Terminology Standard v 0.1 Contact [email protected] with submissions, questions

Mission Statement: To define a shared vocabulary for AI Agents operating across MSP and IT Pro platforms to ensure interoperability, transparency, and reliability.

AI Agents communicate using natural language, just like we do. But in a technical environment, that introduces challenges. One user might type “device,” another might say “asset,” “endpoint,” or “laptop.” These can all mean the same thing... or not. And as these AI Agents pass instructions from one system to another, across different companies and platforms, the risk of misinterpretation increases.

To ensure these AI Agents execute the right actions, every time, we need to align on a shared language. That’s why we’re creating a structured dictionary and relationship model (lightweight ontology) tailored specifically to the MSP and IT Pro world.

This lightweight but powerful framework lets us:

• Standardize how we refer to infrastructure, services, and tools across our ecosystem
• Help AI Agents understand the relationships between devices, policies, tenants, and actions
• Improve accuracy when automating complex tasks across environments
• Prepare our platform and our partners for a future of autonomous service delivery

By leading the creation of this MSP and IT Pro specific terminology standard, we’re ensuring that our AI Agents, and the tools they interact with, speak the same language. It’s a foundational step toward smarter, more secure, and more efficient IT operations.

Terminology Template Format:

Canonical Term: [standardized object or action]

Definition: [clear description in MSP context]

Recognized Synonyms: [list of alternate words that map to this]

Relationships:
• [term] [relationship] [other term]
• Prompt Examples:
• [realistic user query]
• Agent Execution:
• [description of how agent interprets and responds]

Prompt Examples: [clear things a user may type into a chatbot or one AI Agent may pass to another including the term and an action or output. ]

Agent Execution: [examples of how an Agent would parse an instruction using the term

Current Terms

• N-able
• Distributor
• Partner
• MSP
• Organization
• Service Organization
• Site
• Location
• Tenant (customer)
• Integration
• External Vendor
• Asset (Device)
• Server
• Account
• User
• Template
• Resource
• Data
• Cloud
• Software
• Package (see also Patch)
• File
• Patch
• Vulnerability
• Hardware
• Disk
• Firewall
• Performance
• Warranty
• Certificate
• Product
• Service
• Script
• License
• Backup
• Storage
• Notification
• Report
• Network
• IP
• Workflow
• Task
• Agent
• AI Agent

Definitions

N-able

Canonical Term: N-able

Definition:
The top-level platform node representing N-able as the vendor and service provider. This is a unique, singular entity in the model. It serves as the source of platform-wide policies, agent frameworks, product definitions, and integration standards. All tenants, services, and tools exist beneath or are associated with this node.

Recognized Synonyms:

• Platform
• Vendor
• Parent org (avoid — use N-able)
• Global provider (context-dependent)

Relationships:

• N-able provides Product
• N-able defines Terminology Standard
• N-able publishes Template
• N-able manages Partner
• N-able supports Partner
• N-able manages Distributor
• N-able owns AI Agent Framework

Prompt Examples:

• “What monitoring tools does N-able provide for remote access?”
• “Does N-able have a default patching policy template?”
• “List all automation policies published by N-able.”

Agent Execution:
When the term “N-able” is used in a prompt, the agent interprets it as a query against global or platform-wide definitions that originate from the N-able platform node. The agent resolves requests by:

• Accessing the centralized N-able catalog of tools, templates, and product definitions
• Filtering results not tied to any specific tenant or MSP configuration
• Returning default (global) policies, tools, scripts, or documentation
• Providing N-able-wide metadata rather than tenant-specific data

Distributor

Canonical Term: Distributor

Definition:
A Distributor is an organization authorized by N-able to resell N-able products and services. Distributors may operate under their own branding, manage multiple MSPs or sub-distributors, and may either directly support their downstream entities or rely on N-able for support. Distributors typically have visibility and limited control over the environments of their managed MSPs.

Recognized Synonyms:

• Partner (ambiguous – avoid unless context is clear)
• Channel Partner
• Aggregator (in legacy or specific sales contexts)
• Reseller (not exact. do not use. Resellers do not typically support products or access interfaces)

Relationships:

• Distributor resells Product
• Distributor supports MSP
• Distributor manages SubDistributor
• Distributor has Visibility into Tenant
• Distributor follows PolicyTemplate from N-able
• Distributor isManagedBy N-able

Prompt Examples:

• “List all MSPs under the distributor ‘AlphaTech’.”
• “Does the distributor ‘EuroSecure’ manage support for its MSPs or is it handled by N-able?”
• “Deploy the onboarding script to all new tenants under the distributor ‘CloudBridge’.”

Agent Execution:
When the term “Distributor” is used, the agent interprets the request as applying to an intermediary layer between N-able and one or more MSPs. The agent will:

• Query the organizational hierarchy to identify associated MSPs or tenants
• Check configuration settings to determine whether the Distributor or N-able owns support responsibilities
• Determine available product sets, policies, or automation workflows that the Distributor can access or propagate
• Restrict or allow actions based on permissions assigned at the Distributor level

Partner

Canonical Term: Partner

Definition:
A Partner is a general term used by N-able to refer to any authorized entity that consumes, resells, or delivers services using N-able products. A Partner may be an MSP (Managed Service Provider), a Distributor, or another type of entity in the N-able ecosystem. In most cases, the term "Partner" is used to describe an MSP that manages end-customer environments and directly delivers IT services.

Recognized Synonyms:

• MSP (when referring to a service-delivering Partner)
• Reseller (context-specific — can refer to Distributors or MSPs)
• Service Provider
• VAR (Value-Added Reseller – legacy or sales term)
• Account (not recommended, see Canonical Term: Account)

Relationships:

• Partner manages Tenant
• Partner uses Product
• Partner consumes Policy from N-able or Distributor
• Partner owns Agent installed on Asset | Device
• Partner isSupportedBy N-able or Distributor
• Partner provides Service to End Customer

Prompt Examples:

• “Which partners have not enabled 2FA on their admin accounts?”
• “Deploy the monitoring baseline to all partners using N-central in North America.”
• “What policies has the partner ‘SecureSys MSP’ applied to its tenants?”

Agent Execution:
When a prompt refers to a “Partner,” the agent determines the role based on context:

• If the Partner manages customer environments and devices, the agent treats it as an MSP
• If the Partner acts as an intermediary (e.g., Distributor), the agent identifies upstream/downstream relationships
• The agent then scopes the action or query to the Partner’s role:
  o For MSPs: devices, scripts, tenants, policies
  o For Distributors: managed entities, visibility controls, policy propagation
  o For all: support responsibilities, license usage, product configuration

The agent enforces role-based boundaries, ensuring actions and data queries align with the Partner’s permissions and relationships in the system.

MSP

Canonical Term: MSP

Definition:
An MSP (Managed Service Provider) is a type of Partner responsible for delivering IT services to one or more customer environments (tenants). MSPs use N-able products to monitor, manage, secure, and automate infrastructure for their clients. They typically own the relationship with the end customer, administer devices, apply policies, run scripts, and handle service tickets.

Recognized Synonyms:

• Partner (when the context is service delivery)
• Service Provider
• IT Provider
• Admin Organization (rare/legacy — avoid)

Relationships:

• MSP manages Tenant
• MSP owns Asset | Device
• MSP applies Policy
• MSP runs Script
• MSP isProvisionedBy Distributor or N-able
• MSP isSupportedBy N-able or Distributor
• MSP provides Service to End Customer
• MSP has AccessControl for Agent

Prompt Examples:

• “List all tenants managed by the MSP ‘BrightNet Solutions’.”
• “Check for outdated antivirus software across all MSPs in the Western region.”
• “What scripts are frequently used by the MSP ‘DataSecure’ during onboarding?”

Agent Execution:
When a prompt refers to an MSP, the agent identifies a Partner entity that actively delivers services to customer tenants. The agent will:

• Resolve the MSP’s list of managed Tenants
• Access the MSP’s configuration (e.g., available scripts, policies, tools)
• Determine which Devices, Agents, and Users fall under the MSP’s control
• Execute actions or retrieve data within the scope of that MSP's environment only
• Respect role-based access boundaries for visibility, control, and policy enforcement

Organization (See Service Organization)

Canonical Term: Organization

Definition:
This term should probably be avoided due to ambiguity. An Organization is a general-purpose container representing any structured business or entity within the N-able ecosystem. It may refer to a Distributor, MSP, Tenant, or even N-able itself, depending on context. “Organization” is primarily used in internal models, system hierarchies, or data structures to abstract the roles of different actors in the platform. In end-user contexts, the more specific terms (MSP, Tenant, etc.) are preferred.

