Marketing and AI are getting married, and it’s a shotgun wedding. Marketing technology firms and the big AI platforms are furiously launching new capabilities that they say will help brands know us, feel our pain, understand our frustrations, intuit our desires and always send the right message at the right time to give us exactly what we want.
But is that true? Or will consumers feel outgunned and manipulated by brands with seemingly god-like powers of knowing too much?
At its Forge conference in Las Vegas last week, customer engagement platform Braze unveiled a sweeping set of AI and platform innovations designed to transform how brands interact with their customers. The centerpiece: AI agents: autonomous systems that can analyze terabytes of data, optimize campaigns, and craft content on behalf of marketers.
Everyone’s making agents. OpenAI recently released agentic capabilities inside ChatGPT, and yesterday announced an agent builder at its DevDay event. Competitor Anthropic, which makes the Claude LLM, announced its Agent SDK yesterday too, as did the AI voice startup Eleven Labs. AI agent use in customer service grew 22X this year alone, according to Salesforce, which is also building and shipping agents. Marketing software HubSpot just launched 15 agents. Adobe, which offers a marketing cloud with dozens of software tools for marketers, launched even more agents, and these are just a few of the companies launching AI agents.
Braze’s focus, like Agent Smith multiplying hundreds of times in the second Matrix movie, is giving every marketer super-powers via their own teams of intelligent, autonomous agents. Think winback agents (AI agents to get you back after you left. Cart abandonment agents (AI agents to coax you to buy that thing that you decided not to. Next purchase agents (AI agents to study your past behavior and intuit what you might buy next). And, of course, customer service agents and support agents and just about any kind of AI agent a brand can dream up for writing emails, sending push notifications, shooting you a quick chatty SMS or WhatsApp, and more.
We used to say “there’s an app for that” to reference the impression that there’s apps for literally everything imaginable.
Now, apparently, we’re approaching a future in which “there’s an agent for that.”
So we’re rushing towards a world in which most of brand communication to customers will be not just automated–which it has been for years–but tailored, customized, and even personalized by smart AI agents to help you get what you want. Or, perhaps, to make you buy more.
What’s that world going to look like?
Braze CMO Astha Malik says it’s going to make brands smarter about what they send you. That it might result in fewer messages and less spam.
“I want different things from different companies,” she told me. “That’s because I’m complicated.”
In other words: the out-of-stock beauty product she wanted but couldn’t get: yes, she wants an immediate notification via SMS or a push notification. The home furnishings she was just browsing? Nope, don’t even think about sending a real-time message. Maybe email instead.
On the face of it, that’s great for everyone. Consumers get what they want, brands get more sales, and everyone’s happy.
Of course, the cost is giving brands that you like and buy from your personal information. Contact information, preferences information, and–of course–all the data that you using their websites and apps reveals about who you are and what you like.
Not everyone at the conference was certain they liked the trade, including Dweeti Fanshawe, a marketer at EasyJet:
“Because of the nature of where we work, I analyze how they know this,” she told me. “Should I be giving them less data?”
Others were totally fine with the exchange.
Mallory Gordon from Pacify Health told me that she wants brands that she likes and trusts to brands to know, understand, and anticipate her needs. That she has a strong preference for timing-specific, relevant communications, and that she pays attention to companies who show genuine customer understanding.
“I connect with that so much more than just a generalized message of, oh, here’s 20% off of at the store … or a product you didn’t even want,” she told me. “And I think that’s true for a lot of people.”
For Braze product leader Kevin Wang, the key is brands implementing hard-core guardrails and exercising caution in using these newfound powers. Each message has a cost, and getting it wrong will feel like an invasion of privacy.
“It is corrosive to your brand to not be relevant,” Wang says. “I’ve gotten a bunch of really poorly timed push notifications. It just makes me think they’re dumb. And it’s like, I don’t want to work with dumb people.”
So: with great power comes great responsibility.
That’s true from a number of different angles, of course.
Marketing has always been about persuasion, and marketers have always had a lot of emotional tricks up their sleeve like scarcity (“only three left”) or appeal to authority (95% of doctors agree”) or social proof (“millions sold”) and the like.
The question is will AI supercharge this ability to persuade, and could this result in consumer harm? It’s at least worth thinking about.
Duke University’s Ali Makhdoumi says, for example, that companies have a growing information advantage over consumers and that “constant improvements of large language models and machine learning capabilities will likely offer even more opportunities for behavioral manipulation.” That’s mostly in reference to large online platforms, but brands are increasingly capturing and storing more data as well.
The good news is that you’re in control. Don’t like messages from a company? Turn it off in the app preferences, or–last resort–block the company. I’ve gotten some good push notifications from Uber, including discounts, but the company has yet to learn I only Uber when I travel … and so is sending me push notifications as I sit in my home office, making that off button very tempting indeed.
The other good news is that in spite of every advance in AI and agents, it’s still too expensive for brands to simply apply AI agents to every customer. The compute costs would be far too high, Wang said.
Marketing companies like Braze are fully aware of the potential problems, says co-founder and CTO Jon Hyman, which is why they’re applying AI to new features for message prioritization, ensuring brands send fewer but better notifications to their customers.
There’s an ever better solution, however.
Ultimately, we’ll each have our own agents, says Malik.
“Their agents are going to be talking to your agents … the whole notion of like ‘my people can talk to your people’ is going to be ‘my agents are going to talk to your agents.’ And that is going to happen,” says the Braze CMO.
We already see that somewhat in our email inboxes, where Gmail and Apple Mail apply prioritization rules that govern what we see, and what’s a little more hidden. They’re not yet using AI agents, but they are limiting distractions.
Our own trusted and reliable AI agents will fully restore the balance of power.
And ultimately, they could lead to the most efficient kind of market where indeed, you do get exactly what you want, and at–or close to–the price you want to pay, while brands do get a chance to get their messages and offers and communications in front of you.
Or in front of your trusted personal AI agent.
