ServiceNow has announced AI Passport, a cryptographically signed authorisation credential designed to govern AI agents and MCP servers while workflows are running. It is scheduled to be available in the fourth quarter of 2026.[1] The image of a digital badge helps to explain the principle: an agent must be able to justify the resources it consults and the actions it takes, within a scope defined by the company.
An agent’s identity is not enough to define its rights
An assistant that answers a question and an agent that modifies a record do not have the same access needs. As soon as a system uses tools, it is necessary to determine what it can read, write or trigger, and in what circumstances additional approval becomes necessary.
According to ServiceNow, AI Passport makes it possible to check the authorised scope at each step of a workflow involving several agents. The system provides for delegation with progressively narrower rights, targeted revocation and graduated responses that can go as far as stopping the agent.[1]
The challenge is to make these limits effective at the moment of action. An instruction asking the agent to be careful is no substitute for a technical control capable of refusing an operation. Likewise, recording an action in a log does not, on its own, prevent it.
MCP connections are part of the scope to be secured
The MCP protocol connects AI applications to tools and resources. The official documentation points out that servers using authorisation must verify access tokens, in particular that the tokens are actually intended for them.[2] A technical connection to a tool is therefore not a general permission to use it.
For a company, verification must cover the whole chain: the agent, the server exposing a function and the application performing the operation. If any of these components has overly broad rights, the limits displayed in an interface may not match what is actually possible.
ServiceNow announces a collaboration with NVIDIA to enforce these controls in the infrastructure, notably with OpenShell. The Super Intelligence Control Tower is designed to keep actions traceable.[1] The actual coverage of tools and environments will need to be checked during deployment.
Four questions before giving an agent a badge
This editorial grid helps to translate authorisation management into concrete decisions.
| Component | Operational effect | Control question |
|---|---|---|
| Scope | Limit the data and tools that can be accessed. | Do the rights match the assigned task? |
| Delegation | Govern the handover of a task to another agent. | Do permissions remain limited at each step? |
| Revocation | Withdraw authorisations when the situation changes. | Is the withdrawal actually enforced by connected systems? |
| Traceability | Link an operation to its authorisation and its outcome. | Can we understand who allowed the action and what it changed? |
An authorised operation can still be wrong
Take a support agent authorised to consult tickets and draft a reply. This task does not automatically justify resetting an account or changing a network configuration. Those operations should have their own rules, with approval suited to their consequences.
Even within its scope, the agent may misinterpret a request or select the wrong record. Permission controls must therefore be complemented by testing, checking of the result and a recovery procedure. This example is a suggested framework, not a standard configuration announced for AI Passport.
Inventorying agents and matching costs to usage
ServiceNow is also announcing detection of unauthorised AI and AI Cost Management, expected in the fourth quarter of 2026. The aim is to link spending more closely to usage and results.[1] These functions should be distinguished from permission control itself.
For SMEs, start with one agent and one specific task
An SME could look at an agent already used for support or document processing. The pilot would involve identifying its access, removing unnecessary permissions and testing that an out-of-scope action is refused. It would also be necessary to check that rights are withdrawn and that the task can be taken over manually.
Agent security becomes an everyday administration job
This development gives Cybersecurity a central role. Professionals must apply the principle of least privilege: granting only the access needed for a task. They must understand the connections between agents, tools and applications in order to spot excessive permissions.
Their role also involves testing controls. An expected refusal must be checked in practice, as must the revocation of access or the stopping of a task. Logs must make it possible to explain actions without needlessly exposing sensitive information.
Oversight continues as usage evolves. A new connection or a change of task can alter the level of risk. The skill lies in adjusting authorisations, investigating incidents and organising the return to manual control with business teams and application owners.
Control will be judged at the moment of action
AI Passport highlights a key question for agentic AI: how can permissions be made verifiable and enforceable throughout a task? The answer will depend on integration with real systems. A badge will be useful if the associated controls work and if teams can understand, interrupt and correct operations.
