Is there a place for AI in higher education marketing campaigns?

The progress of the internet, data collection, and machine learning have been moving at lightning speed. In 2017, folks were swirling on Big Data. In 2020, people wanted to destroy the third-party cookie (that poor cookie). Now with the release of ChatGPT and other platforms, we’re reeling on artificial intelligence — contemplating how AI might impact our world and the roles we wake up to every day.

For most folks, nothing’s changed yet, at least not in any way with doom-and-gloom connotations (unless you’re a college student copying and pasting an essay explicitly written through AI software).

But in our world of branding and marketing for higher education, AI is a tool we’ve been using for a long time.

We use it in tools like Grammarly that check our emails before we hit “send” — only because we don’t use Word anymore, and Clippy isn’t around to help fix our errors.

AI is also built into advertising platforms that help us do A/B testing for different messages and layouts, and that recommend search terms when campaigns are over- or underperforming. It’s even incorporated into creative tools like Adobe’s Creative Cloud, giving our designers more time for tasks like developing your brand or proposing a new campaign concept — things a robot truly cannot do.

Now let’s pull back the curtain and share some behind-the-scenes information on how our team uses AI to do the big work we do for our clients in the higher ed space.

The Creative (Sarah Martin)

Shortly after ChatGPT’s release, we started to see platforms like Midjourney and OpenAI’s DALL-E pop up. It was only a matter of time before AI would enter the visual space, but I didn’t think it would be so soon. In thinking about AI, my mind immediately shifted from descriptors like “fun” to “scary” and “bad,” and I wondered, “Is AI going to take my job?”

I started digging, I needed to find out if it was really that good. Can AI take something existing and make it better? Can it create a visual brand from some simple prompts?

Short answer: no, it can’t. At least not without a human behind it.

For platforms like Adobe Firefly, DALL-E, or Midjourney to work well, you need to write a descriptive prompt. These prompts can be nearly anything, like “Create a portrait of a dog in the style of Caravaggio” or “Make a flat, photo-realistic wooden texture.”

My results with both these prompts were incredible. The platforms give you different versions, and you can pick the best one. It’s unbelievable.

Then, using Adobe Firefly, I tried, “Make a simple ad using Ohio State University colors and text about in-state tuition costs.” The results for this prompt were… not so impressive. It had no idea what Ohio State (my alma mater) looked like, and didn’t pull any visual or verbal language from its brand. In fact, it didn’t even include any decipherable text, although it looked like it tried.

After entering more prompts similar to this, my thoughts about visual AI soon shifted from “Will AI replace the design process?” to “How can I use AI to benefit my work?”

When creating concepts for a brand, it’s easy to go down rabbit holes, trying to find a specific style for photographs, textures, and other graphic elements. On my own, it can take me most of a day to find the exact type of photo I’m looking for. With the right prompt, however, AI can create a texture, draw an illustration a specific way, or mimic a photograph style in just a few seconds. And with Adobe Photoshop’s new generative expansion tool, I can take a cropped photo and use AI to extend the image beyond its edges to fit a layout perfectly.

Using AI like this — to quickly create simple outputs like graphic elements and specific photos — not only saves time, but it can act as the final puzzle piece that brings a visual language together. That way, we can use AI to our benefit, and not run away from it.

We used AI-generated images in making these creative assets for Carnegie Mellon. For this campaign, we needed to tell the wildest imaginable range of stories in a meaningful way without relying on concrete artifacts. AI helped us achieve this more quickly and easily.

The Placement (Megan Hyde)

Now let’s talk about how AI is affecting ad placement and optimization.

A recent buzzword in the marketing space is “programmatic advertising.” This term, which started to rise in popularity in the mid-2000s, has been used more widely as more DSPs (demand-side platforms) have become more accessible. Programmatic advertising has made it easy to buy ads at scale, using “user profiles” created by AI, based on information like browsing behavior, location, and contextual interests.

Full transparency: we buy ads for clients like this often. It helps us reach the right people at the right time, with the right content, and it absolutely shoulders some of the load during planning and helps us stay consistent for certain audiences. On the flip side, we work very hard to make sure that those user profiles are aligned as closely as possible with our targeted audience segments (with the information available) — because if we don’t, this strategy could go rogue quickly.

We also use AI recommendations in channels like Google Search and Meta. Both of these platforms allow you to place ads dynamically. This means what you upload is not complete ads, but rather separate pieces and elements (images, headline copy, description, and so on). You can also upload multiple versions of any creative assets for testing. From there, the platform shows unlimited variations of these ads to audiences. It optimizes the mix to show the ones that resonate best and drive the most traction toward your goal — like impressions, clicks, or conversions.

These are just a few ways that machine learning and AI have changed the way we think about ad targeting, placement, and creative. Without these tools, testing and optimization would be manual and would take much longer to set up, change, and even report on.

The Takeaways

Since the 1950s, the advertising industry has found ways to automate how content is served to the masses and how to optimize that content based on how people respond — for better or worse. We’ve recently seen large-scale examples where new AI has failed us as consumers. Recently, Forbes came under fire for selling digital ad space that sent users to illegitimate URL destinations. This was missed because AI ad placements weren’t receiving proper oversight. This issue went on for almost a decade and meant billions of dollars (that’s billions, with a B) were wasted. Facebook (before Meta) had similar issues, where ads were routinely missed when AI quality assurance couldn’t apply correct context to advertisements targeting children. Both circumstances have led to an overhaul of the platforms and governance to protect people using the platform and to thwart advertisers working to abuse the system.

In both instances, AI was poorly managed and trusted blindly, which leads us to two key insights:

• AI can’t replace the empathy that’s built creatively into your marketing efforts.
• AI can’t always detect appropriate contexts without humans, which is especially important for industries like higher education, where you’re often trying to reach minors under the age of 17.

To make sure that your campaigns run smoothly and work actively to help you achieve your campaign goals, it’s critical to have the correct team or partner in place to monitor the advertising approach from start to finish — with AI included as a teammate. It’s clear that AI is here to stay.

Leaning on these tools throughout your day will likely save you time, money, and energy if you use them well, but they can create havoc, headaches, and losses if they’re used carelessly.

In our opinion, it’s best to treat AI tools like new friends. Hang out for a little bit and see if your interests line up. If they do, bring them into your inner circle as someone you can count on. If, after a little while, your interests no longer align, let them go to find their place in the world.