The changing relationship between retailers and consumers is a hot topic today. Consumer expectations for high-quality experiences continue to rise while retailers continue to fight for customer mindshare in a noisy, digital market. The solution many retailers have embraced to solve this disconnect is personalization through Generative AI in Retail. Yet the returns on personalization investments have been disappointing. Coresight research finds that 71% of brands and retailers think they excel in personalized marketing, yet only 34% of consumers say they experience excellent personalization from the brands they interact with.
“That’s kind of a pretty phenomenal delta,” says Deborah Weinswig, CEO of Coresight Research, in a recent conversation with Persado COO Assaf Baciu on her podcast, Retaili$tic.
Read on for Persado COO Assaf Baciu’s answers to some of Deborah’s questions (edited for clarity and brevity) on how retailers can leverage Generative AI in retail to close the gap between the experiences they offer and customer expectations.
How can Generative AI in Retail or Motivation AI — a kind of Generative AI for language — help retailers meet customer expectations?
We’re part of a big wave in technology called Generative AI that in the next two or three years will transform the enterprise in various places, from how you educate and onboard employees to how you write letters to communicate internally or with consumers.
The world has caught up with us, because Persado started about nine years ago. Persado Motivation AI uses Generative AI for language — or natural language generation — to help companies talk to their customers as if they know them personally. This means we’re able to generate language for every customer communication, for example, for acquisition, for apps, for cross-selling, customer relationship management, loyalty, customer service, and communications. We do this across channels, such as email, web, mobile, and social to drive or motivate customers to act. This is extremely important always, and even more so in the current economic climate.
How does Persado Motivation AI help retailers personalize the customer experience using Generative AI in Retail?
Imagine me, as a human, talking to somebody. If I know them, I know how to use language and emotions and storytelling to interact with them. As humans we do that all the time, one-to-one.
Persado allows companies to do that via Generative AI in a personal way and create a retail customer experience that is better, more accurate, and with more tailored communications.
The difference is reflected in behavior. Do they interact with what you’re telling them or not? Do they buy based on that or not? Do they become loyal or not?
We see that a lot of brands are somewhere on a path to personalization, but the results are not there yet. There are potentially two or three key reasons for that. One is that they are focused on building their data infrastructure and decisioning, but it’s not yet connected to the experience — as we know, consumers don’t experience data, they experience content. So if whatever you invested in does not translate to an experience the customer sees, it is meaningless and therefore does not have an impact; it misses the expectation of the consumer.
The other massive area is language. So assuming I personalize to some extent offers to a group of customers. And let’s say my Motivation AI decisioning engine surfaces 20,000 customers to send a proposition for a given product. These 20,000 people will get the exact same experience, language-wise.
If you use AI to personalize the language instead, you’ll not only give them a better experience, but you’ll get 41% more conversions on average. So despite all the efforts in personalization that relates to data and matching products to people, matching people to language is not often implemented, and that is a gap.
How do you get there? What’s the first-party data you need in order to figure out who needs to see what message?
When you use Generative AI in retail, it will generate language and variations. In our case, we have a language model built specifically for enterprise communication, and it has certain elements of language tagged in it. Those language elements are matched to the brand’s first-party data. For example, what did customers look at? What do people buy? If it’s demographic data (what age are they, what marital status), the AI is able to match the attributes the brand has with the language attributes tagged in our AI model and determine the best language for a given customer. And then it appears in the experience. So a brand talks to the customer as if it knew what he would like to hear about. That’s why we call it Motivation AI: it speaks the language of that person’s motivations
The difference in business conversions between using Motivation AI and essentially writing message A (or potentially a message A and B and comparing the two), the difference between them is 41% — that’s the difference we’ve measured on average with our customers with statistical significance over the last eight years.
It’s a sizable impact. I’ll give you an example. A midsize retailer today that has a growing online business with about 20% penetration online or digital would get between $20 million and $80 million in incremental revenue from motivating customers with language. In comparison, what most people do today is write something: they create A and B and that goes to market, and that’s it. The next time they write, they start from zero. There is no aggregated knowledge that is accumulated to guide the next generation. But knowledge of what worked to motivate is where Persado excels.
How should marketers adapt their strategy to address the likely challenges of 2023?
I think that companies should have a very good view of the channels that today drive acceptable returns in terms of customer behavior. So do customers interact on that channel or not? Do they buy from that channel or not? If so, go all-in on these channels. There is always a pecking order. You may have email and web and mobile and social and other areas. Not all drive the same value. I would focus on the top three, top two, or maybe top four.
I would definitely look to investments that bring immediate return and in a greenfield area. What we see that we think is a mistake today is that people replatform. They go from X to Y, from Salesforce to Adobe, from Adobe to Salesforce. And this replatforming at most provides incremental impact.
There are other underutilized ways to deliver value. Look at the hundreds or thousands of communications going to millions of people. When I ask a retailer, “Who wrote the message?” and they say, “It’s Johnny or Jenny,” I say, there you have zero technology. And yet there is proof that if you change the language, it’ll have an impact, and an immediate one. There are potentially other areas with more than incremental potential in the future. It’s now, and it’s in a greenfield area. This is what I would look for.
Can Motivation AI help on non-digital channels?
When companies use us on their digital channels, it’s measured. Ours is a machine learning engine and it learns how to talk to customers more broadly. Because we have this Generative AI language model that knows how to identify emotions that work as well as narratives and stories, we aggregate those to insights. And those insights are summarized in a brief that is given to the CMO on the brand side and to the agencies to guide them on production of communications that are above the line.
If I were the CMO and I was about to potentially invest in some Super Bowl campaign, I would look at this data to highlight the right emotions and other language factors to use for the target audience and identify how to convey that to them. Then, creatives can take that and materialize that to an ad. So that’s how we do it: for the areas we don’t touch directly with Generative AI in retail, we can influence via analytics.
To hear more from Deborah and Assaf, visit the Retaili$tic podcast page to listen to their entire conversation.
To learn more about Persado, reach out to us.