Going for an “AI-First” strategy is seductive. AI promises to do more with less. Plus, its makers say that this technology will change the world.

But taking the plunge and transitioning into an AI-run organization is, let’s say, a gamble.

For example, using it for design means the AI “thinks” about it in the same way, whether it’s a logo, a social media graphic, or a marketing campaign. (Which is not at all.)

This leaves you with a pile of “almosts,” “not-quites,” and design results ranging from uncanny valley to just plain wrong.

Now this might sound like we’re anti-AI, but we’re really not. We just understand that, if you want to create truly effective designs, there are certain use cases where it makes sense to bring in a human touch.

The biggest limitation of AI is that it can’t read your mind

The amount of information that AI has is gargantuan. For instance, ChatGPT is trained on trillions of parameters of data. But even with all of this information, if you delegate a creative task to AI, it won’t have the one thing it needs to execute it well out of the gate—context.

AI doesn’t know about the meeting you just had in the hallway with a colleague, or what your competitor said at an industry conference, and it won’t recognize how to translate your brand identity for unique, one-off campaigns.

This is why you get “meh” results.

AI doesn’t think. Instead, it hallucinates, takes shortcuts, and fills in context gaps with generic training data.

This results in an incredibly frustrating feedback loop that looks a little something like this:

– You write a prompt that makes sense to you. (e.g., Make two direct mail pieces to promote our latest furniture sale. They should use fall colors and have a minimalist style. Use the attached brand guidelines.)
– The AI, lacking context, spits out a generic design.
– You give it another prompt with feedback about the design. (e.g., I don’t like the way this looks. It’s too [bright/warm/too minimalist]. Try again.)
– The result you get back is slightly different, but not any better than what you got originally.

This cycle continues because AI has no idea why you don’t like it. While the original prompt said it was for an upcoming fall sale and you provided a set of brand guidelines for the AI to reference, you didn’t specify who the direct mail pieces were for or the business problem you were solving.

It creates an annoying loop of multiple inputs, leading to slightly different, yet still wrong outputs. This is frustrating, yes, but there’s a fix.

Smarter delegation with AI and humans

We aren’t telling you to abandon AI outright. We’re suggesting you match the task to the right tool or person, since every design piece isn’t equal in terms of execution, time, or strategy.

Some tasks are perfect for AI. Meanwhile, other projects would benefit from having a person managing them alongside AI. But the most valuable marketing pieces will require human thinking and implementation.

So how do you decide which one you need for those creative tasks? We recommend using these tiers to determine when to use AI and when to bring in humans.

Tier 1: AI-Only

These are the quick, usually production tasks, you can knock out with AI using simple prompting and minimal direction. You’re giving it straightforward inputs like “Replace the sunrise background and ocean in this image with mountains at sunset.”

Example use cases: Resizing large batches of images, removing backgrounds, and converting file types.

What do you get out of it? Saving money and time. It handles high-volume production projects, freeing up your (and your team’s) bandwidth for more important work.

Tier 2: AI + Human

Think of this as using AI as a co-pilot for jobs that would normally take hours, days, if not, weeks to complete. Because AI is doing the heavy lifting for you, it cuts this time down considerably.

But more importantly, you’re giving direction, refining, and making the final call on the design piece. AI creates the foundational layout, and you give it a good polish to make it ready for prime time.

Example use cases: Creating blog or ad banners, generating first-draft icon designs, and building product use imagery.

What do you get out of it? Again, time-savings. Using AI as your co-pilot can potentially get you a design result that’s 80% there, which leaves you or a designer with 20% to complete.

Tier 3: Human-Only

These are the designs that you can’t afford to let AI create because they are invaluable to your business. For these particular projects, you need humans involved from start to finish. These are the experts who will understand how to interpret the nuances and subtleties in creative direction that AI cannot.

Example use cases: Designing your brand identity, website design, or marketing campaign strategies.

What do you get out of it? Value creation for your company, brand protection, and human-led strategies that are unique, audience-focused, and based on specific, achievable objectives.

Choosing the right tool (and person) for the job

When used as a co-pilot, AI can help save time you would otherwise spend on production and smaller design tasks. But it’s important to understand where the technology succeeds and where it falls short.

By using AI for what it’s good at (production), you can put more time and energy into what you and your creative team are good at. You know, strategy.