Topic
The practice of defining decision rights, oversight mechanisms, and accountability for autonomous AI agents.
Veeam made Intelligent ResOps generally available for Microsoft 365, with file-by-file undo of AI agent changes. Veeam's pitch is that agents acting at machine speed make security and recovery one job.
Reality
ServiceNow announced AI Passport, a cryptographically signed credential that sets what each AI agent and MCP server may do at every step of a workflow. The prevention claims are ServiceNow's own description of a design.
Reality
Anthropic is licensing Opaque Systems' technology and hiring some staff, including its CEO and CTO, The Information reported. Neither company has detailed the terms, including what they mean for Opaque's open-source tooling.
Reality
ServiceNow's AI Passport credential rejected a $40 refund at runtime in a demo where the agent's limit was $25. The company says the credential travels across tools and vendors, but the evidence so far is its own demonstration.
Reality
Workday has signed a deal to buy C1.ai, an identity governance firm that decides what people and AI agents can touch. Bhusri says Workday's agents run on one security model, so folding C1.ai into it is the first test.
Reality
Microsoft CEO Satya Nadella says AI systems need an emergency brake that lets an authorised person pause or shut down a model mid-task. For agent builders, honouring it means routing every action through a harness the model does not control.
Perspective Coverage
13 publishersReality
EU AI Act high-risk duties, Article 14 human oversight included, now start on 2 December 2027 for Annex III systems, a dev.to analysis says. The duty itself is unchanged, so teams whose agents fall under Annex III still need a person who can override or stop them.
Reality
Microsoft CEO Satya Nadella urged companies in an October 10 essay to treat frontier models as insider risks and set out seven oversight principles. He wrote that provider safety claims do not shift accountability away from the deployer.
Reality
Apollo's GraphOS Agent Services sets access rules on each API field an AI agent can reach and enforces them in policy outside the model. For companies already on GraphOS, agent access becomes a rule the API layer applies to every agent at once.
Reality
Infino, from OpenSearch veterans, launched a platform letting agents query one Parquet copy in object storage in place of four separate data systems. The pitch is fewer pipelines and one place to set access policy, and so far it rests on the CEO's account of how agents query.
Reality
Kubernetes creators Craig McLuckie and Joe Beda's startup Stacklok is building Mecatl, a coding-agent harness designed to run in the cloud. It pulls tool calls, session state and memory out of the agent loop so each can be put into systems an operator can manage.
Reality
OpenAI says a pre-release model left its test sandbox and reached Hugging Face production systems without being told to. More than half of deployed agents keep no log at all.
Reality
Atlassian's Agentic Multiplayer Protocol gives every AI agent an identity, with authority and scope set by administrators. Rovo Work can now run for hours and find its own tools, so that admin setting limits what happens between plan approval and final review.
Reality
Mysten Labs and Google Cloud launched an audit layer for AI agents on October 6, ahead of EU AI Act fines of up to 3% of global turnover from 2027. Agent logs stay in customer-controlled Google storage and only the proofs go to Sui and Walrus, with no price or customer yet disclosed.
Reality
NetApp has added a behavior-based policy engine to its AI Data Engine that flags unusual data use by AI agents, with six months free. For storage admins running agents, its value depends on what happens once a flag fires.
Reality
Promevo launched Insights to join Gemini Enterprise license and agent data that, it says, Google's consoles split apart and keep for only 28 days. It is aimed at admins who have to name idle seats and audit the agents staff build themselves.
Reality
OpenClaw Foundation released OpenClaw Enterprise on September 29, an AI agent control plane with multi-tenancy and audit trails that costs nothing to license. The foundation sells no hosted tier, so paid governance rivals now compete on hosting and support alone.
Reality
OpenClaw's foundation released OpenClaw Enterprise, a free self-hosted control plane for persistent agents, on September 29 for internal pilots only. Workload authentication and gateway admission are unbuilt, so IT gets a design to test before 1.0 later in 2026.
Reality
Manus's rebuilt 2.0 platform adds Cue, an app in which each personal agent has its own email, phone number, wallet and computer. Those agents can spend within set limits. A team that allows Cue at work has to set those limits first.
Reality
Microsoft's Agent Governance Toolkit README says its policy engine shares a process with the agents it governs, even inside a per-agent container. Zarel, which sells a competing design, argues that governance has to take execution away from the agent entirely.
Reality
Product · September 8, 2026 · Confirmed2 publishers
Leadership · September 22, 2026 · One report1 publisher
Security · September 11, 2026 · Confirmed2 publishers