As AI becomes increasingly embedded in the process of optimising ad campaigns, marketers have expressed concerns that their competitors stand to benefit from the success of their campaigns, by using the same platforms that learn from the performance of their ads. For streamr.AI, a startup whose tech allows SMEs to create AI-generated video ads and launch them on streaming TV, that’s sort of the point.
“If you see a really well performing Meta ad and you want to prompt GenAI to make a similar ad that was performs well on another platform, we’re going to let you do that,” streamr.AI CEO Jonathan Moffie tells VideoWeek. “I think that’s great for competition. You can’t stop it, and I think you just lose if you do.”
Today the company has announced the integration of Google’s Veo 3 AI video generator into its creative engine, enabling the platform to generate 1080p video ads with voiceovers, music and brand customisation. The Google model creates eight-second video chunks, which streamr stitches together into a 15- or 30-second ad. Its AI agent then makes recommendations around budget and targeting, and the customer can enter their credit card details and launch the campaign.
Moffie notes that using Google Veo 3 directly, users would need to enter prompts themselves in order to generate video. But using streamr, customers can enter their business name or website URL, and the ad is generated. They can then edit the imagery, calls to action, script, music or template, translate the ad into over 32 different languages, and pull in social assets.
“A lot of D2C companies, like e-commerce brands, really like taking a viral Instagram reel and then launching that on TV,” adds Moffie. “So we’ve made it super easy to pull in your whole Instagram library, and then you can actually make a CTV commercial from Instagram.”
“Zuck stole my thunder”
The company launched just over a year ago, with a business model that involves white labelling its tech to CTV publishers, broadcasters, ad tech firms and agencies, using its streamr.ai platform as a “storefront for the Magnites and TripleLifts of the world to see if they want to explore white labelling us or not.” The startup now has 13 white label deals in place, with the Veo 3 model being used by advertising nonprofit AdGood, as well as an unnamed demand-side platform (DSP) and a UK broadcaster, enabling SMEs on their platform to generate ads using Veo 3.
And while the end goal looks similar to Meta’s plans to automate the entire advertising process using AI – an announcement that has raised questions over the role of agencies in an AI-driven future – streamr’s business model involves working alongside agencies, at least for the timebeing.
“Zuck stole my thunder with that announcement,” jokes Moffie. “I wrote a post about our blueprint for building the AGI (Artificial General Intelligence)-powered ads manager back in January. So I’ve got receipts!”
The company ultimately aims to build a platform that “doesn’t just assist with creating ads but completely owns the process from start to finish”, running potentially infinite versions of ads to every demographic and behavioural audience, while constantly monitoring and optimising campaigns in real-time.
“I do believe it can get to this point,” comments Moffie. “I think almost the entire economy can get to this point, if you really look at the rapid pace and innovation in AI. But in the meantime, what we’re doing is just reducing friction drastically for agency teams.”
He says agencies currently use streamr as a way to generate leads by generating preview TV ads for prospects, alongside campaign recommendations, in order to start conversations with new-to-TV brands.
“We’re like the SME top-of-funnel, basically a new advertiser acquisition tool,” says Moffie. “Then they can actually upsell them once they have the conversation started. So I think that’s a good path for agencies – but it’s going to be bumpy, no doubt about it.”
GenAI goes mainstream
Meanwhile industry participants appear to be catching up with that rapid rate of change. Last week, WPP Media released its ‘This Year Next Year’ report, which shed light on how quickly AI is predicted to become the norm in advertising. Back in 2020, less than 40 percent of respondents thought it was likely that companies would rely on AI to produce the majority of creative content by 2030. Now that figure has risen to over 70 percent.
On the flipside, a report from the World Federation of Advertisers last year found that 80 percent of marketers are concerned about their agency partners using GenAI on their behalf. And while larger advertisers are particularly concerned about the potential risk to brand reputation, Moffie notes that “SMEs are just so excited to be able to get on TV for the first time.” He uses the example of a coffee shop client who was so intent on running during the ‘March Madness’ college basketball tournament that they were prepared to blow their entire $1,500 budget in a matter of hours.
But the company also works with larger advertisers, and Moffie says using a brand’s assets to generate video (for example using its image-to-video model) helps safeguard against low-quality output or hallucinations – as opposed to the surrealist imagery that can come from prompting GenAI.
“There’s been some cool stuff on X and on LinkedIn, where people are leaning into the more surreal, looks-like-GenAI aspects of it, but we’re really into making it look like your business,” he comments. “We’ve already worked with some pretty massive brands, and they don’t care as long as it looks like their business.”
And with the addition of Veo 3, that output is “only getting better” in terms of speed and quality. “I’m not naive, it’s not winning a Cannes Lions award – this year,” says Moffie. “But do I believe it’s going to get there? Totally.”
Follow VideoWeek on LinkedIn.
