Automation Anywhere Launches EnterpriseClaw With Cisco, Nvidia, Okta, and OpenAI to Govern Claw-Style AI Agents
Automation Anywhere announced EnterpriseClaw on May 19, a platform built with Cisco, Nvidia, Okta, and OpenAI that wraps claw-style AI agents in enterprise governance controls. The product targets organizations that want autonomous agents operating across desktops, cloud platforms, on-premises systems, and secured networks, all under centralized policy enforcement.
What EnterpriseClaw Does
The platform runs on Automation Anywhere’s Process Reasoning Engine (PRE) and Contextual Intelligence Graph, according to the company’s press release. Each integration partner contributes a specific layer: Cisco AI Defense and DefenseClaw provide security controls purpose-built for agents, Nvidia contributes its open-source OpenShell runtime alongside NIM microservices and Nemotron models for on-premises deployments, Okta handles cross-agent identity management and authentication, and OpenAI provides GPT-5.5 model access.
Agents deployed through EnterpriseClaw run in managed containers behind firewalls with local access to files, apps, browsers, and terminals. They can hand off tasks between each other and combine outputs across workflows. Enterprises set policies, access controls, guardrails, and agent credentials, all enforced locally on-device, with telemetry, audit logs, and LLM usage tracking.
“For AI to have a transformational impact on business, it needs to be able to do work where the work actually happens,” Automation Anywhere CEO Mihir Shukla said in the announcement. “Many claw-style AI agents are incredibly powerful, but most were designed for isolated environments or individual users.”
EnterpriseClaw is available in preview now, with general availability expected later in 2026.
The Differentiation Question
Not everyone is convinced. Jason Andersen, VP and principal analyst at Moor Insights & Strategy, told Computerworld that Nvidia’s existing NemoClaw open-source stack already offers essentially the same governance capabilities. “Which begs the question: If you are already using Nvidia’s, why choose this?” Andersen asked. He noted that the Cisco and Okta integrations will likely interest their existing customer bases, but those products already work with other tools.
Manish Jain, principal research director at Info-Tech Research Group, offered a more favorable assessment. “The collaboration between Nvidia, OpenAI, Okta, and Cisco adds to the credibility of the proposition of trusted infrastructure, identity, and security layers,” he told Computerworld. Jain characterized OpenClaw’s original launch as a tool that “did not meet enterprise-grade product standards,” pointing to data leaks and uncontrolled agent behaviors as evidence that governance tooling was inevitable.
The Governance Layer Race
EnterpriseClaw enters a market that has formed rapidly since OpenClaw went viral in November 2025. Technology analyst Carmi Levy told Computerworld that OpenClaw “shifted the notion of AI from something we chat with to something that actually gets work done,” creating a key step in “replacing human capital with technological capital.”
That shift created the governance vacuum EnterpriseClaw aims to fill. OpenClaw operates as a client-server model across many different LLMs with no built-in governance capabilities, as Andersen noted. Nvidia launched NemoClaw to address the same gap. OpenClaw itself published a five-point security roadmap in May. The question for enterprises is no longer whether agent governance is needed, but which vendor’s stack will become the default.
Automation Anywhere’s bet is that the answer involves multiple vendors collaborating rather than any single provider owning the full stack. Whether that consortium approach wins over Nvidia’s integrated offering will likely depend on which enterprise buying centers, security teams versus infrastructure teams versus IT operations, end up controlling agent deployment budgets.