Recognized Synonyms:

• Org (internal shorthand)
• Business Unit (when referring to internal structures)
• Entity (generic or API-level use)
• Account (ambiguous – avoid unless referring to billing)

Relationships:

• Organization mayBe N-able | Distributor | MSP | Tenant
• Organization has SubOrganization
• Organization manages Users
• Organization owns Devices | Assets
• Organization applies Policy
• Organization has AccessLevel

Prompt Examples:

• “Which organizations have more than 50 active scripts deployed?”
• “List all organizations managed by the distributor ‘TechBridge Global’.”
• “Does this organization have access to the advanced automation catalog?”

Agent Execution:

When a prompt refers to an “Organization,” the agent interprets it as a flexible node in the hierarchy. It resolves the entity’s role based on its metadata:

• If it’s at the top of the hierarchy with publishing rights → N-able
• If it resells to others → Distributor
• If it delivers services → MSP
• If it receives services → Tenant

Service Organization

Canonical Term: Service Organization

Definition:
A Service Organization is a hierarchical container used within on-premises N-central deployments to group and manage multiple customer environments under a single Partner (MSP or IT Pro). It acts as an intermediary layer between the Partner and the Customer (Tenant) and allows for administrative separation, reporting, and delegation of services across logically distinct client accounts. Service Organizations are primarily used for internal structuring of accounts, enabling MSPs to manage multiple customers without cross-contamination of access or policy.

Recognized Synonyms:

• Sub-Customer (legacy or colloquial)
• Organization Unit (in internal structuring discussions)
• Client Organization (context-dependent)
• Department or Branch (when mapping internal clients or sites)

Relationships:

• Service Organization isManagedBy Partner
• Service Organization contains Customer (Tenant)
• Service Organization groups Devices, Policies, and Users
• Service Organization inherits or overrides Policies from Partner
• Service Organization isLoggedInTo by AssignedUser or Admin

Prompt Examples:

• “List all service organizations under ‘CoreTech Systems’.”
• "Deploy the updated monitoring policy to all customers in the ‘Gov Services’ service organization.”
• “Which users have access to more than one service organization within ‘SecureOps MSP’?”

Agent Execution:
When a prompt includes the term “Service Organization,” the AI Agent will:

• Resolve the organizational hierarchy: Partner → Service Organization → Customer
• Identify all objects (devices, users, policies) scoped to the Service Organization
• Apply or evaluate configuration or automation at the Service Org level
• Enforce boundary conditions (e.g., policy inheritance, user access isolation)
• Return filtered results or execute actions within the correct hierarchy node

Recommendation:
Use Service Organization when referring to an internal logical grouping layer in N-central’s on-prem environments. Always specify:

• The owning MSP or IT Pro
• Whether you are targeting all customers within the org or a specific subset
• The scope of action (reporting, policy application, access control)

Site

Canonical Term: Site

Definition:
A Site is a physical or logical subdivision within a Tenant (Customer). It may represent a physical location (e.g., office, data center), a virtual group (e.g., cloud region, department), or any operational boundary used for organizing and managing Devices | Assets. Sites provide structure for automation, reporting, and policy scoping beneath the customer level.

Recognized Synonyms:

• Location
• Office (when physical)
• Department (when logical)
• Subtenant (rare — avoid)

Relationships:

• Site belongsTo Tenant
• Site contains Device | Asset
• Site has Policy
• Site mayContain Subsite (if nested modeling is supported)

Prompt Examples:

• “Check if the New York site for ‘Harbor Industries’ has any offline devices.”
• “Apply the new compliance policy to all sites in the customer ‘CityGov’.”
• “List software installed on devices in the Finance department site.”

Agent Execution:
When a prompt refers to a “Site,” the agent:

• Resolves the site within the scope of the specified or inferred Tenant
• Filters Devices, Policies, and Scripts to those associated with the Site
• Applies actions (e.g., scans, updates, reports) only to entities within the Site boundary
• If no Site is specified, defaults to all Sites under the Tenant unless scoped otherwise

Location

Canonical Term: Location

Definition:
Location refers to the physical or geographic place associated with an asset, user, or organization. It can represent where a Device is physically deployed, where a User typically operates, or where a Site is situated. While similar to a Site, “Location” focuses strictly on geography (e.g., city, region, country, GPS), whereas Site can be a logical grouping. Location data is useful for geofencing, compliance, asset tracking, threat analysis, and assigning regional policies.

Recognized Synonyms:

• Geo (informal)
• Physical Address
• Office Location
• Region (when scoped to country-level or zones)
• Area (context-dependent — avoid unless defined)

Relationships:

• Device isLocatedAt Location
• Site isLocatedAt Location
• User operatesFrom Location
• Location belongsTo Tenant
• Location mayHave RiskProfile or ComplianceZone
• Location influences PolicyScope

Prompt Examples:

• “List all devices currently located in Germany.”
• “Apply the EU compliance policy to locations within the EMEA region.”
• “Which locations show unusual login activity after business hours?”

Agent Execution:
When a prompt refers to a “Location,” the agent will:

• Resolve the geographic identifier (e.g., country, city, GPS tag, custom label)
• Filter relevant entities (devices, users, sites) assigned to or operating from that location
• Consider location-aware policies (e.g., regional compliance rules, threat intelligence mappings)
• Enable geolocation-sensitive workflows, alerts, or risk analysis
• Return context-specific responses, such as compliance status, threat levels, or regional usage trends

Tenant (customer)

Canonical Term: Tenant

Definition:
A Customer is an organization or individual that receives IT services from an MSP. In technical terms, this is represented as a Tenant — a logically isolated environment within an MSP’s management system. Each Tenant contains its own devices, users, policies, and configuration, and does not have visibility into other tenants. “Customer” is the business-facing term; “Tenant” is the system-facing canonical term.

Recognized Synonyms:

• Customer
• Client
• End Customer
• Site (legacy or context-specific)

Relationships:

• Tenant isManagedBy MSP
• Tenant has Device
• Tenant has Sites
• Tenant has Policy
• Tenant isMonitoredBy Agent
• Tenant has SLA or ServiceAgreement
• Tenant belongsTo Distributor (indirectly, via MSP)

Prompt Examples:

• “Which customers haven’t checked in within the last 24 hours?”
• “Apply the new antivirus policy to all tenants in the finance sector.”
• “List all devices owned by the customer ‘Lighthouse Legal Group’.”

Agent Execution:
When the prompt includes “Customer,” the agent maps this to a Tenant object in the system. It will:

• Resolve the correct tenant based on name or ID
• Scope queries or actions to that tenant’s isolated environment
• Retrieve or apply policies, scripts, or configurations owned by the MSP
• Enforce data and action boundaries so one tenant cannot affect another

Integration

Canonical Term: Integration

Definition:
An Integration is a defined connection between an external vendor’s product or service and one or more N-able products. Integrations allow data exchange, automation, event handling, and workflow coordination across systems. They may be developed by N-able, by the external vendor, or by a partner, and can vary in complexity from simple scripts and API calls to deeply embedded multi-function toolsets.

Recognized Synonyms:

• Plugin (legacy or context-specific)
• Extension
• Integration Module

Relationships:

• Integration connects External Vendor to Product
• Integration exposes Action
• Integration calls N-able Action
• Integration consumes Data or Event from External Vendor
• Integration mayBeManagedBy N-able | Vendor | Partner

Prompt Examples:

• “What integrations are available for ticketing systems like Autotask and ConnectWise?”
• “Enable the Bitdefender integration for all MSPs under ‘Global IT Solutions’.”
• “Does the SentinelOne integration support triggering remediation actions from within N-central?”

