Dow Jones tests automated RFP processing

Dow Jones is using generative AI to streamline how it responds to incoming RFPs (requests for proposals) from advertisers, a task traditionally handled by humans due to the complexity and high stakes involved. Kedar Prabhu, VP of ad product and technology at Dow Jones, explained that their mid-nine-figure advertising business depends heavily on responding to RFPs with customized media plans. While this has historically required a hands-on approach, Dow Jones is now looking to AI to help automate the process of interpreting RFPs and crafting suitable responses—though not replacing media planning itself.

The company began exploring automation two years ago, seeking out B2B tech vendors to assist in responding to RFPs more efficiently. However, no existing solutions could fully address their needs, prompting Dow Jones to build their own AI system. The challenge? RFPs are far from standardized; they vary in language, format, and complexity, meaning that AI couldn’t simply be fed these documents and expected to produce perfect responses. After determining the cost of building such a system in-house, they decided to proceed with a proof-of-concept model, training it on a dataset of past RFPs and responses over a period of months.

Currently, Dow Jones is testing a "zero-shot" AI model, which is designed to process RFPs without needing specific examples for training, making it adaptable to diverse formats. Over the next six months, they plan to roll out further iterations, with the goal of demonstrating improvements in both efficiency and effectiveness in how they handle RFPs. While the AI may not be automating media planning itself, it has the potential to revolutionize how quickly and accurately Dow Jones can respond to advertisers, driving both revenue growth and cost savings.

Full story at Digiday.