Securing the agentic enterprise with data lineage
MIT Technology Review
AI agents handling automated sales workflows without human oversight risk compounding errors across multiple data sources.
An automated sales workflow demonstrates the potential risks of agentic AI systems when multiple agents interact without human supervision. A sales representative's request for pricing leads one AI agent to pull restricted and archived data. A second agent then uses this unverified output to draft a proposal, highlighting the challenge of data lineage in autonomous enterprise environments.



