<div class="gn-article"><div class="gn-hero gn-reveal"> <div class="gn-hero__image"><img src="https://cdn.prod.website-files.com/687a235da6861294eec73166/6a4ce7aae06c2c679d937bba_igor-omilaev-gvqlabgvb6q-unsplash-6a4cdfb19609c339871987.jpeg" alt=""></div> <div class="gn-hero__head"> <span class="gn-kicker"><span class="dot"></span>Intelligence</span> <h1 class="gn-title">Mediocrity at Scale: The AI Creative Problem Agencies Can Now Prove They Solve</h1> <div class="gn-meta"> <strong>The GO Network</strong> <span class="pip"></span> <span>7 July 2026</span> <span class="pip"></span> <span>3 min read</span> </div> </div> </div> <div class="gn-body"> <p class="gn-lede gn-reveal">A single phrase kept surfacing across panels and press coverage at Cannes Lions 2026: mediocrity at scale. It described a specific worry among senior marketers and creatives, that as more of the industry adopts the same handful of AI models trained…</p> <h2 class="gn-reveal">The Evidence Behind the Phrase</h2> <p class="gn-reveal">The mediocrity at scale concern is not just festival mood music. A creativity benchmark study by Springboards.ai, involving nearly 700 marketing and advertising professionals judging more than 11,000 pairwise comparisons across major large language models, found that human evaluators largely could not tell the models apart. Win rates across models clustered narrowly between 50 and 55 percent, meaning the entire field of leading AI tools produces creative work of roughly indistinguishable quality.</p> <p class="gn-reveal">Separate academic research into AI-assisted creative writing found a related effect at the level of the individual writer. Studies have shown that when writers use the same generative AI assistant, their output becomes measurably more similar to other writers using that same assistant, even though each individual's work may read as more polished in isolation. Diversity falls at the population level even as competence rises at the individual level.</p> <p class="gn-reveal">Put together, the picture is consistent: AI creative tools are very good at lifting the floor and very poor at protecting the ceiling. Everyone using the same tools, trained on the same data, tends toward the same middle.</p> <div class="gn-stats gn-reveal"> <div class="gn-stat"><span class="gn-stat__num">700</span><span class="gn-stat__label">Marketing and advertising professionals in the Springboards.ai creativity benchmark study.</span></div> <div class="gn-stat"><span class="gn-stat__num">11,000<em>+</em></span><span class="gn-stat__label">Pairwise comparisons across major large language models.</span></div> <div class="gn-stat"><span class="gn-stat__num">50–55<em>%</em></span><span class="gn-stat__label">Win rates across models, clustered narrowly, indicating roughly indistinguishable quality.</span></div> </div> <div class="gn-divider gn-reveal" aria-hidden="true"></div> <h2 class="gn-reveal">Why This Matters More at Pitch Stage Than Delivery Stage</h2> <p class="gn-reveal">Brands are increasingly capable of generating competent first-draft creative in-house, using the same tools agencies use. That closes the gap on baseline competence, which used to be a large part of an agency's pitch to a client.</p> <p class="gn-reveal">What it does not close is the gap on genuine distinctiveness, the quality that separates work a category will remember from work that reads as adequate but forgettable. If every competitor in a category is running the same class of AI tools, the brand that stands out will be the one whose agency can demonstrably push past the model's default output, not the one that simply has access to the tool.</p> <aside class="gn-quote gn-reveal"><q>AI creative tools are very good at lifting the floor and very poor at protecting the ceiling. Everyone using the same tools, trained on the same data, tends toward the same middle.</q></aside> <p class="gn-reveal">That distinction is becoming a live pitch differentiator. Clients are starting to ask, implicitly or explicitly, why they should pay an agency for creative thinking that AI can approximate for free. The honest, provable answer is the one agencies have not yet built the evidence base to make.</p> <div class="gn-divider gn-reveal" aria-hidden="true"></div> <h2 class="gn-reveal">The Research Angle Nobody Has Claimed Yet</h2> <p class="gn-reveal">Almost all of the current commentary on AI creative sameness is qualitative: festival panels, opinion columns, a handful of academic studies aimed at writing rather than advertising specifically. Nobody in the agency-facing market has yet built a standing, comparative measure of creative distinctiveness that an agency can point to in a pitch room.</p> <p class="gn-reveal">That is the proprietary research opportunity. A structured distinctiveness audit, run category by category, that scores a brand's AI-assisted creative concepts against its direct competitors on measurable similarity, not subjective taste. Done properly, this becomes a diagnostic an agency can sell before a single creative concept is produced: here is how similar your category has already become, here is the distinctiveness gap available to whoever closes it first.</p> <p class="gn-reveal">The methodology does not need to be complicated to be credible. A repeatable scoring approach, applied consistently across a sector, benchmarked over time, is enough to become a reference point the market has not yet established. Whichever agency network builds and publishes that benchmark first owns the framing of the conversation.</p> <div class="gn-divider gn-reveal" aria-hidden="true"></div> <h2 class="gn-reveal">What This Means in Practice</h2> <p class="gn-reveal">Agencies leaning into AI purely for production speed are optimising for the wrong variable. Speed is now table stakes, available to any competent in-house team with the same tools.</p> <p class="gn-reveal">The commercial opportunity sits one level up: proving, with a repeatable and defensible method, where AI-assisted work is converging inside a category and where a client's current creative territory already sits inside that convergence. That proof is what justifies a premium for creative direction rather than creative production, and it is currently uncontested ground.</p> <aside class="gn-callout gn-reveal"> <div class="gn-callout__label">What this means for you</div> <h4>The distinctiveness gap is the agency's commercial opportunity.</h4> <ul> <li><strong>Speed is table stakes.</strong> Agencies leaning into AI purely for production speed are optimising for the wrong variable. Speed is now available to any competent in-house team with the same tools.</li> <li><strong>Nobody has claimed the benchmark yet.</strong> Nobody in the agency-facing market has yet built a standing, comparative measure of creative distinctiveness that an agency can point to in a pitch room.</li> <li><strong>First mover owns the framing.</strong> Whichever agency network builds and publishes that benchmark first owns the framing of the conversation.</li> <li><strong>Make it a diagnostic you can sell.</strong> Done properly, a structured distinctiveness audit becomes a diagnostic an agency can sell before a single creative concept is produced.</li> </ul> </aside> </div></div>