If you work with AI regularly, you might find yourself in a sentiment silo. You see the benefits and challenges of AI firsthand, but your audience often has a very different perspective.
While most experts see AI as a benefit, the general public is more likely to view it as harmful or feel unsure about how it will impact them at all.
That disconnect matters. Success doesn’t come from embracing AI blindly, but from using it thoughtfully. The key is to use transparent, explainable tools that enhance your advertising without alienating your audience.
This guide explores the distinction between effective and disingenuous AI ad campaigns. What are the risks? What does a better approach look like? And how can you use AI as an assistant, not a replacement, to deliver campaigns your audience will actually appreciate?
You might remember Coca-Cola’s 2024 AI-generated holiday ad. Known for experimenting with new tools and formats, Coca-Cola took a bold step, but this time, it didn’t land. Viewers criticized the ad’s composition, scale, pacing, and emotional resonance, calling out the awkwardness caused by AI video generation’s current limitations.
You may also remember the notorious clash between expectation and reality in Willy’s Chocolate Experience in February 2024. The promotion included creative with barely intelligible text crafted by AI. It was foreshadowing for a poorly executed event:
In both cases, the common thread was clear: AI wasn’t used to support human creativity. It was doing the heavy lifting. And when AI’s presence is too obvious, especially in emotional or artistic content, audiences notice. It feels impersonal, low-effort, and off-brand.
A simple rule of thumb: if the AI-generated element is obvious to viewers and doesn’t meet the brand’s expected quality, it is likely to feel disingenuous and may do more harm than good.
Coca-Cola will recover. But most brands don’t have that kind of resilience. If you lean too heavily on AI in your campaigns, you risk:
Irrelevant or inappropriate placements: Without proper oversight, AI might optimize for short-term performance, placing ads in low-quality or off-brand environments that hurt credibility.
Subpar creative: AI can generate visual assets quickly but often lacks brand nuance. Many creative tools are trained on examples from thousands of brands. If your ads feel generic, or your audience can tell they’re AI-generated, you risk damaging trust.
Over-automation: Fully autonomous campaigns might chase surface-level metrics like clicks but miss context and long-term strategy. This black box problem, where AI makes choices without explanation, leaves marketers in the dark.
Tone-deaf personalization: AI can personalize at scale, but without human review, messages can come off as intrusive or misaligned. Think of ads referencing a one-off purchase or poor timing during sensitive moments.
Reputational risk: Your audience is more AI-savvy than you might think. Passing off AI-generated work as entirely human-made is a gamble that can quickly erode brand trust.
The risks are real, but they don’t outweigh the rewards. Many prominent brands are already using AI successfully and thoughtfully.
Heinz famously produced an AI campaign that transparently plays with the nature of AI. Heinz asked AI image generators to "draw ketchup," and AI consistently created images resembling Heinz's classic product package.
This clever campaign demonstrated how iconic Heinz is — even AI recognizes their signature bottle and logo. It was a creative use of AI that reinforced brand recognition rather than fabricating it:
AI creative doesn’t have to be the final product, either. Nike, often seen as much a technology company as an apparel brand, used Midjourney to generate thousands of initial images for their video ad.
These AI-generated visuals served as a storyboard for the ad, with human designers layering in their edits. And most importantly, Nike was transparent about using AI.
This hybrid approach allowed for massive creative scale while maintaining a human touch. AI provided a baseline, and humans took it to the next level while following Nike’s brand guidelines:
Check out how my friend Jak Bannon and Gibson Hazard used AI to help visualize one of the dopest Travis Scott Nike commercials ever made!! 2025 is our year. https://t.co/W0fbLqQpED
— Dave Clark (@Diesol) December 26, 2024
These examples show how brands can use AI to enhance creative, not replace it. That is the sweet spot: creative scale and efficiency with human oversight and storytelling.
The same principle applies beyond creative. When thoughtfully applied across the full programmatic campaign lifecycle, AI can elevate performance without sacrificing control.
