Diana Intelligence Corp, a San Francisco startup backed by Y Combinator and General Catalyst, launched Diana on April 15. The product is the first fully managed, enterprise-grade implementation of the OpenClaw framework delivered as a SaaS, embedded directly into Slack.

The core pitch: every employee gets their own isolated, dedicated AI agent in seconds. No local setup, no infrastructure management, no wrestling with the security risks that have defined OpenClaw’s reputation in enterprise circles throughout 2026.

The “Boss” Layer

Diana’s differentiator is what the company calls the Boss, a human governance mechanism that maintains oversight and control over each agent’s actions. The product site describes a prompt ingress scanner that detects adversarial injection attempts and an outbound traffic audit console that blocks PII, credit card data, and API key exfiltration in real time.

According to Diana’s product page, the system has scanned over 1.2 TB of data with a 99.9% compliance rate and 142 violations caught. The Governor, as the monitoring layer is called, runs 24/7 on every agent instance.

This directly addresses the operational risk that has dominated OpenClaw coverage this month. The Qualys attack path analysis published April 14 showed how a single OpenClaw vulnerability could chain into full domain controller compromise. Diana’s response: don’t let enterprises run OpenClaw themselves.

Slack as the Delivery Surface

Rather than building another standalone interface, Diana embeds entirely within Slack. The “Give Your Employee An Employee” tagline captures the positioning: each knowledge worker gets a dedicated agent, not a shared chatbot. The Slack-native approach eliminates the onboarding friction that has kept OpenClaw adoption concentrated among developers and power users.

The team behind Diana includes engineers from MIT, Carnegie Mellon, and Yale, according to the product site.

The Enterprise OpenClaw Market Takes Shape

Diana’s launch arrives on the same day that TechCrunch reported Microsoft is building a proprietary OpenClaw-like agent for enterprise Copilot customers with tighter security controls. Two competing approaches to the same problem: Diana wraps the original framework in governance; Microsoft builds its own from scratch.

The enterprise agent governance market is crystallizing around simultaneous product launches this week, with multiple vendors shipping competing approaches to the same problem: how to get agent capabilities into production environments without accepting unmanaged risk.

The Managed OpenClaw Bet

Diana is betting that OpenClaw’s architecture is correct but its operational model is wrong for enterprises. The open-source runtime is powerful, but nine CVEs in April alone, plus the Qualys attack chain demonstration, have made self-hosting a liability for any company with a security team that pays attention.

Whether Diana can convert that liability into a subscription remains the open question. For enterprises that want OpenClaw’s capabilities without its operational overhead, Diana is the first credible managed option on the market.