AI has become a go-to brainstorming partner for marketers. These days, you don’t have to look very far to find someone using AI to help shape a marketing strategy. ChatGPT and Claude can produce campaign recommendations in seconds, making them an attractive starting point for brainstorming and planning. The real question is: How useful is AI when it comes to forward-thinking marketing strategy?
As a self-described AI skeptic, I try and give AI tools all the opportunity in the world to actually make a noticeable difference in the job that I do before writing them off. I haven’t seen anything that’s made me a believer yet, but I’m open to the experience. With that in mind, I’ve decided to run a little experiment with two of the most popular AI coworkers nationwide: ChatGPT and Claude. I asked each platform the same question I ask myself regularly when planning campaigns for our clients:
What are the three most important things higher education marketers advertising digitally should focus on over the next year?
Before looking at their answers, I wrote down my own. After more than two decades in digital marketing—including over 11 years focused exclusively on higher education—I wanted to see where AI aligned with real-world experience… and where it didn’t.
Answers from a Human (Rich Funk): What I’d Tell Any Higher Ed Advertising Team Right Now
Here’s how I answered the question before looking at either AI response.
1. Know Which AI Features to Turn Off
It’s going to be just as important to know what AI tools not to use as it will be which ones to use. In platforms like Google Ads and Meta Ads, more and more AI features are opted into by default, sometimes giving the platform control over your messaging, your images/video, and even the landing page traffic eventually reaches.
A lot of these are a big no-no when it comes to higher ed brand standards. Many other features can easily raise prices or spend too much budget on irrelevant traffic if you aren’t careful, so these kinds of features need to be understood and set properly.
2. Don’t Judge Campaign Success by Lead Volume Alone
Don’t define success and failure solely by leads generated.
Don’t get me wrong: Turning paid traffic into leads is still the main goal, but it’s getting harder and harder to turn prospects visitors into leads on their first visit (or second, or third in a lot of cases).
With multi-device and multi-touch attribution being such a challenge, success for campaigns shouldn’t be limited to just the leads it produces. There are other metrics (time on site, brochure downloads, etc.) that can also show value and relevance that should be a regular part of your performance reporting.
3. Teach Google’s AI What a Good Lead Looks Like
Feeding leads back into Google is about to be super important.
With the rise of Demand Gen and Performance Max, Google Ads prefers to take the wheel and find prospects for you without letting pesky things like “keywords” get in their way. Because of this change, Google needs to know a good job from a bad one.
Re-uploading the leads you get from both Demand Gen and Performance Max campaigns back into Google and telling them those are the kinds of leads that you want more of is one of the few ways for us to steer the direction of these more self-driving campaigns.
ChatGPT’s Answers:
For higher education marketers, the next 12 months are likely to be defined less by finding new advertising channels and more by adapting to three major shifts: AI-driven search and discovery, first-party data, and proving return on investment.
- Build for AI-driven student discovery—not just Google Search. Instead of simply ranking webpages, AI systems synthesize answers from authoritative content across multiple sources. Higher ed marketers should prioritize things like creating comprehensive program pages that answer real student questions as well as publish faculty insights, career outcomes, FAQs, and comparison content. The institutions that become trusted sources of information are more likely to be surfaced by AI-powered search experiences.
- Make first-party data your competitive advantage. Privacy changes continue to reduce the effectiveness of third-party audience targeting. Institutions that rely heavily on purchased audiences will likely see diminishing returns.
Instead, invest in collecting and activating first-party data:
- CRM integrations
- Inquiry forms
- Event registrations
- Campus visits
- Webinar attendance
- Email engagement
- Application behavior
The schools with the best data infrastructure will generally advertise more efficiently than those spending more on media alone.
- Measure enrollment impact—not marketing activity. Marketing leaders are facing increasing pressure to demonstrate business outcomes, not just campaign metrics. Move beyond reporting impressions, clicks, leads, etc. Instead, focus on things like cost per application, enrollment rate, and student lifetime value by program. This often requires closer alignment between marketing, admissions, and institutional research so that advertising performance can be connected to actual enrollment outcomes.
Claude’s Answers:
1. Optimize for “Search Everywhere,” not just Google
Students no longer start (or stay) on a search engine. The most impactful trends include optimizing simultaneously for Google, AI assistants, TikTok, and Reddit rather than treating SEO as a standalone channel. That means:
- GEO/AEO (Generative/Answer Engine Optimization) so your programs get surfaced correctly when students ask ChatGPT, Gemini, or Google’s AI Overviews open-ended questions like “best career paths in healthcare” — these ads and mentions are now triggered by the full conversational context of a query, not just keyword matches.
- Real presence on Reddit and TikTok, since more than 88% of prospective students now use two or more platforms during their search, and Reddit specifically plays a reinforcing role — 56% of students say something they saw there confirmed their choice of school.
- Accepting that a drop in website traffic doesn’t automatically mean marketing is failing — trust is forming upstream, off your site, more than it used to.
2. Let AI personalize creative and landing pages — but feed it well
AI is increasingly used to personalize ad creative and landing page experiences based on individual browsing behavior — e.g., a student who’s viewed your CS program page might get shown a peer testimonial video instead of a generic campus tour. The catch: this only works if you supply a wide variety of creative assets and have genuinely strong landing pages to personalize toward. Relatedly, platforms like Performance Max are pushing marketers from manual targeting toward oversight — success now depends less on hand-building campaigns and more on feeding the platform strong creative signals and knowing when to intervene, since lead quality and program fit still matter in a way they don’t for e-commerce.
3. Shift measurement from cost-per-lead to cost-per-enrolled-student
With 61% of enrollment marketing budgets now going to digital while spending growth has stalled, leadership is scrutinizing ROI harder. Institutions are being pushed to track full-funnel performance — cost per enrolled student rather than cost per lead or inquiry — using metrics like email-attributed applications, engagement by funnel stage, and yield rate by source. This also means moving past last-click attribution: tracking assisted conversions and cross-channel influence, since social and paid awareness channels often support conversions well before a form is ever submitted.
Human Advice vs. AI Advice
I think the answers we got are a great example of “close, but not quite” when it comes to relying heavily on AI for strategic advice.
Both LLM’s said that moving from a cost per lead to a cost per enrollment reporting structure is more valuable. I don’t necessarily disagree with this. In a perfect world, every school would be looking at these kinds of figures. But there are two real-world challenges to this kind of setup that AI most likely wouldn’t be aware of, not being a living breathing person working in higher ed.
Enrollment Data Isn’t Always Practical
- Number one, enrollment cycles are extremely long. Decisions can take months (sometimes years) and can sometimes take a few cycles to have a big enough sample size to base decisions on.
- And second of all, this kind of tracking is sometimes just not possible depending on data-sharing information, especially when working with an agency that may not be able to access that kind of data on a regular basis.
First-Party Data Is Only Part of the Story
Additionally, AI loves to suggest collecting first party data and tweaking messaging based on that info over time. It’s great advice, and it’s something that a lot of advertisers don’t actually do. But in higher ed, a lot of that is already being done. That kind of first party data is usually aggressively collected and organized in a CRM, which puts them way in front of the average e-commerce Google Ads user. The bigger challenge is connecting those systems, maintaining clean data, and activating that information effectively within advertising platforms.
The Verdict: Advertising Strategy Still Requires Human Judgment
Overall, the advice given by both platforms is sound and has solid reasoning. But as always, it’s important to take that advice as merely a broad suggestion. They can identify broad industry trends and summarize best practices, but it takes a human touch to bring strategic judgement. In the end, AI suggestions are meant for everyone. Your agency’s answers should be specifically tailored to you.