When you use AI well, the benefits compound:
Better efficiency: AI automates tedious tasks, freeing up human time. This includes automating bidding strategies, quick audience segmentation, and real-time budget allocation, ensuring resource allocation is optimized.
Smarter targeting: AI can process and segment large datasets in real time. What might take a team days, AI can handle in seconds.
Creative scale: Use AI to adapt assets across formats, brainstorm new concepts, and scale production of creative. This is especially useful for iterative testing.
Insightful decisions: AI can surface opportunities for optimization and explains the rationale behind suggestions, turning data into actionable intelligence.
Brand safety: When trained on your voice and visual identity, AI becomes a second set of eyes, helping you QA content and maintain brand integrity at scale.
We think we’ve developed a good formula at AdRoll for implementing AI in ad campaigns.
These are the three steps we use to work with AI tools, but limit their full control. Our philosophy is that AI should work alongside you, augmenting your intuition and streamlining your process, not replacing your input.
Use AI tools to experiment against low-stakes goals and see how they perform. Many will be capable of minor changes and optimizations with your oversight, but the key is to begin with a goal in mind.
For instance, if you run programmatic ad campaigns, AdRoll's platform provides AI tools that are perfect for initial experimentation:
Performance optimization suggestions: Our AI analyzes campaign data and provides actionable suggestions for better performance. These aren't black-box decisions; they're informed recommendations designed to improve your outcomes.
Audience targeting suggestions: AI helps identify and refine audience segments through contextual keyword recommendations, ensuring your campaigns reach the most receptive consumers.
Creative streamlining tools: AdRoll's AI Ad Builder supports creative development in ways that enhance, rather than replace, humans:
CTA and headline suggestions: Our AI can pull relevant calls to action and headlines directly from your website, providing tailored suggestions for your ads. This saves time across multiple ad variations.
Image expansion: Have a great lifestyle image but need it to fit multiple ad sizes? Our tool can help expand and apply that image across an entire set of ad formats, dramatically cutting down on replication time while maintaining visual integrity.
These features allow you to use AI for ad optimization and campaign setup support, rather than relying on it for net-new, potentially generic creative that alienates your audience.
A common concern with AI is the lack of transparency in its “thought process.” How does it know what it knows? How is it making decisions that impact your campaigns?
Our AI addresses this by providing clear rationale for its suggestions:
When our AI detects an opportunity for improvement in your campaign (based on your objectives), you'll see a sparkle symbol next to relevant metrics in your dashboard.
Hovering over this symbol prompts an AI recommendation. You can then engage with our AI assistant in natural language to explore the rationale behind the suggestion. This conversational approach provides the "why" behind the "what," giving you the confidence to implement or refine the recommendations.
Beyond optimization suggestions, our AI also offers prompts for missing setup steps or suggested campaigns based on your goals.
This is particularly effective: self-serve users who set up campaigns with the support of AdRoll's AI assistant are 50% more likely to launch campaigns with all recommended and standard ad sizes.
Contrast this with "black box" automation found in some ad platforms, where you might see performance gains but can’t determine why, making it difficult to replicate AI success. That’s hard to explain to stakeholders!
AI can actually be a good way to validate your ideas. One of the most obvious ways to use AI for idea validation is to ask an LLM questions for better context or to review your ideas.
Another method is to train external AI models (like custom GPTs) to understand your brand's guidelines. You can then use these models to "test" or "QA" human creative or even other AI-generated content against your brand standards.
Or, as another option: Within the AdRoll platform, our campaign creation assistant guides you through proper setup. Even if you're setting up a campaign by yourself, you can ask the AI assistant questions to double-check that everything is correctly implemented. This pre-launch QA helps prevent common human errors.
Ultimately, avoiding a "bad" AI campaign comes down to keeping humans in the loop. AI is moving mountains for advertisers that are using it properly, but it thrives when guided by humans, who have a deep understanding of your brand's unique identity and audience.
As AI continues to evolve, the focus will shift from "can it do this?" to "how can it do this better with human partnership?" The future of effective advertising will always rely on collaboration between humans and AI.
Last updated on August 20th, 2025.
