Workday, an enterprise HR, finance and IT platform, has launched a dedicated AI research team focused on technical challenges associated with deploying artificial intelligence in enterprise environments.
The Pleasanton, Calif.-based company said Workday AI Research will investigate areas including persistent agent memory, agent orchestration, reward overoptimization, recommendation systems and adaptive resource control. Its research already has been accepted by conferences including the International Conference on Machine Learning and the International Conference on Learning Representations.
“As AI agents evolve to remember context and take action on behalf of employees, enterprises are facing complex challenges around privacy, auditability, efficiency and enterprise-grade accuracy that off-the-shelf models simply cannot solve,” said Gerrit Kazmaier, president of product and technology at Workday.
Among the team’s recent findings, researchers developed a selective approach to agent memory that improved precision by 12 percent and overall memory quality by approximately eight percent in testing. Another study found that dividing complex tasks among specialized agents improved accuracy by nearly six percent.
Research into agent memory deletion also found that information an agent was instructed to forget remained recoverable from previous summaries about 20 percent of the time. According to the company, the research team concluded that fully removing information requires deleting every summary containing it.
Workday has also unveiled a PhD fellowship that will provide selected doctoral students with $50,000 in annual research funding, mentorship and collaboration with Workday researchers.
