Automation Anywhere launches EnterpriseClaw

Automation Anywhere’s EnterpriseClaw launch is one more sign that the enterprise AI race is moving away from “which model is smartest?” and toward “who can let agents act inside real systems without breaking governance.” Automation Anywhere said on May 19 that EnterpriseClaw was developed with Cisco, NVIDIA, Okta and OpenAI to let companies run AI agents across cloud platforms, desktops, on-premises systems and secured enterprise networks under centralized orchestration and control. Okta said its role is to apply the same authentication and least-privilege authorization to agents that companies already apply to human users. (prnewswire.com) That matters because this is not just about chatbots answering questions. Automation Anywhere is pitching agents that can do work where enterprise work already lives — inside internal apps, desktops, networks and operational systems. The company said EnterpriseClaw is meant to make “claw-style AI agents” usable in enterprise operations, with Cisco, NVIDIA, Okta and OpenAI supplying pieces tied to infrastructure, security, identity and models. (prnewswire.com) The Dell/OpenAI announcement points in the same direction. OpenAI said on May 18 that it is partnering with Dell Technologies to bring Codex to hybrid and on-premises enterprise environments so companies can deploy AI coding agents closer to the data, systems and workflows they already control. OpenAI described Codex as one of its fastest-growing enterprise products and said the partnership is aimed at customers that need deployment options beyond a cloud-only setup. (openai.com) Put those two announcements together and the pattern is clear: vendors are trying to move agents from demo environments into governed enterprise infrastructure. In practice, that means agents are being positioned nearer to internal codebases, documents, identity systems and workflow engines, rather than being kept at the edge as standalone assistants. That inference is supported by Automation Anywhere’s emphasis on centralized control across enterprise systems and OpenAI’s focus on hybrid and on-premises deployment with Dell. (prnewswire.com) OpenAI’s outreach to startups adds a third piece. The Economic Times reported on May 20 that OpenAI is offering $2 million worth of tokens to each startup in the current Y Combinator batch, and Sam Altman said in a post that he was excited to see what “tokenmaxxing startups” build and how they use the tools internally. The offer is designed to push more young companies to build products and workflows on OpenAI’s stack. (economictimes.indiatimes.com) The operational question now is less about whether agents can generate output and more about how companies assign responsibility when agents take actions inside production systems. Automation Anywhere’s own framing stresses governance and control, while Okta’s role centers on authentication and least-privilege access. Those details point to three requirements that enterprises will likely treat as table stakes: a clear initiator identity, an audit trail of what the agent did, and a human owner who can approve, dispute or roll back the result when something goes wrong. That is an inference from the product architecture and partner roles described in the announcements, not a direct quote from any one company. (prnewswire.com) The immediate takeaway is that enterprise AI is being packaged less as a standalone model subscription and more as a controlled execution layer tied to identity, infrastructure and internal systems. Automation Anywhere is selling orchestration across enterprise environments; OpenAI and Dell are selling proximity to enterprise data and workflows; and OpenAI’s YC token offer is aimed at getting the next generation of startups to build on that stack early. (prnewswire.com)