The San Francisco startup emerges from stealth with Mayfield backing and a pitch that treats ad creative as a continuous learning loop, not a quarterly deliverable.
Every B2B marketing team knows the problem. A campaign launches, the creative is fresh, the targeting feels right, and then, slowly, it starts dying. Audiences tune out. Click rates fall.
The agency comes back for a creative refresh and the cycle begins again. Matt Jayson calls this “decaying ads,” and it is, by his account, a structural failure of how digital advertising is built: campaigns that start losing effectiveness the moment they go live, because the feedback loop between what customers actually say and what the ads actually say is too slow.
On Wednesday, the startup Multiply emerged from stealth with $9.5 million in funding to tackle that problem. The round was led by Mayfield, with participation from Sorenson Capital, Instacart co-founder Max Mullen, and Josh Woodward, Google’s VP of Labs and Gemini, the executive credited with building NotebookLM and overseeing Google’s flagship AI app.
Executives from HubSpot, Braze, Brex, Sierra, and Common Room also joined the round.
Multiply’s pitch is that modern B2B companies are already sitting on the data they need to run far better advertising, they just aren’t using it. Sales call recordings, CRM pipelines, and closed-won deal data contain precise information about why customers actually buy.
Multiply’s system plugs directly into those sources and uses a suite of AI agents to translate them into continuously improving ad campaigns on Google Search and LinkedIn.
Hundreds of structured experiments run in parallel each week, testing messaging, audiences, and creative, with winners scaled and losers cut automatically.
The agent architecture breaks down into five components. A Customer Insights Agent extracts language from sales calls to personalise ad copy. An ICP Agent analyses closed-won deals to tighten audience targeting.
A Quality Score Agent tunes keyword alignment and copy for Google’s ranking signals. A Creative Design Agent refreshes imagery on a weekly cycle. An A/B Testing Agent runs the experiments and identifies what’s working.
Human media buyers sit above all of it, providing brand oversight and compliance review, the “hybrid” in what Multiply describes as a hybrid AI-plus-human agency model.
Jayson, who previously worked at Google in user acquisition and then at Brex as Head of Product for core experiences, describes the gap the company is trying to close: the insights that land deals, the specific objections, the competitor comparisons, the language that actually resonates, rarely make their way back into ad campaigns quickly enough.
His co-founder and CTO, Ashish Warty, spent five years as SVP of Product and Engineering at HackerOne and held senior engineering roles at Airship and Dropbox.
“Modern companies already have all the data needed to create radically better ads,” Jayson said in a statement. “Sales conversations, CRM systems, and pipeline outcomes reveal exactly why customers buy, yet those insights rarely make their way into ad campaigns fast enough.”
The timing is deliberate in another sense. Multiply’s infrastructure is, the company says, already being positioned for ChatGPT advertising, a format that OpenAI has signalled it intends to launch but has not yet released at scale.
The argument is that the same campaign learning systems built for search and social can extend into conversational and AI-driven ad formats as they emerge. That is a forward-looking claim that will depend entirely on how those platforms eventually structure their ad products.
“There is a major shift happening in the $50 billion B2B advertising market,” said Patrick Salyer, Partner at Mayfield and a Multiply board member, in a statement. “Service-as-Software is redefining how companies grow, and Multiply has built the first AI model for B2B advertising.”
The $50 billion market figure comes from Mayfield’s own framing and has not been cross-referenced against independent market data.
Multiply is, in essence, making a structural argument about where the ad agency model breaks down: not in creative execution, but in the speed of the feedback loop.
Whether a $9.5 million AI stack can fix that faster than incumbents adapt is the question its pipeline metrics are presumably meant to answer.
You must be logged in to post a comment Login