The Judgment Gap Between MarTech & Talent

‍

Marketing technology has always been good at speeding up work, but has our education on what constitutes “good” increased and evolved in tandem? AI can tell you what worked, but who makes the call on whether the work is actually good?

Jacey Berg, SVP of Strategy and Planning at Media Bridge Advertising, a women-owned full-service advertising agency, has spent nearly two decades in the strategy and planning discipline. In our interview, Jacey articulates a concern that simmers below the martech conversation and AI: this technology is arriving at precisely the moment when companies are reducing junior roles, limiting training budgets, and expecting immediately usable work. AI is accelerating production which is having an effect on the environments in which professional judgment develops.

The challenge for marketing organizations is no longer simply teaching people to use AI in martech, but ensuring that teams still acquire the experience needed to question what AI produces.

Judgment Requires Room to Be Wrong

Jacey traces her own development of professional instinct not to a training program or a methodology, but to an environment. Her first agency gave her bounded decisions she could own, defend, and revisit. In that role, Jacey’s boss offered feedback that distinguished between personal preference and actual error.

"If we were working on something, she would say, ‘this isn't how I would've done it, but you did a really good job’ versus ‘you did this wrong.’ I had room to chase theories and be wrong," says Jacey.

That environment produced a feedback loop that’s very familiar to science fans: form a hypothesis, see what happens, explain the reasoning, receive criticism from someone with more experience, revise the understanding, try again. Looping through the scientific method is how exposure becomes expertise. It requires time, low-stakes opportunities to fail, and someone senior enough to give a correction worth receiving.

Jacey specifically mentions what breaks the learning loop. It isn't ego, though ego plays a role, but pressure.

"The modern expectation is that everything is done perfectly,” explains Jacey. “We've decided there's no room to be wrong, chase ideas, try something different. That pressure leads to defensiveness."

Jacey goes on to explain why it’s hard to develop judgement in an environment of defensiveness. A marketer whose primary response to criticism is self-protection learns to produce safer work, not better work. The loop then switches to “how quick can I get this off my plate?” instead of asking "what can I learn from this?"

There’s an organizational implication Jacey points out directly, because judgement tends to get framed as a talent problem when it is often an environment problem. Companies describe good instincts as an individual quality — someone either has them or doesn't. 

Jacey's experience suggests judgment develops where people can be wrong without being punished for it, and stalls where every assignment must be immediately client-ready, polished, and unambiguous.

When AI enters this context, it makes the organizational/environmental problem harder to see.

Competent-Looking Is Not the Same as Good

Generative AI can imitate the surface characteristics of expertise, but how do we evaluate expertise? The output might be structured, plausible, and complete enough to pass initial review — especially from someone who hasn't built the judgment to evaluate it critically. 

A recent MIT study found that over-reliance on generative AI leads to reduced memory retention and reduced critical thinking. Jacey's explanation of the mechanism is precise: "It's because you're having someone else do all the work to think. And generative AI doesn't think. It pulls information from a variety of sources, it finds patterns, combines ideas. That's not thinking and ideation — that's pattern recognition."

The output looks finished before anyone has decided whether it is good. A recommendation can be factually coherent and still be strategically useless — missing the client's actual objective, the constraints that matter, the competitive context, the trade-offs the organization is willing to accept. None of those things are in the prompt.

Jacey doesn't argue against using AI. She uses it constantly, but for specific, bounded tasks: finding a peer-reviewed study, sizing an audience, navigating regulatory updates in healthcare. "I'm not having it build from A to Z," she says. "AI helps me review the process but it’s not the final producer."

"We have to really remember to use our expertise and our critical thinking, not to just take generative AI at face value," Jacey says. "Go back to basics, go back to the brief, brainstorm. Marketing is best when it's collaborative, a combination of arts and science."

Her practical instruction: "Don't assume it's right. Don't assume it's smarter than you because it's a computer, and push back on generative AI the way you would with a peer or a junior employee."

And the standard she holds herself to: "We first have to value our own expertise before someone else can value it. If you want your expertise to be valued, you need to feel it is valuable as well — and not just replicable in ChatGPT. I think some people are starting to forget that once upon a time they knew what good looked like."

Investing in Judgment, Strategy, and Relationships

Jacey's challenge isn't to the efficiency argument for AI, but what happens with the efficiency once it's captured, and if efficiency is confused for effectiveness.

"I can't tell you how many times I ask clients, 'What does success look like to you?' And they'll say, 'Oh, click-through rate.' And I have to tell them, that's one way to measure success, but there are others," says Jacey. "Are you looking for an increase in unaided awareness? Incremental leads? More efficient CAC?" 

A media KPI is not a business objective. AI-driven optimization makes the gap between the two easier to ignore. Jacey cites a specific example, a Meta Advantage+ campaign can improve cost-per-click indefinitely while the underlying objective goes unexamined. 

"Anyone can use AI to run a Meta Advantage Plus campaign, letting it optimize every facet. You might get a more efficient CPC," she says. "But is that impacting the bottom line? Is it actually signaling success with your business objectives?"

The same logic applies to how organizations are treating their people. Jacey describes a shift she's watched accelerate: companies requiring employees to incorporate AI tools while cutting the training budgets and headcount those tools were supposed to free up. 

"They'll just proactively reduce head count on what AI should save them. So now you need to keep everything on the rails with fewer people while learning and implementing a new skill set." Her counter is blunt: "You have to invest in training your people. Batteries run out of power, and you have to continue investing in growth. It's not about eliminating junior employees in favor of AI work, but evolving how you're leveraging the next generation."

The marketers Jacey sees using AI well aren't using it to move faster through the same work, but expanding their visibility. 

"If you're just using AI as a way to do a task for you and just do it faster — it kinda takes a little bit of your soul away," she says. "The marketers who are using it to be better are using it as the lever to make them think bigger." Theses teams are applying AI enabled analytics to their reporting, spending less time managing the numbers, and more time thinking about what the numbers mean.

The category of work Jacey most wants marketers to protect is the kind that produces no deliverable and appears in no dashboard: building relationships across clients, vendors, and the industry at large.

"Go to happy hours after work, go get coffee, do outreach on LinkedIn, join a networking group, practice awkward small talk, deal with rejection, practice emotional reactions," she says. "You want to be able to navigate the human messiness — because I think that's just going to be even more valuable with computers doing the automation."

The Work MarTech Can't Complete for You

The durable advantage for an agency or brand will not come from access to the fastest content generator or the most automated media platform. Tools are available to any competitor willing to pay for them. But how do you differentiate? You might opt to build technology in-house, but what are the hidden costs of building your own AI marketing stack?

The true advantage will come from teams that develop judgement. Marketers capable of interrogating outputs, challenging objectives, explaining trade-offs, and navigating the human complexity surrounding business decisions. 

The capacity to judge properly takes time to build, requiring environments where people can be wrong, seniors who give criticism worth receiving, and organizations willing to treat development as an ongoing investment rather than a solved problem.

If AI takes over the work through which marketers once developed judgment, where will the next generation of judgment come from?

That's the organizational question the industry is in the midst of struggling with. But Jacey's read on the practitioners who are already getting it right, is that they aren’t treating AI as a finish line or a threat. There are marketers, she says, who are using this moment to become dramatically better at what they do — spending less time managing numbers and more time understanding what the numbers mean, less time on research and more time on building relationships. 

"The marketers who are using technology like AI to be better," says Jacey, "are using it as the lever to make them think bigger, and expand their range of judgment."

Listen now: Spotify | YouTube | Apple

Connect with Scott: LinkedIn | Website

Subscribe to What Gets Measured for more conversations like this.

‍