Critical infrastructure use case · OpenAI + ServiceNow + Microsoft 365
Field technician assistant
Technicians search long manuals and past service notes while working on equipment whose configuration may differ from the standard example. Technicians receive source-linked answers from authorized manuals and asset history, in the language they use on site.
Systems involved
What each system contributes
25 actions
- Chat Completions8
- Files4
Manual retrieval uses the configured knowledge-source interface through an API, database or Rules + Code step. The model step runs against the AI provider you choose, including Anthropic, Google Gemini or a model you host yourself. Anything outside a pre-built action runs through an API, protocol or database call, or as Rules + Code in Python, Bash, Go, Java, .NET, PHP, PowerShell or JavaScript/TypeScript, with the same approvals and audit.
Architecture
One governed run
How it starts: A technician asks an asset-specific question through the designated Microsoft 365 interface.
Step by step
How the process runs
- 01
Identify the equipment
Resolve the technician's permitted ServiceNow asset or configuration-item reference and available service history.
- 02
Retrieve the right material
Select approved manual passages matching the equipment version through the configured knowledge-source interface.
- 03
Explain in the technician's language
The selected model drafts an answer grounded in those sources and separates documented instructions from unresolved uncertainty.
- 04
Check operational boundaries
Rules enforce access restrictions and route restricted or safety-sensitive recommendations to the responsible supervisor.
- 05
Return and record
Microsoft 365 delivers the answer with source references, while an approved ServiceNow action can record the reviewed service note.
Control
Human decision and exception path
Human decision
The named supervisor approves any recommendation that requires a controlled operational change or overrides documented procedure. The assistant does not actuate equipment, grant access or approve a technician's authority to perform a task.
Exception path
- Unknown equipment versions or missing manuals return a clarification request instead of generic operating instructions.
- Conflicting sources stop the recommendation and escalate to the technical owner.
- Unsupported or safety-critical advice is withheld pending review, with the unresolved question retained.
Evidence
What gets audited
- Technician question and authorized asset context.
- Manual versions, service-history references and retrieved passages.
- Model answer, language and any supervisor decision.
- Delivered response and approved ServiceNow note reference.
Nexus runs on premises, in private or sovereign cloud, or air-gapped. Credentials stay in the vault, roles decide who can approve, and the execution record stays inside your deployment boundary.
Sovereignty and deploymentRelated
Patterns, connectors and next steps
- Ivanti-ServiceNow integration
Integration pattern · Cross-team service work stays linked across Ivanti and ServiceNow without competing ticket ownership.
- OpenAI connector
25 pre-built actions
- ServiceNow connector
55 pre-built actions
- Microsoft 365 connector
45 pre-built actions
- Human-in-the-loop
Where approval belongs in a running workflow.
- Infrastructure alert to incident and remediation
Another critical infrastructure use case.
FAQ
Questions about this use case
- Does Nexus replace any of these systems?
- No. Each system keeps its own records. Nexus runs the process between them and records each decision.
- How does the run start?
- A technician asks an asset-specific question through the designated Microsoft 365 interface.
- What if a step is not covered by a pre-built action?
- Manual retrieval uses the configured knowledge-source interface through an API, database or Rules + Code step. The model step runs against the AI provider you choose, including Anthropic, Google Gemini or a model you host yourself. Anything outside a pre-built action runs through an API, protocol or database call, or as Rules + Code in Python, Bash, Go, Java, .NET, PHP, PowerShell or JavaScript/TypeScript, with the same approvals and audit.
- Can the manuals and model remain inside our boundary?
- Yes, when the model and knowledge store are hosted locally; a Microsoft 365 interface still requires approved connectivity and is not part of a fully air-gapped deployment.
- Does translation change the technical authority of the answer?
- No; translated guidance remains tied to the original source, and unclear terminology or restricted actions are escalated rather than reinterpreted as permission.
Use case review
Discuss this use case
Bring your version of the process. We will map the systems, approvals and evidence it needs.