purchase of X" or "Since you liked Y, you might also enjoy...", “Still thinking about Z? Here are some similar options.” 

3. Regularly audit the AI models your business is using for biases and update them to reflect cultural nuances.

It’s important to ensure that communication is not only as personal as possible but also appropriate and culturally sensitive.

For example, advertising food delivery during fasting hours in Ramadan across Muslim countries would be poorly timed. Instead, promoting these services before dawn or after sunset shows respect for cultural practices and can increase conversions and brand loyalty, as the right people receive your message at the right time. 

4. Allow customers to choose their personalization level 

Every person is unique. For some, even being addressed by their full name can feel intrusive, while others may prefer highly tailored experiences, where they don’t need to manually input details like delivery address, clothing size, or the number of people in the family when ordering dinner.

Therefore, to manage these different preferences, a good approach is to give customers a choice in how personalized they want their experience to be. For instance, customers could choose whether they prefer to receive emails with general updates and news, personalized offers based on their previous purchases, and/or recommendations based on popular trends in their region. 

Ultimately, the key principle is as follows, “yes” to personalization and “no” to over personalization. However, since there is no clear definition of what over personalization means, the best options are to ask your customers directly about their preferences, as well as experiment and iterate. 

It should also be mentioned that even though AI tools like ChatGPT can easily help add a personal touch to various communications, implementing AI models at scale and in a responsible manner requires the expertise of professionals.  

After all, you need to give it access to various types of sensitive data (e.g. customer, financial, legal) and allow it to make decisions based on that data. As your company grows, it becomes physically impossible to manually control and review everything your AI is doing.  

This is exactly where Youwe experts can help.

Get in touch to discover how AI can support the achievement of your business goals today!