Agentic AI has spent a year auditioning for a job in advertising creative: write the script, cut the video, swap the assets, do it in hours instead of months. What has been missing is proof it actually moves a business result. That proof arrived quietly in the form of a Cuisinart food processor video.
Conair Corporation, the consumer products company that owns Cuisinart, ran a controlled test pitting a video from Amazon Ads’ agentic tool, Creative Agent, against a traditionally produced version. The AI-made video won on both engagement and cost, and it did so on a timeline traditional production cannot match. That single test is now the clearest public signal yet that agentic creative tools are moving from beta demos into the part of the marketing budget that actually gets measured.
What Creative Agent actually does
Amazon Ads launched Creative Agent as an AI-powered creative partner built into Creative Studio, its ad-production tool. Built on AWS technology using Amazon Bedrock foundation models, including Amazon Nova and Anthropic’s Claude, the tool takes an advertiser through research, concept brainstorming, storyboarding, and final production through a conversational interface. It can generate multi-scene video with animation, voiceovers, and music, plus matching display formats, while giving the advertiser editing control at each stage.
“This is about more than speed, it’s about giving every advertiser and agency access to the kind of strategic, high-quality creative support that once only large brands could afford,” said Jay Richman, vice president of Amazon Ads, in the company’s announcement. The tool is currently in beta and available in Canada, China, France, Germany, Italy, Spain, the United Kingdom, and the United States, at no additional cost to advertisers. Amazon Ads says professional creative development that once required specialized design skills, and could run into the tens of thousands of dollars per project, is now reachable by advertisers of any size through a conversational interface, with granular editing control preserved at every stage of production.
The Conair test, by the numbers
Conair’s e-commerce team, working with agency Global Overview, used Creative Agent to produce a 15-second video promoting the Cuisinart food processor and ran it head-to-head against a conventionally shot version over a 30-day window in April and May. The AI-assisted video delivered 18% higher detailed page views and a 14% reduction in cost per detail page view compared with the traditional version.
The speed gap was just as significant as the performance gap. Amazon’s tool produced the finished video in roughly four weeks, against the three to six months a comparable video typically takes to move from brief to delivery. Creative Agent generated an estimated 80% of the finished video; the remaining 20% still required human correction, mostly fixing logos, brand colors, and on-screen text that the AI got wrong.
“We’re moving faster than some of our peers on this,” said Justin Swenson, Conair’s senior vice president of e-commerce. “There is definitely something pretty compelling here.” Kelsey Smithuysen, Conair’s director of Amazon advertising, said the AI-made version connected better with e-commerce shoppers, largely because it emphasized ease of use in ways the brand-produced version did not prioritize.
What this means for the marketing leader
The Conair result reframes the agentic-creative conversation away from novelty and toward a budget line marketing leaders already track: cost per outcome. An 18% lift in detailed page views paired with a 14% cost reduction is a number a media planner can put in a business case, which is a different conversation than “the AI made a decent video.”
It also resets expectations about labor, not headcount. The 80/20 split between machine output and human correction suggests the near-term role for creative teams is quality control and brand-standard enforcement, not raw production. That is a real shift in job function for in-house creative and agency teams alike, and it changes what marketing leaders should be hiring and training for over the next budget cycle. Teams built around briefing external production vendors will need to rebuild around briefing and correcting an AI system instead, a different skill set than the one most in-house creative teams currently hire for.
The timeline compression matters for a different reason: it changes what is possible to test. A four-week creative cycle instead of a four-month one means brands can run more creative variants per quarter, against more SKUs, in more markets, without a proportional increase in production budget. For marketers managing large catalogs, especially in retail and e-commerce, that is the more durable advantage.
What to watch next
One case study is not a category-wide verdict. Amazon has an obvious incentive to publicize a favorable result for its own tool, and the correction rate on brand elements is a real quality-control cost that offsets some of the speed gain. Marketing leaders evaluating agentic creative tools, from Amazon’s Creative Agent to competing offerings, should ask vendors for the same two numbers Conair now has: cost per outcome versus a traditional benchmark, and the actual human-hours needed to bring AI output to brand standard. Those two figures, not demo reels, are what will separate durable tools from beta hype over the next few quarters.
Amazon’s own ad infrastructure is expanding in parallel with this creative push. The company recently extended Amazon Retail Ad Service internationally through a deployment with Asda, and it has been layering agentic capability across its ad stack more broadly, including the agentic ads it embedded into Alexa+. Creative production is simply the latest link in that chain, and the marketing leaders who benefit most will be the ones who treat it as a measurable channel, not a demo.
Source: Amazon Ads
