AI Claws: The Power of Autonomous Agents for the Enterprise

The last few months have been abuzz with a new type of agent. Popularized by OpenClaw, these AI Claws have captured the imagination of developers and business leaders to become the fastest-growing open source project in Github history.

But alongside the excitement comes very real security concerns such as sensitive data leaked to the public internet, supply chain vulnerabilities exposed through open source dependencies like LiteLLM, and growing awareness that these powerful agents aren't yet ready to run as-is in enterprise environments. For many organizations, AI Claws agents have become the latest entry on the enterprise risk register rather than a tool teams can actually use.

Today, we're changing that. We've released a new labs package to our Agentic App Store that we see as a critical first step toward bringing these agents into a framework where customers can use them productively and securely.

What is an AI Claw?

At their core, AI Claws are broadly capable, long-running agents that can handle a variety of tasks and improve over time. What makes them distinctive is a combination of capabilities that together represent a real step forward in what agents can do.

Unlike traditional agents that follow pre-defined workflows, claws can decompose ambiguous goals into subtasks on their own, persist context across long-running sessions, and adapt when their first approach doesn't work, including by writing their own tools mid-task. They can use the OS and many of its capabilities (browser, terminal, file system, etc.), log information in a simple memory file, and enable simple chat-based user interaction.

In short, they behave less like a chatbot and more like a capable colleague who can sit down at a computer and get things done. They also represent a next, higher level on the autonomy scale from what customers like ours are already using today with Automation Anywhere’s goal-based agents.

What can AI Claws do for the enterprise?

Because AI Claws can adapt to unfamiliar problems and develop new approaches on the fly, their potential in the enterprise spans a broad range of use cases that exceed what even some of the most advanced and well-engineered agents could do only 6 months ago. Any time you need to investigate, analyze and take action on a complex problem that might require a high degree of adaptability and development of new skills, AI Claws may help.

Here are just a few representative examples across IT, Financial Services and Healthcare:

• Infrastructure vulnerability research - An agent scans CVE databases, cross-references them against the organization's active software inventory, assesses exploitability in context, and delivers a prioritized risk brief to the security team, with remediation recommendations already mapped to the affected systems.
• Investment thesis research - An investment manager describes a thesis in plain language. The agent scours SEC filings, earnings transcripts, analyst reports, internal sources, and market data, surfacing supporting and contradicting evidence, and returns a structured research memo with citations.
• Clinical evidence synthesis - Before a complex case review, an agent searches PubMed, clinical trial registries, and internal case archives for relevant studies, synthesizes the evidence by outcome and patient profile, and delivers a structured literature brief to give clinicians in-depth research in minutes rather than days or weeks.

Putting AI Claws to work with orchestration and management

It’s important to start by recognizing that not every job needs an AI Claw. Many times a more well-defined goal-based agent, AI skill, or automation is the right tool for the job. The real power comes in combining them.

AI Claws have quickly proven their value for individual users looking to automate personal workflows on their own devices. But realizing their impact in the enterprise is a different challenge entirely. These agents need to be managed, deployed in controlled environments, and orchestrated as part of larger processes.

That's where our platform comes in. With this labs package on our app store, teams can manage agent definitions, runtimes, and deployments in a secure, governed environment. They can orchestrate Claw agents alongside other agents, automations, and human decision points within end-to-end processes.

Adding AI Claws opens up new use cases for highly complex tasks in the enterprise. Early customers are especially interested in root cause analysis for IT issues that require working across many systems and complex data to find what is often a needle in a haystack.

See AI Claws in action as they power enterprise automation — coordinating intelligent agents to solve complex, multi-step challenges with centralized governance and control.

Addressing security concerns today and into the future

More work is required to make AI Claws fully enterprise-ready from a security standpoint, and that work is actively underway with us and a team of partners. Credential management and fine-grained access controls, tighter tool-level permissions and guardrails, security scanning, comprehensive audit logging, and agent identity are all areas we are actively developing.

But waiting for perfection means missing the window to learn. Platforms like ours provide the foundation to start using these agents productively and securely today, while our offering and the broader ecosystem continues to mature around them.

Integrating AI Claws into your agentic process automation strategy

Think of AI Claws as a powerful new instrument in the orchestra, best suited today for tasks like research and analysis in controlled environments with well-defined access boundaries. Coupled with our goal-based agents, automations, AI skills, and people in the loop, they introduce new possibilities for businesses to take on some of their most complex and time-consuming challenges across IT, service operations, finance, and beyond.

We're just getting started. Join us at Imagine 2026, May 19-20 in Dallas, to hear how we're building this capability deeper into our platform and making it available for every business. This labs release is just the opening chapter, join us to see what's next.

You can try out our new labs package by downloading it from the Agentic App Store.