OpenAI's Frontier wants to manage your AI agents - it could upend enterprise software, too

OpenAI announced Frontier, an enterprise platform for building, deploying, and managing AI agents, pairing forward-deployed engineers with customer teams and emphasizing agent security (identities, permissions, and guardrails) and a shared semantic layer. The report compares Frontier to Palantir’s ontology-driven model and Anthropic’s Claude Cowork, noting potential partnerships and competitive friction with firms like Palantir and Palo Alto Networks. It argues Frontier could shift the enterprise software landscape by making LLM-based agents the primary user interface, potentially diminishing the importance of traditional software suites.