Agent Execution:
When a prompt refers to an “Integration,” the agent will:

• Identify the integration metadata from the N-able Integration Catalog
• Resolve the associated products, available actions, and required permissions
• Determine whether the integration is installed, active, or requires onboarding
• Enable or invoke integration-specific functionality, such as data syncs, policy mappings, or scripted triggers
• Route follow-up actions to the proper toolset (e.g., PSA ticketing, security alert feed, reporting system)

External Vendor

Canonical Term: ExternalVendor

Definition:
An External Vendor is a third-party company that provides software, hardware, or services which integrate with N-able products. These vendors are typically integration partners and may offer functionality in areas such as security, backup, PSA (Professional Services Automation), billing, remote access, or threat detection. External Vendors participate in the N-able ecosystem by exposing APIs, actions, or event streams that can be called by or respond to AI Agents, scripts, and workflows.

Recognized Synonyms:

• Vendor
• Integration Partner
• Third-Party Provider
• Tool Vendor (context-specific)

Relationships:

• ExternalVendor provides Integration
• ExternalVendor owns ExternalSystem
• ExternalVendor supports Action
• ExternalVendor mayBeCertifiedBy N-able
• ExternalVendor collaboratesWith MSP
• ExternalVendor integratesWith Product

Prompt Examples:

• “What actions are available from the external vendor ‘SentinelOne’ through its integration?”
• “Is the external vendor ‘Acronis’ certified for cloud backup integrations with Cove?”
• “Which external vendors support remediation workflows via API?”

Agent Execution:
When a prompt includes the name of an External Vendor, the agent will:

• Search the N-able Integration Registry for known vendors
• Retrieve associated integrations, available actions, and certified products
• Determine integration status (enabled, not installed, in error) for the current MSP or Tenant
• Route API calls or automation flows to the appropriate external system, respecting authentication and permission boundaries
• Return actionable metadata such as supported event types, playbooks, or remediation capabilities

Asset (Device)

Canonical Term: Asset

Definition:
An Asset refers to any physical or logical entity managed by portfolio products. This includes, but is not limited to, endpoints (e.g., laptops, desktops, servers), edge systems, mobile devices, cloud resources, Entra (Azure AD) identities, and Microsoft 365 services. Assets represent the lowest-level managed object within a hierarchy. Use the canonical term Asset instead of “Device,” which is a limited synonym. Assets can exist in various environments—on-prem, cloud, hybrid—and across tenants, sites, and service organizations.

Recognized Synonyms:

• Device
• Endpoint
• Node (legacy)
• Laptop (ambiguous — use only when specifically identifying form factor)

Relationships:

• Asset has Policy
• Asset belongsTo Tenant
• Asset isMonitoredBy Agent
• Asset runs Script
• Asset hasStatus HealthState
• Asset contains Software
• Asset has RiskScore

Prompt Examples:

• “Reboot the asset named ‘ENG-WORKSTATION-03’ if it's unresponsive.”
• “What’s the patch compliance status of all assets in the ‘Sales’ department?”
• “Run the disk cleanup script on all Windows 10 assets with over 80% disk usage.”

Server

Canonical Term: Server

Definition:
A Server is a physical or virtual computing resource designed to provide services, applications, or data to other devices or clients. In the DevOps Infrastructure Ontology, a Server is a subclass of devopscore:Resource and is characterized by its role in hosting software components, managing workloads, and facilitating networked operations. Servers are integral to IT infrastructures, supporting various functions such as web hosting, database management, and application deployment.

Recognized Synonyms:
• Host
• Node (context-dependent)
• Compute Instance
• VM (Virtual Machine)
• Physical Server

Relationships:
• Server hosts SoftwareComponent
• Server hasHardwareComponent Hardware
• Server isLocatedIn DataCenter
• Server connectsTo Network
• Server dependsOn PowerSupply
• Server isManagedBy SystemAdministrator

Prompt Examples:
• “List all servers hosting the 'Customer Relationship Management' application.”
• “What is the current CPU utilization of servers located in the 'East Coast Data Center'?”
• “Deploy the latest security patch to all web servers running Apache HTTP Server.”

Agent Execution:
When processing prompts involving "Server," the AI agent will:

1. Identify Server Instances:
   o Query the infrastructure inventory to retrieve all entities classified under the Server category.
2. Filter Based on Attributes:
   o Apply filters such as location (isLocatedIn), hosted software (hosts), or hardware specifications (hasHardwareComponent) to narrow down the relevant servers.
3. Perform Actions:
   o Execute the requested operations, such as retrieving performance metrics, deploying software updates, or modifying configurations on the identified servers.
4. Ensure Compliance and Logging:
   o Verify that actions comply with organizational policies and log activities for auditing purposes.

Account

Canonical Term: (Avoid using "Account"; use a more precise term such as Tenant, MSP, Distributor, or Customer)

Definition:
The term Account is discouraged due to its ambiguity. It can refer to multiple different roles in the N-able ecosystem, including:
• A Tenant (a customer managed by an MSP)
• An MSP (a partner of N-able)
• A Distributor Customer (a client of a distributor)
• A Billing Entity (used in financial systems)
Because of this ambiguity, using “Account” in prompts or documentation can cause confusion or incorrect agent actions.

Recognized Synonyms:
• Account (ambiguous – avoid)
• Client (ambiguous – avoid unless scoped)
• Partner (ambiguous – clarify as MSP or Distributor)

Relationships:
Do not use Account in relationships. Instead, use specific terms such as:
• Tenant belongsTo MSP
• MSP isPartnerOf N-able
• Customer isManagedBy Distributor
• BillingEntity paysFor Subscription

Prompt Examples (and corrections):
• INCORRECT: “Get a list of all accounts with over 100 devices.”
• CORRECT: “Get a list of all tenants with over 100 devices.”
• INCORRECT: “Which accounts are overdue on patch compliance?”
• CORRECT: “Which MSPs or tenants are overdue on patch compliance?”
• INCORRECT: “Send onboarding materials to new accounts.”
• CORRECT: “Send onboarding materials to new MSPs and distributors.”

Agent Execution:
When “Account” is used in a prompt, the AI agent will:
• Attempt to infer the intended object (Tenant, MSP, Customer, etc.)
• Prompt for clarification if multiple interpretations are possible
• If inference succeeds, resolve the term to the correct object type and execute accordingly
• This behavior increases the risk of misinterpretation and is not recommended

User

Canonical Term: User

Definition:
A User is an individual identity that interacts with or is managed by the ecosystem. The term User may refer to two distinct but related concepts. In the DevOps Infrastructure Ontology, a User is typically a human actor who interacts with resources and services, often through a UI, command line, or API. A user may represent different roles depending on the context, including an MSP administrator, IT technician, end user (someone logging into a managed device), or N-able support personnel. A user may also be a leaf node representing a person managed as part of a customer environment. Users can be associated with specific tenants, organizations, roles, or permissions, and their activities may be logged, audited, and constrained by policies.
It is not recommended to use the term User without additional context.

Recognized Synonyms:
• Admin (when referring to privileged users)
• Technician (for operational users)
• End User (device or service consumer)
• Operator (in automation/DevOps contexts)
• Identity (generic — use carefully)

Relationships:
• User isAssignedTo Tenant, MSP, or Site
• User hasRole (e.g., Admin, Viewer, Operator, Support)
• User isAuthenticatedBy IdentityProvider or DirectoryService
• User performs Action or Task
• User accesses Product or Platform
• User isLoggedInVia Asset | Device or WebSession

Prompt Examples:
• “List all users with admin access to tenant ‘BlueNet Solutions’.”
• “Has the end user ‘alex.jordan’ logged in from multiple locations recently?”
• “Disable the user account for ‘tech.msmith’ across all N-central-managed environments.”

Agent Execution:
When a prompt refers to a “User,” the AI Agent will:
• Resolve the identity context (e.g., MSP admin, end user, support tech)
• Look up associated permissions, roles, tenant ownership, and login history
• Take actions such as listing access, modifying roles, disabling users, or flagging suspicious activity
• Use multi-tenant scoping rules to ensure actions are authorized and logged appropriately
• Disambiguate prompts like “user logged into the asset” vs. “user with N-able product access”

Template

Canonical Term: Template

Definition:
A Template is a reusable configuration object that defines a set of settings, rules, or characteristics to be applied to one or more assets, such as devices, scripts, users, or policies. Templates are commonly used to enforce consistency across environments and to streamline the deployment of standardized configurations. In N-able products, templates may be used for onboarding, automation, security settings, or monitoring configurations.

