The new Workday AI Research team will focus on the practical friction points that often stall enterprise adoption, including multi-agent orchestration and the reliability of AI memory. By bridging the gap between academic machine learning and real-world HR, finance, and IT workflows, the company intends to standardize how businesses manage AI guardrails.

Recent internal studies highlight the urgency of this work. Researchers discovered that standard 'forget' commands frequently fail to fully scrub data, with information remaining recoverable in one out of five instances. Other findings show that selective memory systems can perform 31% faster than conventional models while retaining 97% of essential context. To sustain this pipeline of innovation, Workday is establishing a $50,000 PhD fellowship, pairing doctoral candidates directly with company scientists to tackle the complexities of large-scale, enterprise-ready AI.