Agents at Work
Cognizant’s Claims Automation Release Tests the Business Case for Governed Agents
A healthcare claims workflow announcement shows how agentic automation is being packaged around bounded procedures and human exceptions.
3 minutes
OpenAI’s Ironclad Collaboration Points to a Harder Test for Computer-Use Agents
OpenAI is turning complex contracting workflows into training and evaluation tasks. The larger question is whether agents can reliably navigate specialized business software.
3 minutes
Before You Build an AI Agent, Fix the Process
Before building an AI agent, map the workflow’s inputs, decisions, exceptions, and outputs. Clear processes make automation easier to evaluate and govern.
7 minutes