Microsoft Taps Yobi AI to Challenge Data Garden Monopolies

Microsoft and behavioural AI pioneer Yobi have revealed a partnership that aims to dismantle the data monopoly long held by social media and search giants. 

By integrating Yobi’s massive behavioural foundation models directly into the Microsoft Azure ecosystem, the duo is handing mid-to-large-scale enterprises the same predictive crystal ball previously exclusive to platforms like Meta and Google. 

This represents a shift from deterministic marketing, which reacts to what a user just searched for, to predictive marketing that anticipates what a user will want next based on deep behavioural patterns.

What’s special about Yobi?

While LLMs like OpenAI’s GPT-4 are trained on text to predict the next word in a sentence, Yobi’s 700bn-parameter Behavioral Foundation Model is trained on actions. 

It ingests billions of consented data points including purchases, store visits and marketing conversions to predict the next likely action in a consumer's journey. 

Historically, only companies with their own walled gardens, which are closed ecosystems where they track every click could build models this large. 

By moving to Azure, Yobi can now leverage Microsoft’s AI infrastructure to run these proprietary models for third-party enterprises.

This partnership addresses a major pain point in the industry known as ‘bottom-funnel exhaustion’.

Most digital advertising today is stuck in the lower funnel, meaning brands spend billions competing for the same shoppers who have already expressed intent, such as someone who just searched for running shoes. This drives up costs and often merely sustains demand rather than growing it. 

Yobi’s behavioural AI allows brands to find net-new audiences at the top of the funnel by identifying patterns in shoppers before they even start their search, allowing enterprises to acquire high-lifetime-value customers at a fraction of the cost of traditional search ads.

According to Max Snow, CEO of Yobi AI, understanding and predicting customer intent is a competitive necessity, but enterprises today face a massive data disadvantage.

“Microsoft Azure offers infrastructure and best-in-class tooling that allows Yobi to train proprietary 700B parameter models –  unlocking for companies like Wolverine the ability to optimise sales without compromising privacy,” he says.

Finding the shoppers search ads missed

Footwear giant Wolverine Worldwide, the parent company to Merrell and Saucony, served as the primary proof of concept the idea that personalisation at scale can unlock outsized returns by reaching high-value consumers that traditional advertising channels often miss.

Wolverine used Yobi to find shoppers that legacy channels were missing entirely. The result produced some of the strongest incremental returns – revenue that would not have happened otherwise – the company has seen outside of paid search.

One of the biggest hurdles for enterprise AI is governance, as borrowing behavioural data is a legal minefield in an era of strict privacy laws. Yobi solves this by creating privacy-preserving customer representations.

“By combining Yobi’s consented behavioural data with Microsoft’s cloud and AI platform, organisations like Wolverine Worldwide are gaining critical insights that improve customer engagement and drive meaningful business growth,” says Judson Althoff, CEO of Microsoft’s commercial business. 

“This partnership reflects our commitment to building AI solutions that are both innovative and responsible, with trust and privacy at the core.”

Essentially, the AI understands the intent of a consumer segment without ever exposing their personal details or personally identifiable information. 

For Azure customers, this has become a turnkey solution where enterprises can purchase Yobi through the Azure Marketplace to centralise their own permissioned data and activate Yobi’s signals in real-time. 

This partnership signals that the data wars are entering a new phase where the advantage is no longer about who has the most users, but who has the best model to predict what those users will do next.