In an AI-Saturated Summer, Judgment Became the Most Important Marketing Skill

This summer, I taught digital marketing in a classroom where AI was already everywhere.

Not theoretically: Students were using it. Marketers were using it. Organizations were trying to figure out what it meant for search, content, social media, customer behavior and their own jobs.

Before the course began, I asked students what they wanted to understand.

The questions came quickly:

• How should marketers use AI?
• How is it changing search?
• How do you know whether the information it produces is accurate?
• How much content is too much content?
• What makes people trust a brand now?
• Which platforms still matter?
• How do you know whether something actually worked?

The questions looked like they were about technology.

By the end of the term, I thought many of them were about something else: Judgment.

AI made producing an answer easier

That is the obvious change.

A marketer can now generate an audience profile, content calendar, competitive analysis, list of keywords, campaign concept or draft strategy before lunch. Tasks that once required hours of searching and synthesis can happen in minutes.

That is not trivial. AI’s extraordinary advantage is compression.

It compresses the distance between a question and something that looks like an answer.

But the classroom made the other side of that equation impossible to ignore.

The ability to produce something quickly does not tell you whether the thing is right.

• Or useful.
• Or based on credible evidence.
• Or appropriate for this customer.
• Or strategically connected to the business problem.
• Or even solving the problem you thought you had.

Speed without verification just accelerates mistakes.

The work changed when the questions changed

One of the most interesting parts of teaching marketing is watching what happens when students stop asking: What should we do?

And start asking: Why would we do it?

A weaker strategy starts with tactics.

• We need TikTok.
• We need SEO.
• We should use influencers.
• We need an AI strategy.
• We should run paid search.

A stronger strategy asks:

• Who exactly are we trying to reach?
• What are they trying to accomplish?
• What do they already believe?
• Where are they getting information?
• What could make them hesitate?
• What evidence tells us this is the actual problem?
• What would have to change for the strategy to work?
• What would we measure to find out whether we were right?

Those questions are considerably less glamorous than prompting an AI tool. They are also where most of the value lives.

A buyer persona can now be wrong faster

Buyer personas provided a particularly useful example.

Give an AI tool a company name and some context and it can produce an impressive-looking persona almost instantly.

• Name.
• Age.
• Occupation.
• Goals.
• Frustrations.
• Media habits.
• Maybe even a stock image and a quote.

It looks “complete.” But that is precisely the problem.

A polished artifact can create the impression that a question has been answered when the underlying assumptions have never been tested.

The persona may be plausible, but plausible is not the same as true.

The better student work moved beyond demographic decoration to examining the decision itself.

• What is this person actually trying to accomplish?
• What creates uncertainty?
• What information do they need?
• What do they trust?
• What alternatives are they considering?
• What prevents action?
• What changes their mind?

AI did not make the buyer persona obsolete, it made unverified personas easier to manufacture.

That raises the standard for the marketer.

The same thing happened with customer journeys

Funnels look wonderfully orderly in a presentation.

Awareness ➡️ Consideration ➡️ Conversion

The real people inside them are considerably less cooperative.

• They search.
• They leave.
• They ask someone.
• They watch something.
• They forget.
• They come back.
• They read reviews.
• They ask AI.
• They compare.
• They abandon the process.
• They return three weeks later from another device.

My students repeatedly confronted the difference between the journey they wanted customers to follow and the one customers were likely to follow.

That distinction matters because marketing plans become much stronger when students stop designing around a diagram and start designing around behavior.

The marketer’s job is not to force people neatly through a funnel.

It is to understand enough about the decision to remove unnecessary friction along the way.

Trust kept showing up where I wasn’t explicitly teaching “trust”

This happened throughout the course.

• We would be discussing search and end up talking about credibility.
• We would discuss social media and end up talking about reputation.
• We would discuss customer journeys and encounter uncertainty.
• We would discuss AI and end up asking whether the answer deserved to be believed.

By the time we reached privacy, security, ethics and reputation, the pattern was difficult to miss.

