TBNTech
Making the case for context-aware AI
Publication: MIT Technology Review
Author: MIT Technology Review Insights
Source published: 22/10/2026, 10:00:00 UTC
Briefing created: 22/10/2026, 10:23:08 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 22/10/2026, 10:20:24 UTC.
Briefings
Enterprise AI lacks operational context when transitioning from experimentation to mainstream operations, leading to significant performance deficits. Devavrat Shah, chief scientist at Celonis, argues that AI agents require specific business context to function effectively, as generalized principles fail in specialized scenarios. Without this context, AI risks costly mistakes, resource waste, and missed constraints. Shah advocates for a phased approach starting with specific business decisions to measure value before scaling, emphasizing that decision intelligence depends on connecting AI to existing enterprise processes.
Key points
- Enterprise AI agents currently operate without operational context, causing performance deficits when applied to specialized scenarios.
- AI agents require specific business context to understand workflows, operate within guardrails, and produce actionable recommendations.
Context
Devavrat Shah, chief scientist of enterprise AI at Celonis, explains that while enterprise AI can automate tasks and generate recommendations, it lacks the operational context needed to make correct decisions. This context is essential for connecting AI to existing enterprise processes and data, enabling agents to understand the role of each task and the business conditions surrounding it.
Limitations
The evidence is limited to Shah's interview transcript and mentions of Celonis and MIT Tech Review. Specific quantitative data, dates, or named actors beyond Shah and Megan Tatum are not provided. The article does not contain causal claims or opinions, only Shah's stated arguments and observations.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 22/10/2026, 10:23:08 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeies4axb4ghytp2ceqznnvzwesay6uopoz3qy6tizhxsuq2bdeubua