We talked to 100+ marketing teams about AI. Here's what no one is saying.

In any industry at any moment in time, there are a handful of people whose experience puts them at the intersection of everything important happening. In marketing and AI, Alec Haase is one of those people.

Funny enough, I met Alec over 4 years ago, back in the really, really early days of Hightouch. Back then, he was one of our early customers as the Director of Data Products and Martech at a company called Red Ventures (a massive conglomerate with over 100 digital products).

Long story short, Alec managed marketing technology for more than a dozen brands and ultimately evaluated and standardized 14 different ESPs. And in that experience, he came to a fundamental realization that the future of Lifecycle and Martech wasn’t going to be built inside the CDP or the ESP; it was going to be built on top of all the data (i.e., the data warehouse).

These days, Alec spends his time talking to more lifecycle and martech teams than almost anyone I know. So I wanted to pick his brain on what he's actually hearing because what teams are saying privately, sounds very different from what's being said publicly (especially in my experience as someone who talks to marketers every week).

The good news for you is that we decided to record that whole conversation and share it here.

If you don’t have time to watch the full interview, I did my best to synthesize the most interesting bits of the conversation below.

Earlier this year, Alec hosted a dinner in London for a group of marketing and data leaders. He opened it with what he calls the “Rose and Thorn”, which is an exercise he and his wife do every day where they highlight one good thing and one bad thing.

He asked the room the same question, but framed it around their experience with marketing AI. Surprisingly, the answer that resonated the most with the room was the same answer for both rose and thorn: "We're trying a lot."

It sounds funny at first, but it actually speaks to exactly where most teams are right now. As Alec put it, "Teams are trying. They're trying a lot of things. They're building things. They're experimenting… They're getting budgets. But at the same time, there's a reason that they keep trying, because things aren't necessarily working." The reality is that many teams are still in the early stages of figuring things out.

When Alec maps out what he's seeing across the teams he talks to every week, he describes it as a spray. On one end of the spectrum, you have teams running everything through tools like Claude with fully agentic workflows, and on the other, you have teams just now trying AI for the first time.

The majority are stuck somewhere in the middle, running bolted-on features inside their existing tools. And while those features are easy to use, they don't really change the fundamental workflow.

But the deeper problem is what happens when teams actually try to push AI output over the finish line. "It's really easy to get content to 60 to 85% of the way there... You might be like, wow, this is really good. But then you bring it to your brand design team. They're like, no, this is off." 

Getting something to "the 96 plus percent of the way there, which is necessary to get something live, at least at enterprise scale," is where everything falls apart. So teams feel burnt, they conclude that AI doesn't work for lifecycle, and they move on thinking it’s a technology problem, when they never built the foundation (e.g., the underlying context) the technology needed to actually get results.

The teams Alec sees actually making progress all have one thing in common, and it is that they built the underlying foundation before they touched anything else. In his words: "You have this team that's focused on foundations, brand guidelines, your logos, your fonts, your typographies, your templates, best practices for email. Someone's gotta create that foundation. Someone's gotta create the data foundation, your customer 360, all your events, all your purchase data, all your propensity scores.”

Once that foundation exists, generalist marketing teams can operate agents on top of it end-to-end, and "...they no longer need to know every little minute detail of how to code an email, for example, because the agent can do that." Every single decision feeds back into the system, so "...the agents now have access, and they can see how past purchasers performed, how high LTV users performed, etc." Alec added, "...as teams are starting to use AI more and more, the performance, the insights, and even the ability to produce good creative gets better and better."

If you want know how Alec and I are using AI outside of work, you’ll have to watch our full conversation. Both of our use cases were extremely impactful in our personal lives. 😉