- Category: Automation and AI, ITSM
- Posted by Salil Karkarni
- on
Salil Kulkarni, CSIO at Virima, argues that when an AI agent gets overridden, the issue is rarely the model. It’s a data layer that’s stale, incomplete, or missing the context a human would have caught. Drawing on decades leading technology at Caesars, Las Vegas Sands, and in hospitality revenue management, he sets out four conditions a data layer must meet before an AI agent can be trusted to act.