Recognized Synonyms:
• Profile
• Blueprint (informal)
• Configuration Template
• Policy Template (specific case – see separate term)

Relationships:
• Template appliesTo Device
• Template defines Settings
• Template isOwnedBy MSP or N-able
• Template mayBeAssignedTo Site or Tenant
• Template isBasedOn Policy or Script

Prompt Examples:
• “Apply the Windows onboarding template to all new laptops at the Boston site.”
• “List all templates owned by the MSP ‘CoreNet Services’.”
• “Does the antivirus compliance template include real-time protection settings?”

Agent Execution:
When a prompt refers to a “Template,” the AI agent will:
• Retrieve the template definition, including its scope, settings, and owner
• Determine eligible targets (e.g., Devices, Sites, Tenants)
• Apply the settings defined in the template to those targets, unless overridden by policy exceptions
• Log any deviations or errors during application

Resource

Canonical Term: Resource

Definition:
A Resource is a general term that refers to any manageable object within the N-able ecosystem. This includes physical or virtual components such as devices, users, software, network elements, storage, and compute instances. In internal models (and in the DevOps Infrastructure Ontology), “Resource” is often used as an abstract class or grouping term for things that can be monitored, configured, automated, or queried. It is not recommended for use in AI prompts due to its vagueness.

Recognized Synonyms:
• Asset (if referring to devices – use “Device” instead)
• Object (generic – avoid unless technically scoped)
• Entity (context-specific – avoid in human prompts)

Relationships:
• Resource mayBe a Device, Server, User, Process, Script, or NetworkComponent
• Resource has Status
• Resource isManagedBy Agent
• Resource belongsTo Site or Tenant
• Resource isTaggedWith Label or Category

Prompt Examples:
• INCORRECT: “Check the health of all resources in the customer ‘Brightline’.”
• CORRECT: “Check the health of all devices in the customer ‘Brightline’.”
• INCORRECT: “What resources are consuming the most bandwidth?”
• CORRECT: “What devices or servers are consuming the most bandwidth?”

Agent Execution:
When the term “Resource” is used, the agent:
• Attempts to infer the intended object type based on context (e.g., device, server, user)
• Groups results across multiple types unless otherwise specified
• May return an ambiguous or overly broad result, reducing the usefulness of the output
• Logs a semantic warning if the target object class is unclear

Recommendation:
Avoid using the term “Resource” in end-user prompts or agent-facing commands. Instead:
• Use Device, Server, Script, Policy, User, etc., to specify the object type
• Reserve “Resource” for internal system modeling or development documentation where abstraction is required

Data

Canonical Term: Data
(Use with specificity—prefer concrete terms like Log, Metric, Event, or Configuration when possible)

Definition:
Data refers to any structured or unstructured information collected, stored, or generated by N-able systems and integrated tools. This includes telemetry, logs, metrics, configuration records, alerts, ticket history, and external feeds. While "Data" is technically accurate, it is too broad for effective agent interaction. Prompts should specify the type of data required to ensure accurate retrieval or action.

Recognized Synonyms:
• Information (too vague – avoid)
• Output (context-specific – use with care)
• Result (use for returned data from a function or query)

Relationships:
• Data isGeneratedBy N-able, MSP, Device, Script, or Integration
• Data isStoredIn Repository or Platform
• Data isConsumedBy Agent or Automation or Policy
• Data isTaggedWith Type (e.g., metric, alert, configuration)
• Data belongsTo N-able or MSP or Tenant or Site

Prompt Examples:
• INCORRECT: “Get all data for this tenant.”
• CORRECT: “Get performance metrics and alert logs for this tenant.”
• INCORRECT: “Analyze the data for policy violations.”
• CORRECT: “Analyze event logs and compliance reports for policy violations.”
• CORRECT: “Retrieve CPU and memory usage data for device ‘ENG-LT-05’ for the past 7 days.”

Agent Execution:
When “Data” is used in a prompt, the AI agent will:
• Attempt to infer the correct data type based on additional context (e.g., time range, source, object)
• Search available data sources (e.g., telemetry, logs, monitoring tools)
• Return multiple types if ambiguity remains, or issue a request for clarification
• Default to commonly accessed types (e.g., metrics and logs) if unspecified

Recommendation:
Avoid using the general term “Data” in prompts to AI Agents unless the context is unambiguous. Instead, specify:
• Metrics – for performance, usage, thresholds
• Logs – for events, errors, audit trails
• Alerts – for triggered thresholds or anomalies
• Configuration – for settings, policy states, software versions
• Ticket Data – for incident and service histories

Cloud

Canonical Term: Cloud
Definition:

Cloud refers to infrastructure, platforms, or services hosted in a remote, internet-accessible environment rather than on-premises. In the N-able context, "Cloud" is typically used to describe:
• N-able’s own cloud-native products (e.g., Cove, Cloud Commander, NC Hosted)
• Third-party integrations with public cloud providers (e.g., AWS, Azure, GCP)
• The hosting location or operational model for managed assets, policies, or storage
Because the term is broad, it should be accompanied by a qualifier (e.g., Cloud Backup, Cloud Integration, Cloud Device) to ensure clarity.

Recognized Synonyms:
• SaaS Platform (when referring to N-able-hosted tools)
• Hosted Environment
• Public Cloud (AWS, Azure, etc.)
• Cloud Service Provider (CSP)

Relationships:
• Cloud hosts Service or Product
• Cloud contains Asset or Instance
• Cloud isIntegratedWith N-able or ExternalVendor
• Cloud provides Storage or Compute
• Cloud runs Automation or ScheduledTask
• Cloud isAccessedBy Agent or User

Prompt Examples:
• INCORRECT: “Is this customer in the cloud?”
• CORRECT: “Is this customer using N-able Cove Cloud Backup?”
• INCORRECT: “Scan all cloud resources for vulnerabilities.”
• CORRECT: “Scan all Azure-hosted virtual machines for vulnerabilities.”
• CORRECT: “List all devices managed through Cloud User Hub that belong to tenant ‘BlueSteel LLC’.”

Agent Execution:
When a prompt includes “Cloud,” the AI agent will:

• Attempt to classify the intended meaning:
  • Is it referencing a product (e.g., Cove)?
  • A hosting model (e.g., AWS-based device)?
  • A data source (e.g., cloud backup logs)?
• Resolve actions or queries using the correct integration, platform, or cloud service context
• Validate scope: Is the object hosted in the cloud? Is it managed via a cloud product?
• Return results filtered by cloud product, provider, or hosting model

Software

Canonical Term: Software

Definition:
Software refers to any executable program, application, or system component that performs functions on a computing platform. In the DevOps Infrastructure Ontology, Software is a subclass of devopscore:Resource and encompasses entities such as operating systems, applications, services, and scripts. Software can be installed on physical or virtual devices and is essential for enabling various IT operations and services.

Recognized Synonyms:
• Application
• Program
• Service (when referring to software-based services)
• Package (in the context of software packages)

Relationships:
• Software isInstalledOn Device
• Software dependsOn other Software
• Software isManagedBy Agent
• Software canBe Patched
• Software hasVersion Version
• Software isConfiguredBy Configuration

Prompt Examples:
• “List all software installed on device ‘ENG-LT-05’.”
• “Update the antivirus software to the latest version on all servers.”
• “Check if the CRM application is running on the New York site.”

Agent Execution:
When a prompt refers to “Software,” the AI agent will:
• Identify the specific software entity based on name, version, or other attributes.
• Determine the devices or environments where the software is installed.
• Perform actions such as installation, update, configuration, or status checks.
• Ensure that dependencies and configurations are managed appropriately.

Package (see also Patch)

Canonical Term: Package

Definition:
A Package is a discrete unit of software that can be installed, upgraded, or removed on a device or system. It typically contains executable code, configuration files, and metadata needed to deploy software. Packages are distributed through package managers (e.g., MSI, RPM, DEB, EXE) and may represent full applications, libraries, agents, or drivers. In MSP environments, packages are often deployed via scripts, automation workflows, or remote management tools.
Code which is an update may also be referred to as a Patch.

