AI eCommerce Personalization: What's New?

“This new AI technology is a game changer.”

“Imagine being able to outsource all your drudge work and focus on what you actually love about your job.”

“If you don’t adopt AI now, you’ll quickly be outcompeted.”

Do any of these sentences sound familiar to you? AI may be buzzy now, but for eCommerce marketers, it’s frankly old hat. People were saying all this and more way back in 2018 when I worked at an AI-powered eCommerce startup. And even then, it wasn’t really cutting edge. Today, it’s table stakes at best.

AI in eCommerce personalization is a pretty simple proposition. You need to make smart and intelligent recommendations to customers. You can do that manually with tons of related spreadsheets, or you can let AI do it for you, a la Amazon.

Today, AI can do a lot more. For example, AI has the potential to intelligently design new product types based on popularity. It can draft real-time personalized messages using text gen. You have at your disposal actually decent AI chatbots to help customers get the help they need. AI can keep track of and even predict stock levels.

Let’s talk about what eCommerce companies really need to be doing with AI.

A quick definition of AI

Before we get into that, I want to define how I’ll be using the term “AI” and how it relates to eCommerce personalization. While there is a temptation to think of it as a conscious, almost-human robot that lives in your computer, today’s AI is more akin to a faster way of analyzing spreadsheets.

In the context of eCommerce personalization, AI is not a smart robot that can know and understand things. It’s just a logic tree. For example, when I say, “AI can act as a customer service agent,” what I mean is that the AI will input the keywords of the customer message and output the most commonly used words in response based on its training data.

Take that as a warning – AI is not omniscient or conscious. But you can use that as a strength, too. Think of all the annoying, nitty-gritty, repetitive things you do every day that a computer can do better. That’s where AI shines.

Be cre-AI-tive

Now we can get into the promised meat of the article. If you take nothing else away from this article, it’s this: AI can do some pretty amazing things. If you’re limiting yourself to simple “If people buy X, they often buy Y” banners, you’re limiting yourself and your growth.

Here are a few directions to take AI.

Product demand

It may seem farfetched to think that AI can predict consumer preferences on such trivially different products such as Crest or Colgate toothpaste, but it’s not. New research from Harvard Business School and Microsoft shows that AI can already use customers’ tastes to predict whether they’d pick Crest or Colgate toothpaste.

This kind of study would typically cost up to $20,000 and several months of work. Now, “we can get those answers in under 15 minutes,” said James Brand, an economist for Microsoft.

Analyze product popularity

A while ago, I noticed that there was a new scam on Twitter. Artists frequently post their art on Twitter, and their fans often express their adoration by saying, “Put this on a t-shirt!” Unethical scammers used AI to quickly intake social chatter, and every time this happened, a bot would auto-reply with a link to that particular piece of art, pasted onto a t-shirt design.

You can take a less morally ambiguous route by using natural language processing to analyze reviews and social mentions around your brand and products. Then you can use that as inspiration to offer a product where you see demand.

In the future, we may even see AI + dropshipping to instantly personalize product suggestions based on customer input.

Personalized chatbots

I had a very unusual experience recently: I was able to resolve a customer issue by using the site’s chatbot. Pretty wild, huh? It took me about three minutes and my problem was completely resolved by the end, with no human contact. This is night and day compared to my previous chatbot experiences, which were frustrating at best.

A few tips if you choose to implement AI chatbots:

• Be clear that the customer is speaking with an AI-powered chatbot.
• Analyze the input to create FAQ content or blog content if the same question is coming up time and again.
• Always have the option to escalate to a human operator.

If nothing else, the rise of ChatGPT has shown us that AI can write extraordinarily well, at least in short bursts. While it can’t write a whole article at the click of a button without human prompting and instruction, it can easily chat with customers. This kind of personalization is often overlooked because it’s related to messaging, not sales, but it’s a great way to make customers happy. A specific message targeted to a customer’s problem is way better than a generic message that probably won’t help them.

Real-time, personalized messaging

In pre-AI times, marketers relied on simple behavioral triggers, like automated abandoned cart emails. With AI, you can do a lot more.

For example, imagine a customer is looking at a t-shirt product page and it’s sunny out in their location. AI can generate a personalized message saying, “Enjoy 10% off with SUNNYDAYS code!” Or, if stock is running low, AI could say, “Hey, it looks like we’ll run out of this soon - act now if you’re interested.”

This integrates really well with AI chatbot technology, as customers are already getting used to developing a rapport with AI chatbots.

Don’t leave monAI on the table

AI in personalization goes far beyond simple product recommendations.

Even this article just suggests a few things that could be outsourced. To really make the most of AI in eCommerce personalization, just think to yourself every time you do something manually: “How could I get AI to do this for me?”

AI is here to stay. If you haven’t already fully explored it, there’s no excuse left not to take advantage of the many tasks AI can do better, faster, and more efficiently, freeing you up to do the creative marketing work you enjoy.