GenAI Personalization / Commerce Today Episode 131
===

[00:00:00] Hi everyone. Welcome to this episode of Commerce Today. It is officially summer, so I'm having a little bit of fun. Broke out the summer shirts and trying to beat the heat a little bit. But today we are gonna talk about a hot topic in e-commerce, and that is generative ai. And specifically, we recently talked about how generative AI is actually writing code for your e-commerce projects.

Now we're gonna talk about personalization a perennial topic in e-commerce and on commerce today, but the interesting shift I'm seeing is that you don't need a big, heavy customer data platform necessarily anymore to have really strong personalization on your site. Generative ai. Is really taking over that personalization niche.

And that's super important because 89% of e-commerce leaders surveyed said that personalization would be crucial to their success in e-commerce the next three years. They must be listening to commerce today. I'm guessing the other [00:01:00] 11% haven't found us yet. But as all of y'all know personalization is how you have to stand out with e-commerce these days that.

People will go to Amazon, they'll go to Walmart. Unless you can really personalize, speak to them, have a brand that resonates with them and connects with them. However, this is really hard to do historically, like no one even tried for true one-to-one personalization where each user was getting their own completely bespoke experience.

Especially since the vast majority of users don't log in and they may never log in. They may always use guest checkout or they may not log in until they're checking out. So if you're just going off of that, it can be really hard. Like you don't necessarily know which customer this is to personalize off of their historical data.

So that meant that the traditional personalization platforms that have been around the past few years, they were really bulky. They required a lot of costly infrastructure and it was just a lot of work, a lot of manual work, developing the content, targeting the content, et cetera. [00:02:00] But generative AI is coming to the rescue and is completely changing how this works.

So it can analyze massive data sets in real time, both from. Your own customer behavior on your site and that of other brands that are using these tools. And it can work in real time to infer the shopper's preferences. So we may not know that this is Jane Doe from Michigan, but we can probably figure out that people like this particular customer.

Are highly likely to be looking for X, Y, and Z. So this is probably why 72% of US digital retail leaders said that generative AI and specifically generative AI driven personalization is the top trend impacting their bottom line right now. So it's really creating this wave to where even small and mid-market businesses are able to give.

Amazon, like personalization without a really expensive customer data platform and all that, heavy legacy system work. So we're going to explore [00:03:00] how AI driven personalization works. Again, I told y'all last week a little bit about how things we're working with AI driven coding. Now we're gonna switch to the other side of the e-commerce world and look at marketing and personalization.

Also gonna go through some of the trends and stats and I'll have links in the show notes to all of these stats and give you some real examples. So first up, let's look at how generative AI is making personalization accessible. Basically, a lot of brands, especially the small to mid-size brands, you're able to go from data scarce to data rich.

So used to. You would need a customer data platform or an e-commerce platform that was gathering lots and lots of data to power these personalization systems. However, generative AI can gather, store and learn from customer info across every touch point, every system, every platform, everything you're already using, and it's able to infer a lot more from a lot less [00:04:00] data so you can have a unified view of your customers without a formal CDP.

And AI has gotten really good at analyzing those first party signals. So basically the clicks, the context, the purchase history of the customers on your site to predict what even the anonymous visitors might want. That's the big difference, is again, like I said at the start of the show so many people don't log in, so they're anonymous Visitors used to personalization systems didn't do much for those people, but these generative AI systems are able to do quite a bit, even when you're not logged in.

So over 70% of brands say that AI is gonna fundamentally change their personalization approach because it can react in real time. It's if you were able to have. One employee for every single person browsing your site. And that person actually would go around and customize the site in real time based on what they were seeing that person do.

So it's that one-to-one personalization. But of course you don't have to, hire hundreds or thousands of employees. You have AI [00:05:00] employees, basically a AI agents that can do that for you. So Gen ai, if you've seen the new tools that are built into Adobe commerce as a cloud service. Which I wish that had a shorter name.

Have you've seen the tools built into that? It can basically create product descriptions, product images. It can take your base product images and it can make changes to them. And we're not talking about make this image look a little bit better. We're doing things like, Hey, I'm selling a backpack, but this user, I can tell, is someone that really likes hiking.