Trust was not another topic in the curriculum, it was running through nearly all of them. That led to a question I kept returning to: What would make an audience hesitate to trust an organization online—even when the marketing appears polished?

It is an uncomfortable question, but also incredibly useful because so much of marketing today looks excellent—but still creates doubt.

• A beautifully designed healthcare site can make claims that are technically accurate but difficult for a patient to interpret.
• A polished social presence can generate attention without providing evidence.
• An AI-generated explanation can sound authoritative while flattening important distinctions.

A company can be visible everywhere and credible nowhere.

Trust is not the decorative layer we add after the marketing strategy, it fundamentally changes whether the strategy works.

AI clarity and human clarity are mostly the same discipline

One of the stranger developments this year is the growing industry around making organizations understandable to AI.

There is real work to do there.

Entities need to be clear. Expertise needs to be attributable. Information needs structure. Relationships between people, organizations, topics and evidence need to be legible across systems.

But teaching this summer reinforced something more fundamental for me. Many organizations do not primarily have an AI-clarity problem. They just have a clarity problem.

• Who are you?
• Who is this for?
• What problem do you solve?
• What makes your expertise different?
• What evidence supports your claims?
• Where can someone verify them?
• What should the person do next?

Those are questions a human needs answered, and they are increasingly questions a machine needs answered, too. We should absolutely learn how emerging discovery systems work. But there is a danger in applying sophisticated optimization to fundamentally unclear positioning. Sometimes the best thing you can do for AI visibility is make the organization easier for everyone to understand.

My students also reminded me that tools are seductive

Marketing has always had this problem.

Every generation gets a new tool that appears capable of solving the discipline.

• Email.
• Search.
• Social.
• Marketing automation.
• Influencer marketing.
• Data.
• Now AI.

The tool changes, but he temptation remains: we mistake capability for strategy. One of the harder things to teach—and one of the harder things to practice professionally—is restraint. A marketer who understands a new tool can tell you what it can do. A strategist also needs to know when it is irrelevant. That distinction becomes much more valuable when the cost of creating something approaches zero.

We can now make more.

• More copy.
• More images.
• More analysis.
• More campaigns.
• More variations.
• More content than audiences could reasonably consume.

The strategic question is no longer merely whether we can make something, it is whether creating it changes anything that matters.

The scarce skill may be moving

For a long time, access to information was scarce, then information became abundant. AI is making synthesis abundant, too and that changes where human value sits.

I do not think the answer is that humans must simply become more creative than machines. That is too easy.

The more important advantage may be discernment.

• Knowing which question matters.
• Knowing when the data is weak.
• Knowing what assumption an answer depends on.
• Knowing when a customer description is generic.
• Knowing what evidence would change your conclusion.
• Knowing when a tactic is fashionable but strategically irrelevant.
• Knowing when a beautifully produced answer should make you more skeptical, not less.

AI can help enormously with the work, but it does not relieve us of responsibility for the judgment.

Teaching is becoming a form of market intelligence

This is one reason I continue to value the classroom: teaching forces precision. Students notice when an explanation does not hold together. They ask questions practitioners sometimes stop asking because an industry convention has become familiar.

And each class provides a live view into what people are struggling to understand as the market changes.

This summer, AI generated many of the questions, but underneath them I kept hearing older, harder ones:

• How do I know what matters?
• Who should I believe?
• How do I tell whether this is working?
• What is the customer actually doing?
• What should I do with all this information?

Technology changes the surface of those questions, but it does not make them less important.

What I am taking into the fall

I started the summer thinking I would spend a considerable amount of time helping students understand new tools.

I did.

But the more important work became helping them interrogate the output.

• Ask a better question.
• Examine the evidence.
• Identify the assumption.
• Understand the audience.
• Connect a tactic to an actual decision.
• Decide what not to do.

Those are not anti-AI skills, they are the skills that make AI useful. And they are becoming more valuable precisely because producing a plausible answer has become so easy.

The students did not convince me that marketers need to become less technical. They convinced me of something harder: we need to become better judges.

Because in an environment where almost anyone can generate an answer, the competitive advantage may belong to the people who know which answers deserve to survive.