When AI absorbs the creative cycle
Last week, Anthropic launched Claude Design, an unexpected move that triggered a 7% drop in Figma's stock and shook the creative industry. A single product compressed the cycle that sustained the creative industry for two decades: brief, design, prototype, production, delivery. The new Claude Opus model 4.7 generates designs,…

Last week, Anthropic launched Claude Design, an unexpected move that triggered a 7% drop in Figma's shares and shook the creative industry. A single product compressed the cycle that sustained the creative industry for two decades: brief, design, prototype, production, delivery. The new Claude Opus model 4.7 generates designs, slides and one-pagers with direct export to Canva, PDF, PPTX and HTML, plus automatic handoff to code.
The path that used to take weeks across a strategy team, an art director, a UI designer and a front-end developer now fits into a conversation with a language model. Figma understood it before the market did: the stock drop reflected how quickly investors recalculated the value of a collaborative tool when collaboration itself is automated.
But Claude Design is a symptom, not the cause. The structural data point came from Paris: Publicis reported in its Q1 that 86% of its revenue already comes from services powered by artificial intelligence. Eighty-six percent. The oldest advertising holding company in France bills almost nine out of every ten dollars through a stack where AI touches planning, creative, targeting and measurement. Anyone who reads this as a corporate vanity metric is making the mistake Latin American agencies have been making for a decade: confusing an early signal with distant noise.
The traditional creative cycle operated as a linear value chain with a margin at every link. The agency charged for strategy, the production house for execution, media buying for distribution, the analyst for measurement. AI compresses that chain until it is flat. Claude Design eliminates the design-production link. OpenAI's new ads manager — with an interface similar to Google Ads and a ChatGPT Ads pilot starting at US$50,000 — eliminates the intermediated distribution link. And conventional measurement is collapsing too: traditional marketing mix modeling can no longer cope with media fragmentation or inconsistent data across platforms; AI-assisted models incorporate broader signals and deliver actionable readings in shorter cycles.
The shift can be measured in market data. YouTube accounts for 29% of social citations in Google's AI Overviews and projects more than US$10,500 million in advertising revenue for 2026. Brand visibility is migrating from the feed to the generative result: anyone who wants to be cited by ChatGPT or Perplexity needs long-form video content with substantive narrative, because models prioritize sources with verifiable depth over fragments optimized for classic SEO. Positioning is no longer bought with media spend alone; it is built with content density that AI models consider worth citing.
And here is the figure that should keep any brand director awake at night: 49% of AI users say they would try an alternative brand suggested by a chatbot. Brand loyalty, that intangible asset companies cultivate for years with million-dollar budgets, gets redistributed in a thirty-second conversation with a model that optimizes for contextual relevance. Only 46% fully trust the recommendations and 89% verify before buying, but the consideration window has opened. Every query to an AI assistant is an opportunity for an unknown competitor to enter the consumer's consideration set.
For the Dominican and Latin American creative industry, this compression reproduces a familiar pattern. Local agencies that operate as executors of imported briefs — adapting global campaigns, translating copy, adjusting formats — occupy exactly the vulnerable position. They are digital braceros of the creative cycle: they supply skilled labor at low cost without capturing the strategic value of the chain.
When AI automates execution, the margin on local adaptation disappears. Publicis invested early in its own AI stack and today it captures the prize. Regional agencies that wait for the tools to arrive packaged as a monthly subscription will reach the market with no differentiation, competing for the leftovers of a model that has already mutated.
Corporateamnesia® makes the problem worse. Agencies that keep decades of knowledge about brands, audiences and local markets in WhatsApp chains and unstructured Dropbox folders will lose that capital when AI redefines workflows. Undocumented knowledge is knowledge that automation erases without warning.
The window for the region's agencies and creative teams has a precise coordinate: build in-house capacity for AI applied to the creative chain, document accumulated knowledge as a trainable asset, and move the business model from execution by the hour to the delivery of measurable results.
Publicis has already crossed that line. OpenAI already has its own advertising channel. Claude already designs.
The Latin American creative industry now faces a choice: sit down as an operator of these systems, or wait for the next cycle of digital compression to eliminate the link in the value chain it still occupies.