I want the product photography and the product image to show maybe a water bottle and some hiking poles attached to the backpack. But oh, now someone else is looking at this backpack, and this is your stereotypical, soccer mom or soccer dad. And instead of hiking poles, they might need some accessories for going to the kids' soccer games that are in that product image with that backpack.

So being able to generate ads, emails, but especially product descriptions, product images, [00:06:00] tailored to each, not just segment, but each user is a huge breakthrough with these generative AI tools. Another thing that AI is doing is taking what I mentioned with the product images, but doing that for your emails and delivering truly personalized emails down to each individual user.

And just like with the development work that I talked about last week this allows small teams that do a lot more like teams can run leaner and they can get a lot more done than they were able to get done before. Which to me is exciting because I've always loved the classic underdog stories where.

Small companies are able to launch an e-commerce site and take on some of the biggest brands and really have a level playing field. Or sometimes the small companies actually have an advantage over the bigger brands because they are so lean and agile. So I already mentioned a little bit that, sometimes some sites have to, 98% of the traffic never logs in.

So that's the unknown, 98%. And lots of times these older personalization [00:07:00] tools really fell short. The newer AI models that are available for personalization, they can infer intent from contextual data. So sometimes it's something as simple as how long someone's been on the page and how much they're scrolling on that page.

It's almost like an example that I gave a long time ago and a commerce today episode is Netflix, and you've probably seen how Netflix, it's not just. What you thumbs up and what you thumbs down. Netflix obviously is using What shows do you watch? How long do you watch them for? Do you start a show and then stop it and never go back to it?

That type of behavior data to drive their personalization, to show you exactly the shows that you're gonna like the most. TikTok as well does a great job with this, but both those examples, you have to be logged in. Now, these new AI models, they can do this even if you're not logged in off of the single real time session that you are having on their website right then and there.

And it's interesting, this is not just in traditional e-commerce, traditional [00:08:00] products. We're seeing this even with services into it actually. Their CMO was recently quote quoted in one of their earnings calls talking about how. Generative AI is able to answer customer questions in a very personalized way when they're shopping for financial products because the AI is tailoring the responses based on that user's exact context without having to pull like a full profile from a database.

So it's able to, again, AI is watching you and it learns from what you're doing, and it personalizes very quickly. So some of the trends that I'm seeing around this that I think y'all should be aware of and start thinking about. So the adoption of this type of AI personalization is surging.

So majority of companies have actually either. Already I implemented this, or in the process of implementing this, 65% of retailers have a plan to implement generative AI for personalization within the next six to 12 months. McKinsey research is showing [00:09:00] that generative AI is unlocking 240 to 390 billion in retail value through these personalized experiences.

So obviously a lot of executives out there have heard that stat and they are jumping in. Also there has been a consumer behavior shift. Shoppers are starting to embrace AI powered experiences. I think there for a while it was either a kind of weird novelty or sometimes even a creepy experience, but now between the technology getting better and consumers just getting used to it last holiday season.

Generative AI chatbots actually had over 1300% increase in interactions year over year. People in the, for the most part, I know I've seen some horror stories on Facebook, but for the most part, people had good experiences with those generative AI chatbots and they were becoming more and more willing to use those, realizing that when they are useful they can really cut down your wait time and get you the answers you need faster [00:10:00] so you can move on with your holiday shopping.

So recent Adobe survey showed that 39% of US consumers have actually already intentionally used generative AI for online shopping. Basically went out to chat GBT or similar and said, help me shop. And 53% have said that they plan to. By this holiday season people are using it for product research recommendations, deal finding, et cetera.

92% of those that have used AI said that it enhanced their shopping experience. So again, we're getting to that point in the adoption curve where consumers are happy about using ai, they're willing to basically have AI watch them, even though they don't quite realize that's what it's doing. But AI, watch them as they're shopping and personalize these experiences.

Wind shoppers have had these AI powered experiences, whether it was through a tool like chat, GBT or the AI powered personalization they engage more. So they typically and again, this is another Adobe survey that I [00:11:00] will link to in the show notes, but they have found that AI referred visitors are browsing 12% more pages have a 23% lower bounce rate.