Recognized Synonyms:
• Installer
• Module (when part of a larger software system)
• Software Package
• (UNCLEAR, but here we seem to use it interchangeably with PATCH)
• Deployment Artifact (technical)
• Patch (when part of an explicit update to deployed software)

Relationships:
• Package installs Software
• Package isDeployedTo Device
• Package isManagedBy Script or Automation Policy (or Template?)
• Package hasVersion
• Package isSignedBy Vendor
• Package isHostedIn Repository

Prompt Examples:
• “Install the package for Microsoft Edge on all Windows 11 devices.”
• “Which packages are waiting to be deployed on the ‘Sales-Laptops’ site?”
• “Has the antivirus agent package been successfully installed on device ‘HR-LT-002’?”

Agent Execution:
When a prompt refers to a “Package,” the AI agent will:
• Look up package metadata, including name, version, and vendor signature
• Validate compatibility with the target device (e.g., OS, architecture)
• Trigger installation, upgrade, rollback, or removal actions
• Log the outcome and update inventory or compliance records

File

Canonical Term: File

Definition:
A File is a discrete unit of digital data stored on a device or in the cloud. Files can represent executables, documents, logs, scripts, configurations, or any other form of structured or unstructured content. In the N-able context, files may be transferred, monitored, scanned for threats, or used as part of automation workflows. Files exist within file systems, storage volumes, or cloud containers, and may be associated with software, devices, or users.

Recognized Synonyms:
• Document (when referring to user-generated content)
• Artifact (in deployment or CI/CD contexts)
• Payload (in scripting or integration contexts)
• Object (in cloud storage – clarify when used)

Relationships:
• File isStoredOn Device or CloudStorage
• File isScannedBy AVEngine or SecurityTool
• File isGeneratedBy Script or Application
• File isModifiedBy User or Agent
• File isTaggedWith Label or Policy
• File isMonitoredBy Agent

Prompt Examples:
• “Scan all executable files on server ‘NY-EX01’ for known threats.”
• “Upload the installation file to the shared repository and notify the admin.”
• “Retrieve the most recent log file from the device ‘ENG-LT-05’.”

Agent Execution:
When a prompt refers to a “File,” the AI agent will:
• Resolve the file location (e.g., local device, cloud share, log directory)
• Perform the specified action (e.g., retrieve, scan, delete, move, check integrity)
• Use file metadata such as name, type, size, timestamp, and hash to identify the target
• Log actions and validate policy or security rules (e.g., blocking executable uploads in restricted environments)

Patch

Canonical Term: Patch

Definition:
A Patch is a specific update package intended to correct known issues, fix vulnerabilities, or improve functionality in existing software. Patches may be distributed by vendors as part of regular update cycles (e.g., Microsoft Patch Tuesday), or released ad hoc in response to critical security flaws. In the MSP context, patches are applied across multiple tenants, devices, or sites using automation policies, schedules, or manual triggers.

Recognized Synonyms:
• Update (use only when not referring to feature upgrades)
• Hotfix
• Security Fix
• Software Update (general)

Relationships:
• Patch isAppliedTo Device or Software
• Patch isPublishedBy Vendor
• Patch isPartOf UpdatePolicy
• Patch isMonitoredBy ComplianceEngine
• Patch hasSeverity Level
• Patch isScheduledBy AutomationPolicy

Prompt Examples:
• “Apply all high-severity patches to servers in the ‘Finance’ tenant before Friday.”
• “What patches failed to install on the devices at the Boston site?”
• “Generate a report of missing security patches for all Windows 11 laptops.”
• “Patch the application Foo immediately for tenant Jonson & Jonson”

Agent Execution:
When a prompt refers to a “Patch,” the AI agent will:
• Identify the relevant patch based on vendor, product, severity, or classification
• Determine installation status on applicable devices
• Schedule or trigger patch installation (immediate, deferred, or staged)
• Log results, including success/failure per device, and update compliance reports
• If applicable, check for required reboots or follow-up tasks

Vulnerability

Canonical Term: Vulnerability

Definition:
A Vulnerability is a weakness or flaw in software, hardware, or configuration that could be exploited to compromise system security, data integrity, or availability. Vulnerabilities are typically assigned a unique identifier (e.g., CVE ID), a severity score (e.g., CVSS), and may be associated with known exploits. In the MSP context, vulnerabilities are detected through scans, threat intelligence feeds, or vendor advisories and are often tracked and mitigated using patching, configuration changes, or compensating controls.

Recognized Synonyms:
• CVE (Common Vulnerabilities and Exposures identifier – use when known)
• Security Flaw
• Exploit (note: an exploit is a tool/technique, not the vulnerability itself)
• Weakness (less specific – avoid unless referring to misconfiguration)

Relationships:
• Vulnerability affects Software or Device
• Vulnerability isIdentifiedBy VulnerabilityScanner or ThreatFeed
• Vulnerability hasSeverity Score
• Vulnerability isRemediatedBy Patch or ConfigurationChange

Prompt Examples:
• “List all critical vulnerabilities detected on devices in the ‘Healthcare West’ tenant.”
• “Has the CVE-2024-23978 vulnerability been patched across all cloud servers?”
• “Generate a vulnerability report for third-party software used in finance-related sites.”

Agent Execution:
When a prompt refers to a “Vulnerability,” the AI agent will:
• Look up the vulnerability using known identifiers (e.g., CVE IDs) or descriptive keywords
• Query vulnerability scan results or correlated security event data
• Map the vulnerability to affected software, devices, or configurations
• Check remediation status: patched, mitigated, or unaddressed
• Recommend or trigger remediation actions if authorized (e.g., apply patch, isolate device)

Hardware

Canonical Term: Hardware

Definition:
Hardware refers to the physical components of an IT environment, including servers, laptops, desktops, networking gear, storage devices, and peripherals. In the N-able context, hardware is typically associated with managed Devices, and its properties (e.g., CPU type, RAM, disk size, serial number) are collected for inventory, health monitoring, and compliance tracking. Hardware may also include embedded systems and edge devices under MSP management.

Recognized Synonyms:
• Physical Asset
• Equipment
• Machine (context-specific – avoid)
• Host (when referring to physical servers)

Relationships:
• Hardware isPartOf Device
• Hardware hasComponent CPU, Memory, Disk, NIC, etc.
• Hardware isLocatedAt Site
• Hardware isMonitoredBy Agent
• Hardware hasWarrantyStatus or LifecycleStage
• Hardware supports Software

Prompt Examples:
• “List all hardware with less than 8GB RAM across the ‘Education’ tenant.”
• “What hardware components are failing on server ‘NY-DC-001’?”
• “Generate a hardware inventory report for all remote offices.”

Agent Execution:
When a prompt refers to “Hardware,” the AI agent will:
• Query the device inventory or telemetry sources to extract physical specs
• Identify hardware failures, warnings, or threshold violations
• Associate the hardware with the parent device and customer context
• Report lifecycle status (e.g., warranty expired, asset due for replacement)
• Trigger alerts or automations if hardware is underperforming or failing

Recommendation:
Use Hardware to refer to the physical layer of infrastructure. When possible, specify the component:
• Use CPU, Disk, Memory, Motherboard, Power Supply, etc.
• Differentiate between physical and virtual resources when managing hybrid environments
• Avoid vague phrases like “machine info” or “asset details” when directing AI agent actions

Disk

Canonical Term: Disk

Definition:
A Disk is a physical or virtual storage component used by a device to persist data. Disks can be hard drives (HDD), solid-state drives (SSD), or virtual volumes (e.g., VHDs in cloud environments). In the N-able context, disk information is collected for monitoring storage capacity, performance, health, and lifecycle. Disks may be local (installed on a device) or network-attached, and are critical for system stability and application functionality.

Recognized Synonyms:
• Drive
• Storage Device
• Volume (when partitioned or virtualized)
• HDD / SSD (specific types of disks)

Relationships:
• Disk isPartOf Device
• Disk hasCapacity Size
• Disk hasUsedSpace / FreeSpace
• Disk hasHealthStatus
• Disk isMonitoredBy Agent
• Disk stores File or Software

Prompt Examples:
• “Check disk usage on all servers in the ‘Legal’ tenant.”
• “Alert me when any disk has less than 10% free space.”
• “List all devices with SSDs larger than 512GB.”

