Ad creative used to mean a brief, a production timeline measured in weeks, and a team of people between an idea and a finished asset. Marketing technology vendors are compressing that entire chain into something closer to a single prompt, and the companies proving it out are not ad agencies. They are customer data platforms.
The shift: CDPs are becoming ad production tools
Hightouch, a customer data platform, launched Ad Studio in mid April as part of its broader agentic marketing suite. The tool is built around what Hightouch calls a Brand Context Layer: it connects to a company’s data warehouse and ad platforms to see which creative has actually performed, and it connects to a brand’s creative stack, including Figma, DAMs, and Google Drive, to learn a brand’s visual identity from real assets rather than a style guide. From there, Ad Studio runs a three stage workflow: it surfaces a performance signal, generates ad concepts in multiple sizes and languages, and exports the winners straight to ad platforms.
That is the pitch. The proof point is a European fashion marketplace called Otrium, which Hightouch says cut campaign production time by 70%, from four weeks to one, while lifting click through rate by 13% and conversions by 15%, alongside roughly a 10% gain in return on ad spend.
The stunt that made it visible
In June, Hightouch funded a one day Times Square activation to show the tool working in public. It featured ads built with Ad Studio for three of its clients: Sweetgreen, Tripadvisor, and HelloFresh. HelloFresh’s ad leaned on a simple contrast, the cost of takeout versus the cost of a home cooked HelloFresh meal, to make its value case.
Conor Feeney, HelloFresh’s US head of marketing, told AdExchanger that brands are “under pressure to produce significantly more creative than ever before” as attention spans shrink and channels multiply. HelloFresh adopted Ad Studio less to replace its creative team than to run more variations faster. Feeney was explicit that the tool doesn’t “replace the creative thinkers on the team.” What it changes is the iteration loop: marketers can prompt the system in plain language to make copy punchier, shift emphasis from a discount to product quality, or retarget a line to a different audience, and see a new draft immediately instead of waiting on a production cycle.
Hightouch co-CEO Tejas Manohar made a similar point about the skill the tool actually demands. Using Ad Studio requires no coding, he told AdExchanger, but it does require “a talented, creative person” who can articulate why a piece of creative fits a brand and translate that judgment into instructions the AI can follow. That is a meaningful distinction from the “AI replaces the creative department” framing that tends to dominate the debate: the winning workflow still routes through a human who understands the brand, it just removes the weeks of production between that person’s judgment and a finished, deployable ad.
Why this is bigger than one vendor’s product launch
The interesting part is who is building this. A customer data platform entering ad creative is a sign of where CDPs think their value is heading. A CDP already owns the two hardest inputs to good creative at scale: a live read on which messages and audiences are converting, and the pipes into ad platforms to act on that data. Adding a generation layer on top turns a data infrastructure product into a full performance marketing loop, insight, creative, and deployment, without a marketer needing to leave the platform or brief an outside agency for every variation.
That also explains why Hightouch was willing to fund a public stunt for a tool that, as reported, does not yet touch media buying or budget allocation. Distribution of proof, not the underlying tech, is the harder problem for any AI creative tool right now. A billboard in Times Square is a way to make an abstract “agentic workflow” claim tangible to marketers who have heard a version of this pitch from a dozen vendors this year.
It also lands alongside a broader push by marketing platforms to build guardrails around what their AI agents are allowed to do, and a growing body of early performance data on agentic ad creative’s real-world return. Both point the same direction: the category is moving past demos and into vendors publishing numbers they are willing to be held to.
What it means for the marketing leader
For a marketing leader evaluating this category, three things matter more than the demo:
First, ask where the brand context actually comes from. Ad Studio’s differentiator is pulling from a company’s real product catalog, compliance guidelines, and past ads rather than a written brand book, which is the detail most likely to separate a genuinely on brand output from generic AI ad slop. Second, ask who owns revision. HelloFresh’s account of the workflow keeps a human reviewing and directing every draft; any vendor pitch that implies zero human review deserves more scrutiny, not less. Third, treat a single flashy activation as a pilot, not proof. HelloFresh has, by its own account, only shipped one ad built this way so far. The Otrium numbers are real and worth weighing, but one mid market fashion retailer’s results will not automatically transfer to every category or team structure.
The broader signal for the function is that creative production is quietly becoming infrastructure, the same way data pipelines and customer data platforms did a decade ago. Marketing teams that treat AI creative tools purely as a cost cut on agency fees will miss the more useful framing: this is a speed advantage in testing, and speed to test is what compounds into a performance advantage over a quarter, not a single campaign.
What to do next
Marketing leaders evaluating agentic creative tools should pilot on a single, well measured channel first, insist on seeing the brand context sources a vendor actually ingests rather than taking “on brand” as a marketing claim, and set an explicit review step for every AI generated draft before it ships. The vendors racing into this space, CDPs among them, are betting that whoever owns the fastest loop from performance signal to live creative wins the next round of ad tech consolidation. That bet is still early, but it is no longer theoretical.
Source: Hightouch