This personalized conversational guidance is making shopping much stickier. However, if you go back two years, three years I'm gonna have a little bit of an I told you so moment. So sometime ago years ago when I was starting to start talk about chat GPT on this podcast, I pointed out that the biggest thing holding you back from implementing generative AI was not gonna be the technology.

It was not gonna be budget. It was not gonna be time. It wasn't gonna be some technical aspect of the integration. It was gonna be data that you needed to have as much data as possible, and you needed to have it stored in a way that could be easily loaded into these generative AI models. Sure enough, 61% of companies surveyed have said that they actually aren't seeing the [00:12:00] success and the ease of deploying an AI personalization.

Solution as they would like, and as the sales pitch says, because of data accuracy and data access issues, it's that old garbage in garbage out rule. So having clean first party data in a way that is easy to share with the model is absolutely crucial. Good news is the AI tools are getting better and better at dealing with bad data and with data that is.

Harder to access or in more obscure formats. And it can start learning from the data that you do have. Obviously using something like Adobe commerce that has AI built in makes this a little bit easier because Adobe designed how the data is stored. Adobe designed the sensei AI model. It is obviously they've connected the two together.

That works really well. It's when you're trying to use, especially like a. Custom homegrown e-commerce platform and a third party AI personalization service. That's where you run into some really challenging data issues at [00:13:00] times. So looking kinda into what the experts and analysts are saying are gonna happen over the next year or two with AI personalization.

Really interesting report came out from Twilio. We've talked before about Twilio, Twilio segment. It's really one of the best small business and sometimes midsize business customer data platforms out there. And they are saying that the businesses that are looking at generative AI powered personalization as not just a step change, but a true transformation in how they're gonna interact and engage with their customers.

That's who's seen the most success. And they also, Twilio's president was interviewed and he agrees with me. He didn't word it quite that way, but he agrees with me and he said that the key is having a strong data foundation so that the AI can get to your data and is drawing on accurate data.

Basically, you don't need a giant fancy CDP, you just have to leverage the data. You have to feed the [00:14:00] ai. McKinsey has noted that in their engagements, the leading retailers are all rapidly experimenting with generative AI across both marketing and customer service, but they're still, most brands are still figuring out how to scale it out.

Their recommendation was the people they're seeing getting the highest ROI are focusing on a few high impact use cases and doing those really well. So don't say we're gonna put generative ai. And every single function and piece of our e-commerce website, maybe say, okay, we're gonna start with AI, personalized emails and AI chatbots for service.

And then we'll move on to, even things as cool as a full blown AI personalized homepage where every single visitor is getting a truly unique one-to-one personalized homepage experience. Forrester has talked a lot about. Generative AI skeptics and said that 60% of generative AI skeptics by the end of this year are gonna end up [00:15:00] giving up and they're gonna be willing to use generative ai.

So again, I think we are at that tipping point in the adoption curve where people will be comfortable using generative AI in the holiday season this year for 2025. I think by 2026, especially holiday 2026, it will be table stakes. It will be the consumer expectation. One industry leader in total retail actually said that they are seeing generative AI basically as an additional employee and in some cases as several additional employees that can really understand the data and drive personalization trends for your site and for your business.

So a few examples that I saw in my research. So there was one niche retail company with just 25 SKUs that used AI to auto generate product descriptions. And they were able to get their products listed out into all of the marketplaces. So Amazon, Walmart, and all those good [00:16:00] places substantially faster.

Another HomeGoods Etailer had an AI tool that created search synonyms for their search system on their website, and they saw a 2% increase in revenue per search, along with saving about two hours a week in manual work. So it wasn't a huge shift, but it was basically free. When you look at just how inexpensive some of these subscriptions are, you go out there with a $20 a month or sometimes even a free chat GPT account.

You can do things like generate those search synonyms loaded in your site and get that 2% revenue lift. Intuit has an AI shopping assistant that I mentioned earlier. It uses generative AI to handle very detailed, specific financial questions from customers and potential customers with personalized answers.

