The Advertising Paradox
The one thing you need to know in AI today | AI Ready CMO
We’re wrapping up 2025, and over the next few days, we’ll be sharing key insights from our year-end report: AI in Marketing 2025 Annual Report.
The report will be available for all subscribers to download in full for free from the 22nd of December.
AI has created an advertising paradox.
The advertising data from 2025 presented a contradiction: AI-generated creative consistently outperformed human-made work in testing, yet organizations fully automating production without human oversight reported declining effectiveness over time.
Immediate gains were real. AI-driven ads delivered measurably higher conversion rates. Companies using AI for optimization reported substantially higher ROI. Meta’s tools improved performance across hundreds of thousands of ad variations.
AI tools achieved dramatically higher accuracy predicting creative success versus human judgment. By mid-2025, the vast majority of advertisers planned to use generative AI in video strategies, with projections showing AI video could dominate ad inventory by 2026.
Marketers reported creating content faster, with meaningful boosts in awareness and sales. JP Morgan Chase partnered with Persado and saw a 450% increase in ad CTR. Google reported significantly higher ROAS from AI-powered video campaigns.
Yet beneath performance metrics, a strategic problem emerged.
When everyone possessed the same tools and accessed the same training data, individual campaign improvements didn’t translate to sustained competitive advantages.
The Sameness Problem Nobody Expected
The darker reality revealed itself through “creative convergence”—systematic homogenization as AI tools trained on similar datasets produced statistically similar outputs.
By 2025, this reached crisis levels. User-generated content ads—originally valued for authentic appeal—became so formulaic they appeared less authentic than polished brand content. With millions of active stores competing on the same platforms, brands made ads more similar rather than differentiated. Generative AI produced “statistically most likely” versions, which by definition meant generic outputs.
Trust implications extended beyond aesthetics. Only 13% of consumers trusted ads created entirely by AI, while 48% trusted ads co-created by humans with AI support. Research showed people valued objects more when they believed those objects carried traces of human intention.
Toys ‘R’ Us used OpenAI’s Sora to create a brand film that drew immediate criticism for its “cold, uncanny tone.” The principle became uncomfortable: AI tools could predict performance with exceptional accuracy yet paradoxically increased the strategic importance of distinctive creative judgment.
When everyone optimizes for the same algorithm using the same tools trained on the same data, you get convergence. Performance improves. Differentiation collapses. Short-term metrics rise. Long-term brand value erodes.
Why Taste Became Infrastructure
The strategic response wasn’t technological—it was organizational. By late 2025, the competitive differentiator wasn’t access to AI tools (universally available) but systematic cultivation of “taste infrastructure”—organizational capacity to recognize quality and consistently produce distinctly human work.
Human oversight wasn’t optional—it was operational necessity. Without human review, AI generated technically correct content that was off-brand, inappropriate, or subtly wrong in ways testing couldn’t catch until market reception revealed the problem.
Organizations treating AI as creative teammate rather than autonomous replacement achieved superior results. While the majority created content faster, only a minority found AI-generated content more successful than manual work when measuring long-term brand impact versus short-term conversion rates.
Consumer expectations clarified boundaries. While most marketers considered AI critically important for marketing success, consumers demanded transparency and co-creation. The winning formula wasn’t human or machine—it was human and machine in clear partnership.
What This Means for Creative Teams
Your performance marketing team sees AI delivering better CTRs and thinks the creative problem is solved. Your brand team sees creative convergence and knows the differentiation problem just got worse.
Both are right. That’s the paradox.
AI makes optimization trivial. Every brand can now A/B test thousands of variations, predict performance with high accuracy, and automatically allocate budget to winning creative. Performance marketing became democratized—which means it stopped being a competitive advantage.
When everyone can optimize equally well, optimization stops mattering as a differentiator. The new competitive moat is the thing AI can’t do: develop distinctive creative judgment, recognize what makes your brand uniquely yours, understand cultural context that statistical models miss.
The brands entering 2026 treating AI as a creative replacement will keep improving their CTRs while their brand equity slowly erodes.
The brands treating AI as a production accelerator under human creative direction will improve performance while maintaining differentiation.
Your creative director wondering if their job still matters? Their job matters more than ever. Just not for the reasons it mattered in 2020.
The Trust Problem
Consumers don’t trust fully AI-created ads (yet). That’s not a temporary perception problem. That’s a fundamental signal about what audiences value.
People don’t trust AI ads because they correctly perceive those ads lack human intention. The ads optimize for conversion. They don’t attempt to communicate genuine perspective or build emotional connection. They’re statistically likely to perform well and existentially hollow.
You can see this in the Toys ‘R’ Us Sora film. Technically impressive. Strategically terrible. It predicted well in testing and failed catastrophically in market because testing measures response, not meaning.
Your challenge in 2026 isn’t choosing between AI and humans. Your challenge is building organizational capacity to know when to use which, how to combine them effectively, and how to maintain brand distinctiveness while everyone else converges toward algorithmic sameness.
Performance without differentiation is a race to the bottom. AI solved performance. You still have to solve differentiation.
—
We’re publishing our complete AI in Marketing 2025 Report on December 22nd.10 chapters analyzing what actually happened this year—including why most marketing automation initiatives failed and what the 6% who succeeded actually built differently.
Stay tuned. You’ll want to see what you missed.
3 AI Marketing Tools To Try Today
PDF AI
Extract insights from documents instantly—PDF AI reads, summarizes, and answers questions about contracts, reports, and research papers without manual scanning.
TextCortex
Write anything faster with AI that adapts to your style—TextCortex generates emails, documents, and social posts while learning your voice and preferences over time.
MeetAlfred
Automate LinkedIn outreach and multi-channel campaigns—MeetAlfred manages connection requests, follow-ups, and email sequences across platforms, turning cold prospects into warm leads without manual work.