The next big marketing partnership could be hiding in plain sight — somewhere among millions of everyday social media conversations. Researchers at Penn State have developed an artificial intelligence (AI) framework that could help businesses discover unexpected brand partnerships by analysing what consumers say online, potentially transforming how companies identify collaboration opportunities.

Brand collaborations have become an increasingly popular way for businesses to attract attention, reach new audiences and strengthen their market presence. However, identifying suitable partners often involves expensive surveys, extensive research and considerable guesswork. The new AI-powered approach promises to make that process faster, more systematic and better informed.

Published in the journal Information Systems Research, the study combines traditional brand characteristics with consumer conversations on social media. By examining how frequently brands are mentioned together and the context of those discussions, the system identifies promising partnerships that conventional marketing research might overlook.

“Finding brand collaborations is a needle-in-the-haystack problem,” said Keran Zhao, assistant professor of supply chain and information systems at Penn State’s Smeal College of Business. He explained that while managers might consider a handful of familiar partners, AI can systematically evaluate hundreds of brands and explain why particular combinations deserve attention.

To develop the framework, researchers examined personality characteristics, including sophistication and familiarity, for 450 major consumer brands. They also analysed 477,933 Twitter posts from 2020 mentioning multiple brands. AI classifiers filtered out complaints and assessed whether consumers perceived the brands as complementary or competing, creating a network of potential relationships.

The results were encouraging. The system correctly predicted 213 of 251 brand alliances formed after the researchers’ data collection period. In a separate evaluation involving 57 brand managers, AI-recommended partnerships were selected approximately 34% of the time, compared with just 7% for combinations the system had rejected — nearly five times as often.

The researchers also surveyed 350 consumers, who showed a strong preference for the AI-generated recommendations. Together, these findings suggest that social media conversations can reveal meaningful connections between brands, offering businesses an alternative to relying exclusively on traditional market research.

Beyond identifying potential partners, the four-stage framework helps companies prioritise opportunities. Using machine learning, it evaluates social media visibility, consumer sentiment and engagement, then sorts recommendations into categories ranging from high-profile “premier opportunities” to lesser-known “hidden niches” that could appeal to specialised markets.

Generative AI adds another layer by interpreting what consumers actually discuss, including products, promotions, style and sustainability. These insights could help marketers decide whether a partnership would work best through a joint product launch, advertising campaign or another collaborative initiative.

For businesses, the appeal extends beyond saving time and money. Brand partnerships carry reputational and financial risks, making careful selection essential. By turning everyday consumer conversations into practical marketing intelligence, the researchers believe AI could uncover creative alliances that might otherwise never be considered — including some surprisingly successful combinations.

More information: Pankhuri Malhotra et al, Using Information Networks to Discover and Execute Marketing Brand Alliances: A Data-Driven Framework for Actionable Insights, Information Systems Research. DOI: 10.1287/isre.2025.1792

Journal information: Information Systems Research Provided by Penn State