Celonis's Devavrat Shah argues context is key for enterprise AI decision intelligence

Enterprise AI needs operational context to move from experimentation to everyday operations, according to Devavrat Shah, chief scientist of enterprise AI at Celonis. Without understanding business constraints and processes, AI agents risk costly mistakes or excessive resource consumption. The solution involves connecting AI to existing enterprise data and processes, allowing agents to operate within defined guardrails. Shah advises starting with a specific business decision to measure gains in speed and value, noting this could free employees for strategic work, provided there is investment in workforce education.