Agent Execution:
When a prompt refers to a “Disk,” the AI agent will:
• Retrieve storage information from device telemetry or system APIs
• Report on total capacity, free space, used percentage, and disk type
• Monitor for thresholds (e.g., low space, high I/O wait times)
• Trigger alerts or automated cleanup scripts as needed
• Differentiate between physical disks and logical partitions/volumes

Firewall

Canonical Term: Firewall

Definition:
A Firewall is a security system—either hardware-based, software-based, or a combination of both—that monitors and controls incoming and outgoing network traffic based on predetermined security rules. It serves as a barrier between trusted internal networks and untrusted external networks, such as the internet, to prevent unauthorized access and threats. In the DevOps Infrastructure Ontology, a firewall is categorized under network infrastructure components, emphasizing its role in network security management.

Recognized Synonyms:
• Network Firewall
• Security Gateway
• Packet Filter
• Application Firewall

Relationships:
• Firewall controlsAccessTo NetworkSegment
• Firewall isDeployedOn Device
• Firewall hasRuleSet AccessControlList
• Firewall monitorsTrafficFor Threat
• Firewall isManagedBy SecurityPolicy
• Firewall logsEventsTo MonitoringSystem

Prompt Examples:
• “List all firewalls deployed across the 'Finance' tenant's network.”
• “Check the current rule set configured on the firewall protecting the 'NYC-DataCenter'.”
• “Identify any unauthorized access attempts blocked by the firewall in the last 24 hours.”

Agent Execution:
When a prompt refers to a “Firewall,” the AI agent will:
• Identify the specific firewall instance(s) based on the context provided (e.g., tenant, location, device).
• Retrieve and analyze the firewall's configuration, including rule sets and access control lists.
• Monitor and report on traffic patterns, highlighting any anomalies or blocked access attempts.
• Suggest or implement updates to firewall rules in response to detected threats or policy changes, if authorized.

Performance

Canonical Term: Performance

Definition:
Performance refers to the operational efficiency and responsiveness of a system, device, or application. It is typically measured through metrics such as CPU usage, memory consumption, disk I/O, network throughput, and latency. In the MSP and N-able context, performance is used to assess the health of devices, detect bottlenecks, and determine whether systems are operating within acceptable thresholds.

Recognized Synonyms:
• System Health (ambiguous – use when referring to overall status)
• Responsiveness (context-specific – use when referring to end-user experience)
• Efficiency (too vague – avoid in prompts)

Relationships:
• Performance isMeasuredBy Metric
• Performance affects UserExperience
• Performance isMonitoredBy Agent or MonitoringPolicy
• Performance isReportedIn Dashboard or Report
• Performance triggers Alert or Automation

Prompt Examples:
• “Check the performance of device ‘ENG-LT-05’ and flag any CPU or memory spikes.”
• “Generate a performance report for all servers in the ‘West Coast’ site over the last 7 days.”
• “What performance metrics are being tracked for the tenant ‘SecureNet’?”

Agent Execution:
When a prompt refers to “Performance,” the AI agent will:
• Retrieve relevant metrics such as CPU load, memory usage, disk activity, and network traffic
• Evaluate current values against historical baselines and configured thresholds
• Detect anomalies or degradation (e.g., sustained high CPU, low available RAM)
• Recommend or trigger remediation actions (e.g., restart service, notify admin, scale resources)
• Optionally summarize trends in reports or dashboards

Recommendation:
Use Performance when you want a general assessment across multiple resource categories. For specific actions or automation, refer directly to the metric type:
• CPU usage, Memory utilization, Disk I/O, Network latency, etc.
• Avoid vague terms like “slowness” unless you pair it with a concrete object or metric.

Warranty

Canonical Term: Warranty
Definition:
A Warranty refers to the contractual period during which a hardware asset (e.g., server, laptop, disk) is covered for repairs, replacements, or support by the manufacturer or vendor. In the MSP context, warranty information helps inform lifecycle management, asset planning, and service prioritization. Tracking warranty status enables technicians to escalate hardware issues appropriately and manage risk around unsupported devices.
Recognized Synonyms:
• Hardware Coverage
• Support Agreement (may overlap with broader services – use carefully)
• Manufacturer Service Contract
Relationships:
• Warranty isAssociatedWith Hardware or Device
• Warranty has StartDate and EndDate
• Warranty isProvidedBy Vendor or Manufacturer
• Warranty affects ReplacementPlan or LifecycleStage
• Warranty isTrackedIn AssetInventory
Prompt Examples:
• “Which devices have warranties expiring in the next 30 days?”
• “Is the disk in device ‘HR-LT-004’ still under warranty?”
• “List all out-of-warranty servers in the ‘EMEA’ region.”
Agent Execution:
When a prompt includes “Warranty,” the AI agent will:
• Look up asset metadata to determine the warranty start and end dates
• Flag any expired or soon-to-expire warranties
• Associate warranty status with asset age, performance issues, or incident trends
• Suggest replacement recommendations or notify stakeholders based on policy
• Optionally pull warranty verification from vendor portals (if integrated)

Certificate

Canonical Term: Certificate

Definition:
A Certificate is a digital credential that authenticates the identity of an entity (such as a user, device, or service) and facilitates secure communication through encryption. Certificates are issued by a Certificate Authority (CA) and are integral to Public Key Infrastructure (PKI). They ensure data integrity, confidentiality, and authenticity in digital communications.

Recognized Synonyms:
• Digital Certificate
• SSL/TLS Certificate
• X.509 Certificate
• Public Key Certificate

Relationships:
• Certificate isIssuedBy Certificate Authority
• Certificate isAssignedTo Device or User
• Certificate isUsedFor Authentication or Encryption
• Certificate hasValidityPeriod StartDate and EndDate
• Certificate isRevokedBy Certificate Authority
• Certificate isStoredIn Certificate Repository

Prompt Examples:
• “List all certificates expiring within the next 30 days.”
• “Check if the SSL certificate for 'client-portal.example.com' is valid.”
• “Revoke the compromised certificate assigned to device 'ENG-LT-05'.”

Agent Execution:
When a prompt includes “Certificate,” the AI agent will:
• Identify the specific certificate(s) based on context (e.g., domain name, device ID).
• Retrieve certificate details such as issuer, subject, validity period, and status.
• Perform actions like validation checks, renewal processes, or revocation procedures.
• Update relevant systems or repositories with the current certificate status.

Product

Canonical Term: Product

Definition:
A Product is a tangible or intangible item offered by a vendor, encompassing software applications, hardware devices, or services. In the DevOps Infrastructure Ontology, a product is an entity that can be deployed, configured, and managed within an IT environment. Products are often associated with specific versions, configurations, and may be part of a larger service offering.

Recognized Synonyms:
• Offering
• Solution
• Application (when referring to software products)
• Service (in the context of service products)

Relationships:
• Product isProvidedBy Vendor
• Product hasVersion Version
• Product isDeployedOn Device or Infrastructure
• Product isConfiguredBy Configuration
• Product isPartOf ServiceOffering
• Product hasDependency on other Products

Prompt Examples:
• “List all products deployed in the 'Healthcare' tenant.”
• “Check the current version of the antivirus product on all endpoints.”
• “Identify any products with known vulnerabilities in the 'Finance' department.”

Agent Execution:
When a prompt refers to a “Product,” the AI agent will:
• Identify the specific product(s) based on name, version, or associated vendor.
• Retrieve deployment and configuration details across the infrastructure.
• Assess compliance with organizational policies and identify any discrepancies.
• Provide recommendations for updates, patches, or reconfigurations as needed.

Recommendation:
Use Product when referring to distinct offerings from vendors that are deployed and managed within the IT environment. Be specific by including:
• The product name and version.
• The associated vendor or provider.
• The deployment context (e.g., which devices or services it's part of).

Service

Canonical Term: Service

Definition:
A Service is a functional capability delivered through software, infrastructure, or human operations. In the DevOps Infrastructure Ontology, a service may be a system-level process (e.g., a daemon or microservice), a cloud-hosted API, or a business-facing offering (e.g., backup service, monitoring service). In the MSP context, services may be IT services provided to tenants (e.g., endpoint protection), or software-based services running on devices or in the cloud.

