Magnite, a company that helps websites and streaming apps sell video ads, worked with a French ad-buying agency called Amnet to let a computer program handle most of the setup work for a car brand’s streaming TV campaign. The test matters because it shows what happens when software, instead of a person, picks the ad slots and settings, and Magnite says the campaign got built about 70 percent faster with a high rate of viewers watching to the end. Media buyers and planners now have one more real example, rather than a hypothetical one, of what handing tasks to an AI agent actually looks like in practice.
Magnite this week announced the launch of what it describes as its first agentic campaign in the EMEA region, carried out in partnership with the French programmatic trading desk Amnet France. The announcement, dated September 8, 2026, and published from Paris, credits an approximate 70 percent reduction in campaign setup time and a video view-through rate of 95 to a connected television buy built for an unnamed automotive manufacturer.
Magnite (NASDAQ: MGNI) describes itself as the world’s largest independentย sell-sideย advertising company. According to Magnite, the campaign used natural language prompts to let Amnet’s team direct a buyer agent that built and activated a video campaign through Magnite’s ClearLine platform. The company frames the exercise as evidence that agentic tools can shift media buying teams away from manual configuration work and toward strategic decisions.
What Magnite and Amnet built
According to the announcement, Amnet used natural language instructions to operate Magnite’s buyer agent, which in turn built and activated a video campaign through ClearLine, Magnite’s curation and activation platform, on behalf of a car manufacturer that the release does not name. The mechanism connecting the two sides is what Magnite calls Orchestration: a system in which the Magnite buyer agent communicates directly with a corresponding Magnite seller…