Further reading
Four CCCLX reads to explore permissions, security and agent integration in more depth.
Azure Linux 4.0: is Microsoft preparing for life after Windows Server?
Open-source security scales up: IBM, Red Hat and Palo Alto join forces
GitHub Copilot bulks up for business: coding agents come under IT control
AI voice agents: Oracle details a LiveKit architecture on OCI
Sources
- 01
ServiceNow Newsroom (15 October 2026). “ServiceNow advances its operational harness for enterprise AI with AI Passport and expanded governance capabilities”. Announcement of AI Passport and the associated governance capabilities.
- 02
Model Context Protocol (accessed 16 October 2026). “Authorization”. Documentation on access token verification by MCP servers.
Editorial check of 16 October 2026. ServiceNow plans to make AI Passport, unauthorised AI detection and AI Cost Management available in the fourth quarter of 2026. The framework examples and the SME pilot are editorial suggestions, not standard configurations or announced customer results.
PDFs, presentations, audio: Adobe wants to turn Acrobat into a content factory
Acrobat no longer just wants to open your documents For more than thirty years, Acrobat has mainly been used to read, create, edit, sign and share PDFs. Adobe now wants to give it a much broader role. On 9 September 2026, the company announced new capabilities based...
Salesforce plugs Claude into its CRM: Claudeforce prepares for life after the interface
What if the CRM soon no longer needed to be opened? For decades, using a CRM meant navigating between customer records, dashboards, menus and forms. With Claudeforce, Salesforce and Anthropic want to change this logic: rather than forcing users to go into Salesforce...
Kimi K3, Mistral, DeepSeek: Oracle turns OCI into an AI model hub
Oracle wants to make OCI a one-stop shop for AI Oracle is stepping up its efforts in the AI cloud battle by significantly expanding the catalogue of models available in OCI Enterprise AI. On 4 September 2026, the group announced in particular the integration of Kimi...
Agentic Data Cloud: Google rethinks data for the age of AI agents
AI agents are forcing Google to rethink data architecture AI agents promise to search for information, make decisions and act directly within company systems, but this autonomy depends on one essential condition: accessing the right data and, above all, understanding...
Google Cloud tackles the hidden bill of AI agents
AI agents raise a new problem: how much does a task really cost? As companies move from chatbots to AI agents able to work through multiple steps, the question of cost becomes much more complex. A chatbot generally handles a request and generates an answer, whereas an...
Workday, Navan, Avalara: ChatGPT Work moves into business software
ChatGPT Work is gradually leaving the chat window With GPT-6 Astra, OpenAI is no longer simply trying to improve the quality of ChatGPT’s answers. The company wants to bring its AI closer to the applications in which employees actually work. With the arrival of GPT-6...
From chip to cloud: Groq raises 350 million to scale up
Groq now wants to sell cloud rather than just chips Groq is pursuing a major strategic transformation. Long seen as one of Nvidia’s challengers thanks to its LPUs (Language Processing Units) specialised in artificial intelligence model inference, the group is now...
Muse Code: Meta launches its AI agent to tackle giant codebases
Meta wants to bring AI agents into complex software projects Meta is stepping up its efforts in AI-assisted software development with Muse Code, a new agent designed to work on large codebases, where traditional assistants can quickly lose the context needed to...
Qwen3.8-Max reportedly coded for 16 days without human intervention
Alibaba tests how far an autonomous AI developer can go Alibaba claims to have reached a new milestone in agentic software development with Qwen3.8-Max, its new flagship model launched in early August 2026. In an internal test, the model reportedly worked fully...
Two months, $1.1 billion: River AI is already exciting investors
A spectacular raise for a start-up barely out of the shadows Two months after emerging from stealth mode, River AI has just completed one of the most spectacular funding rounds of the year in artificial intelligence. The start-up founded by Igor Babuschkin, a former...