Recognized Synonyms:
• System Service (for processes running on devices)
• Managed Service (for offerings delivered by MSPs)
• API Service (when referring to remote callable interfaces)
• Feature (in software product context — use with caution)

Relationships:
• Service isHostedOn Device or Cloud
• Service isProvidedBy Product or Vendor
• Service isMonitoredBy Agent
• Service isPartOf ServiceOffering or SLA
• Service dependsOn Software or Configuration
• Service isConsumedBy User or Tenant

Prompt Examples:
• “Check if the backup service is running on all servers in the ‘Legal’ tenant.”
• “List all services provided to the customer ‘BrightPath Health’.”
• “Which services are failing health checks in the ‘EU Data Center’?”

Agent Execution:
When a prompt refers to a “Service,” the AI agent will:
• Determine whether the service is a system-level process or a business-level offering
• Identify the devices, tenants, or environments associated with that service
• Query service status (running, failed, degraded, stopped)
• Correlate with configuration, dependencies, or SLA status
• Take remediation steps if authorized (e.g., restart, notify, escalate)

Recommendation:
Use Service when referring to ongoing functional components — whether system-based (e.g., “AV scan service”) or offering-based (e.g., “managed patching service”).
• Clarify the scope: running software or delivered offering
• Include name, location, and context (e.g., product, tenant, device)
• Avoid overloading the term “service” for unrelated tasks or generic support

Script

Canonical Term: Script

Definition:
A Script is a predefined, executable set of instructions used to automate actions on devices, services, or platforms. In the MSP context, scripts are widely used for tasks such as remediation, deployment, configuration, and monitoring. Scripts can be written in various languages (e.g., PowerShell, Bash, Python) and are executed manually, on a schedule, or as part of automation policies triggered by events or conditions.

Recognized Synonyms:
• Automation Script
• Command Script
• Routine
• Playbook (in some contexts — clarify use)

Relationships:
• Script isExecutedOn Device or CloudResource
• Script isTriggeredBy Event or Schedule
• Script isOwnedBy MSP or N-able
• Script isPartOf AutomationPolicy or Workflow
• Script produces OutputData or modifies Configuration

Prompt Examples:
• “Run the disk cleanup script on all Windows servers in the ‘Finance’ tenant.”
• “List all scripts that have failed in the last 24 hours.”
• “Schedule the user offboarding script for every Friday at 6 PM.”

Agent Execution:
When a prompt refers to a “Script,” the AI agent will:
• Locate the script in the approved script repository or library
• Validate execution scope (target devices, permissions, timing)
• Trigger the script as requested (immediate, scheduled, or conditional)
• Monitor for success/failure, capture output, and log results
• Enforce execution limits or policy constraints (e.g., tenant isolation, script approval)

Recommendation:
Use Script when referring to automatable, repeatable task logic. For accuracy:
• Specify script name, language, and execution scope
• Indicate whether it's a one-time action or part of a policy
• Include execution timing or trigger conditions when relevant

License

Canonical Term: License

Definition:
A License is a contractual or system-enforced right to use a product, service, or feature within defined terms (e.g., time, scope, seats, capabilities). In the MSP and N-able context, licenses are required for using specific N-able products (e.g., N-central, Cove) or features (e.g., Take Control, advanced reporting). Licenses may apply at the tenant, device, or user level and are tracked to ensure compliance, billing accuracy, and feature access.

Recognized Synonyms:
• Subscription (context-dependent — more common in SaaS)
• Entitlement
• Access Token (avoid unless referring to security/auth)
• Usage Right

Relationships:
• License appliesTo Product, Feature, or Service
• License isOwnedBy Tenant, MSP, or Distributor
• License isEnforcedBy Platform or BillingSystem
• License isConsumedBy Device or User
• License hasValidityPeriod StartDate–EndDate
• License isTrackedIn UsageReport or BillingRecord

Prompt Examples:
• “List all licenses assigned to tenant ‘SecureNet IT’.”
• “Which devices are consuming Take Control licenses?”
• “Alert me when a license is nearing expiration or usage threshold.”

Agent Execution:
When a prompt refers to a “License,” the AI agent will:
• Look up the license object and its associated terms (type, scope, duration, limits)
• Match usage (devices, users, features) to license consumption
• Report on current usage, available seats, or overages
• Trigger notifications for nearing capacity, expiration, or violations
• Enforce policy: enable/disable features based on license state

Recommendation:
Use License when discussing entitlement to use products or services. Be specific by indicating:
• The product or feature licensed
• The entity consuming it (tenant, MSP, device)
• Whether you're checking usage, availability, or status
Avoid vague terms like “access” or “subscription” unless referring to SaaS billing explicitly.

Backup

Canonical Term: Backup

Definition:
A Backup is a secure, stored copy of data, system configuration, or application state, used to restore functionality or recover from data loss, corruption, or disaster. In the MSP and N-able context, backups may be scheduled, continuous, or event-triggered, and can target endpoints, servers, cloud services (e.g., Microsoft 365), or databases. Backup policies define the scope, frequency, retention, and storage destination of backup operations.

Recognized Synonyms:
• Data Backup
• System Image (if referring to full-machine backup)
• Snapshot (when describing state-based backups)
• Restore Point (when used for rollback)

Relationships:
• Backup isCreatedFor Device, Asset, Server, File, Database, or Service
• Backup isManagedBy BackupPolicy
• Backup isStoredIn Cove or Cloud Storage
• Backup isMonitoredBy Agent or Platform
• Backup isRestoredBy Restore Job
• Backup isOwnedBy Tenant or MSP

Prompt Examples:
• “When was the last successful backup for the server ‘NY-SQL-01’?”
• “List all failed backup jobs in the past 48 hours across the ‘Retail’ tenant.”
• “Restore the backup of user ‘jane.doe’ mailbox from last Monday.”

Agent Execution:
When a prompt refers to a “Backup,” the AI agent will:
• Locate the latest or requested backup instance based on entity, time, and scope
• Validate its status (successful, failed, in progress)
• Trigger restore actions if authorized, using the appropriate recovery method
• Check backup compliance with configured policies (frequency, retention)
• Generate audit logs or reports based on backup history and status

Storage

Canonical Term: Storage

Definition:
Storage refers to the physical or virtual infrastructure used to persist data, including backups, logs, software, configurations, and operating system files. In the MSP and N-able context, storage may include local disk drives, network-attached storage (NAS), cloud storage services (e.g., AWS S3), or specialized backup storage vaults. Storage is a critical resource that supports data retention, recovery, and compliance.

Recognized Synonyms:
• Storage Device (when physical)
• Storage Vault (in backup contexts)
• Repository
• Volume (when partitioned or virtualized)

Relationships:
• Storage contains File, Backup, or Image
• Storage isAttachedTo Device or Site
• Storage isProvisionedBy CloudProvider or MSP
• Storage isMonitoredBy Agent or StoragePolicy
• Storage hasCapacity and UsageThreshold
• Storage isUsedBy Product or Service

Prompt Examples:
• “How much storage is currently available in the backup vault for ‘Tenant-Alpha’?”
• “Alert me when storage usage exceeds 85% on any devices.”
• “List all cloud storage accounts used by the Cove backup service.”

Agent Execution:
When a prompt refers to “Storage,” the AI agent will:
• Query current capacity, usage, and health metrics from telemetry or provider APIs
• Identify the storage type (local, NAS, SAN, cloud) and associated devices or services
• Apply configured policies for thresholds, scaling, or cleanup
• Trigger alerts, archive operations, or storage expansions if needed
• Track usage by tenant, device group, or product to ensure licensing and compliance alignment

Notification

Canonical Term: Notification

Definition:
A Notification is a message or alert generated by a system, policy, or agent to inform a user, technician, or administrator of an event, condition, or required action. Notifications may be triggered by monitoring thresholds, policy violations, automation results, or system events. They can be delivered through various channels such as email, SMS, in-app alerts, or PSA integrations.

