Navigating the Complexities of AI-Driven Personalization in B2B Marketing

Stick the strategy

For all the excitement around GenAI, we’re still in the “Wild West” days of application, as MarketingEDGE previously reported. What’s more, workers face a rising tide of ambivalence — and doubletalk — around the tech: Why are stakeholders so split on whether we’re in a boom or a bubble about to burst? How much do LLMs actually learn, and how much do they parrot copyrighted texts? Do unrealized gains from implementations stem primarily from user error or vendor overhype?

To weather the uncertainty, leaders should anchor their technology initiatives in the fundamentals of good marketing.

“The main thing to recognize through all of this is that your brand is still at stake,” Edelman says. “It’s very easy to erode the trust and any kind of goodwill that you’ve built up from your brand by sending out something that’s inappropriate, that’s inaccurate.”

It’s why he advises his clients and students to start with a strategy “before we get to all the digital craziness.” Using that North Star as your guide, you can then home in on your target audience and unique value proposition; identify specific goals, risks, and opportunities around reaching them; and consider which technologies might help make it all happen.

Ensure that the first piece on brand is really buttoned up, Edelman advises. “For a lot of B2B companies, unlike consumer companies, they haven’t really thought through the full differentiated positioning,” he says. “But with AI, you have to.”

It’s also important, Edelman says, to maintain “rich data” internally (e.g., on how a target buyer interacts with you and the specific services they’ve used in the past) and collect it externally (e.g., on whether a company is expanding or contracting) by monitoring signals from investor calls, press releases and social media. “All of that becomes context, and it’s out there, and you can use that to be much more relevant.”  

What you do with that data will depend on “the nature of the markets you’re going after,” Edelman says. If your organization is similar to Cisco, where you’re selling high-ticket products or services to a limited number of clients, personalization can help you deepen your account-based marketing (ABM) capabilities, ensuring you’re using as much information as possible to generate relevant content for your targets, he explains. As an example, such depth-focused marketers might use a service like the AI-powered SundaySky to create custom videos for potential buyers, covering how a product meets their specific needs. “And so you’re thinking about the right delivery way to really make it feel personalized based on all the information that you have,” he says.

If, however, your business is more akin to Sysco, with higher-volume sales, your use case for personalization is likely closer to that of a B2C organization, says Edelman, meaning AI can help your team better understand what motivates your buyers by testing, learning from and optimizing different messages at scale.

He recalls an initiative from his time as CMO at Aetna, when his team was trying to incentivize plan members to implement healthy behaviors, such as getting their flu shot or taking prenatal vitamins. It’s in cases such as this — when you have a clear idea of what you want people to do but not necessarily how to get them to do it — that GenAI tools can help marketers “test and learn, test and learn, test and learn” how variations of a particular message resonate across segments, he explains.

Avoid risky business

In addition to grounding personalization in solid strategy, it’s important to steer clear of shortfalls. Edelman points to a few standouts:

• Getting too personal. Edelman recalls his wife’s discomfort when, in an attempt to sell her software for her psychotherapy practice, a tech company sent her dog treats because she often posts their pup Coco to her social accounts. “It so offended my wife that they were tracking her personal side and used that to pitch B2B software,” Edelman explains. “It was a complete misfire. And she then spread the word about that misfire and how intrusive it is.”
• Leaving generated content unchecked. You can’t “just let it go without having eyes on what’s going out,” Edelman says. “We’re still not there yet.” He stresses the importance of keeping a human in the loop to ensure an asset is appropriate, accurate, and reflective of your brand before it launches into the world.
• Overlooking bias. In addition to amplifying historical inequities, unchallenged bias can lead to missed opportunities. AI is “circular,” Edelman explains. “It’ll tell you the kinds of customers you should be going after are similar to the kinds of customers you’ve gotten in the past, but it doesn’t necessarily help you expand the markets you go after.”
• Not reading the room. To measure success, it’s important to look at where people are engaging with your personalized content, but also where they’re “actively disengaging,” Edelman says. Signals like poor open or opt-out rates could mean people are put off by, say, an ill-advised dog treat drop.

To make all the pieces of personalization fit, “everybody’s got to play together,” Edelman says. “It takes a village to bring all the data together that you’re going to use for personalization.”

For example, to initiate more intentional relationships with prospects, the executive team might tap marketing and sales to create personalized content in the sales rep’s name, product and billing for data to inform the development process, and customer success to handle any related inquiries.

“The customer may contact you through any channel,” Edelman says. “It’s your brand, and everybody contributes to the brand, and everyone’s at risk if the brand is violated.”