So if a user says, Hey, which of these credit cards is best? It actually looks at the user's data, what they've shared with Intuit, and will give a response tailored [00:17:00] exactly to that user, whether they're logged in or not. It's basically an AI powered concierge that you can give to literally every single one of your customers.

So really interesting things you can do as you scale up. Nike launched an AI pilot doing a generative AI powered shopping assistant and personalization system. They got it launched in three weeks as a proof of concept, and that's probably what excites me the most about generative ai right now.

Kinda the moment we're in summer 2025, is that. You can do things that used to take months or even a year, especially a company like Nike. Launching a new shopping assistant would probably be at least an 18 month project they did in three weeks. Like these generative AI tools are making it so much faster to get these things done and get them launched in live if you're familiar with them and know how to use them well.

So this assistant that Nike launched had real conversations with shoppers. They dynamically adapted recommendations to each user's profile and the real time [00:18:00] responses. It actually went through, classified all 6,300 SKUs that Nike had on their website in just a day and enriched the product data for better AI results.

This basically just shows that if a brand as big as Nike with the red tape, the data, the enterprise level systems that they have can spin up AI driven personalization within weeks and not years. You can do it too, for sure. Amazon, as always, is using AI and is creating more ads. Don't you just love all of the ads that are on Amazon Prime video these days?

What they're doing though with AI right now is actually generating custom product imagery for ads. So they will take existing product photos and put them into realistic scenes. And they've actually increased click-through rates for their advertisers by 40% by using this generative AI technology.

And I'm talking about generative AI technology, and I want to call out that almost all of this can be done with chat GT's, [00:19:00] API. Which means you don't need to go buy, you don't need to go through a big long RFP bidding process. Buy some expensive platform that has a high monthly or annual cost.

You literally go out there and set up a chat GBT account, have one of your developers do a little bit of API work. Next thing you know, you have these auto-generated product images. You have these, autogenerated ads like this technology is a lot more affordable if you approach it in an agile way and you can get these kind of enterprise level results even with the much less expensive and faster deploy technology.

So that is generative AI in personalization, and I think it's really leveling the playing field and just creating so many opportunities for innovation and for experimentation. You don't need a massive data warehouse. You don't need an expensive CDP. You can treat your customers uniquely through these tools and have that one-to-one personalization.

So [00:20:00] look at one area where personalization can move the needle for your brand. Whether that's product recommendations, homepage content images, like I mentioned, email marketing and pilot, an AI tool in just one area. First again. Consumers are embracing AI tools this year, and I think next year they're going to expect them and demand them.

And I don't want to see you left behind because of analysis paralysis or a habit of going out there and just having a long procurement process and buying a really big, older, slower system. Also look at your first party data. Look at the data that you have, make sure you are within all the privacy laws.

Storing and collecting as much first party data as you can in a way that you can easily leverage it with these AI tools. If you have an in-house development team, I would just go ahead and tell 'em, Hey, you can have a hundred bucks a week of chat G-B-T-A-P-I usage. They probably won't even hit that.

It's so affordable to go out and [00:21:00] experiment with. How can we feed our data into chat, GPT and have it return useful, valuable things, whether that's real time personalization for our customers or even just some of those things like search synonyms that we talked about. That would be more for internal use to cut down on how much time your marketing or e-commerce team is spending writing those search centers.

Nips. So personalization, we've talked about it for years, but it is truly here for each and every brand and even one-to-one Personalization, I believe is something that all of you can and should be deploying this year. If you're not sure how, if you're just totally lost, all of this went over your head, find me on LinkedIn.

My name is Joshua Warren. I'm the CEO of creativity. I have a creativity gold background in my headshot on LinkedIn 'cause there's a few Joshua Warrens. I have a link there that says, book an appointment. It's actually for a free 30 minute e-commerce brainstorming session. More than happy to take a look at your existing site.

Talk to you a little bit and give you some very [00:22:00] actionable, immediate ways that you could start using generative AI for personalization in your e-commerce business. So hope you found this helpful, and I will be back next week for another episode of Commerce Today. I.