Large language model-based agents are evaluated for their ability to perform knowledge work tasks on the ServiceNow platform, highlighting a performance gap and the need for further exploration.