Recognized Synonyms:
• Alert (when real-time or critical — use with care)
• Message (too broad — avoid unless contextualized)
• Prompt (in user-interface contexts)
• Signal (not typically used — avoid)

Relationships:
• Notification isTriggeredBy Event, Policy, or Agent
• Notification isDeliveredTo User, Group, or IntegrationEndpoint
• Notification describes Condition or Result
• Notification isPartOf AlertingPolicy or EscalationWorkflow
• Notification isLoggedIn NotificationHistory

Prompt Examples:
• “Send a notification to the admin group when disk usage exceeds 90%.”
• “What notifications were generated during last night's patching window?”
• “Suppress non-critical notifications for tenant ‘Acme Legal’ during maintenance.”

Agent Execution:
When a prompt refers to a “Notification,” the AI agent will:
• Identify the trigger context (event, threshold, policy action, etc.)
• Determine the appropriate recipients and delivery channel
• Format the message using configured templates or dynamic event data
• Deliver the message and log the notification event
• Optionally suppress, escalate, or bundle notifications based on configuration

Report

Canonical Term: Report

Definition:
A Report is a structured output that presents data collected from systems, policies, or agents over a defined time period. Reports are used for analysis, auditing, compliance, customer communication, or internal performance tracking. In the MSP and N-able context, reports may include information on device health, backup status, patch compliance, license usage, or automation outcomes. Reports can be generated on demand or scheduled, and may be exported in formats such as PDF, CSV, or JSON.

Recognized Synonyms:
• Summary (if high-level — clarify intent)
• Export (context-specific — only when referring to file generation)
• Dashboard (use when referring to real-time visual views)

Relationships:
• Report summarizes Data or Metrics
• Report isGeneratedBy Product, Agent, or MonitoringTool
• Report isDeliveredTo User, MSP, or Tenant
• Report isFilteredBy TimeRange, Entity, or Policy
• Report isConfiguredIn ReportingPolicy or Schedule

Prompt Examples:
• “Generate a weekly patch compliance report for tenant ‘SecureTech’.”
• “What reports are available for backup status over the past 30 days?”
• “Email the executive summary report to all distributors by Monday.”

Agent Execution:
When a prompt refers to a “Report,” the AI agent will:
• Retrieve relevant data sources based on the requested topic and filters
• Aggregate, format, and organize the information into the specified report type
• Apply branding, formatting, and delivery settings (if configured)
• Export the report to a file or deliver it through email, UI, or third-party systems
• Log the report generation and delivery for traceability

Network

Canonical Term: Network

Definition:
A Network is a collection of interconnected devices, systems, and services that communicate using standard protocols. In the DevOps Infrastructure Ontology, it includes both physical infrastructure (e.g., routers, switches, cabling) and logical constructs (e.g., IP subnets, VLANs, virtual networks). In the MSP and N-able context, networks are monitored for availability, performance, security, and connectivity, often segmented by customer, site, or service.

Recognized Synonyms:
• LAN / WAN (context-specific — use when defining scope)
• Segment
• Subnet
• Virtual Network (when hosted in cloud environments)

Relationships:
• Network connects Device or Service
• Network isMonitoredBy Agent or NMS (Network Monitoring System)
• Network contains IPAddressRange
• Network isSecuredBy Firewall or AccessControlList
• Network isAssociatedWith Site or Tenant
• Network hasTopology including Nodes and Links

Prompt Examples:
• “Check for latency issues in the network segment at the Dallas site.”
• “List all devices connected to the ‘Guest Wi-Fi’ network.”
• “Is the backup network isolated from production traffic in tenant ‘HealthBridge’?”

Agent Execution:
When a prompt refers to a “Network,” the AI agent will:
• Identify the relevant network(s) based on name, address range, or associated site/tenant
• Retrieve monitoring metrics (latency, packet loss, uptime, etc.)
• Determine what devices or services are connected to or dependent on the network
• Identify policy violations or anomalies (e.g., unauthorized access, unencrypted traffic)
• Trigger diagnostics, isolation, or notification workflows if thresholds are breached

IP

Canonical Term: IP Address

Definition:
An IP Address is a unique numerical identifier assigned to a device or interface within a network, facilitating communication between devices. IP addresses can be either IPv4 or IPv6 and are essential for routing data packets across networks. In the DevOps Infrastructure Ontology, an IP address is considered a fundamental component of network infrastructure, associated with devices, services, and network segments.

Recognized Synonyms:
• IP
• Internet Protocol Address
• Network Address (context-specific)
• Host Address (when referring to device-specific IPs)

Relationships:
• IP Address isAssignedTo Device or NetworkInterface
• IP Address belongsTo IPNetwork
• IP Address isUsedBy Service or Application
• IP Address isManagedBy NetworkManagementSystem
• IP Address hasType (e.g., Public, Private, Static, Dynamic)

Prompt Examples:
• “List all IP addresses assigned to the 'Finance' department's servers.”
• “Check if the IP address 192.168.1.10 is currently active on the network.”
• “Identify any duplicate IP addresses within the 'NYC Office' subnet.”

Agent Execution:
When a prompt refers to an “IP Address,” the AI agent will:
• Retrieve the IP address details, including assignment, status, and associated device or service.
• Determine the network segment or subnet the IP address belongs to.
• Check for conflicts, such as duplicate assignments or unauthorized usage.
• Provide recommendations or actions, such as releasing, reassigning, or monitoring the IP address.

Workflow

Canonical Term: Workflow

Definition:
A Workflow is a structured sequence of tasks or activities designed to accomplish a specific objective within a system or process. In the DevOps Infrastructure Ontology, a workflow represents the orchestration of actions, often automated, that manage and control various aspects of IT infrastructure and software development. Workflows are essential for ensuring consistency, efficiency, and repeatability in operations.

Recognized Synonyms:
• Process Flow
• Automation Sequence
• Task Pipeline
• Operational Procedure

Relationships:
• Workflow consistsOf Actions or Tasks
• Workflow isTriggeredBy Event or Condition
• Workflow isManagedBy OrchestrationTool or AutomationPlatform
• Workflow interactsWith Resources such as Devices, Services, or Applications
• Workflow hasOutcome Result or StateChange

Prompt Examples:
• “Initiate the onboarding workflow for a new employee in the 'Sales' department.”
• “What workflows are currently active for patch management across all servers?”
• “Modify the backup workflow to include encryption before data transfer.”

Agent Execution:
When a prompt refers to a “Workflow,” the AI agent will:
• Identify the specific workflow based on context and naming conventions.
• Execute the defined sequence of tasks, ensuring each step is completed successfully.
• Monitor the workflow's progress and handle exceptions or errors as defined.
• Log the execution details for auditing and compliance purposes.
• Provide feedback or notifications upon completion or if intervention is required.

Task

Canonical Term: Task

Definition:
A Task is an individual, executable unit of work within a larger Workflow. Tasks perform discrete actions such as running a script, checking system status, applying a patch, or sending a notification. In the DevOps Infrastructure Ontology, a task is an atomic operation that contributes to the overall goal of a workflow or automation policy. Tasks may be triggered manually, scheduled, or invoked automatically based on conditions.

Recognized Synonyms:
• Job (context-specific — use in batch or scheduling contexts)
• Step (when referring to a task’s place in a sequence)
• Action (when execution is immediate or user-triggered)
• Unit Operation (less common — use in complex automation flows)

Relationships:
• Task isPartOf Workflow
• Task isExecutedBy Agent or Platform
• Task isExecutedBy Task Manager
• Task isTriggeredBy Event or Condition
• Task modifies or queries Resource (e.g., Device, Policy, Service)
• Task hasResult Success, Failure, or Output
• Task mayDependOn PreviousTask

Prompt Examples:
• “Create a task to restart all printers in the ‘Admin’ department every Sunday.”
• “Which tasks failed during last night’s patching workflow?”
• “Log the output of every task run as part of the onboarding process.”

Agent Execution:
When a prompt refers to a “Task,” the AI agent will:
• Identify the specific task definition, either standalone or as part of a workflow
• Validate prerequisites and execution context (permissions, scope, triggers)
• Execute the action associated with the task
• Capture results, return outputs, and handle exceptions if configured

Agent

Canonical Term: Agent

Definition:
An Agent is a lightweight software component installed on a managed Device (Asset) that enables remote monitoring, management, automation, and communication with N-able or integrated systems. Agents collect telemetry, enforce policies, run scripts, initiate backups, and trigger alerts based on configured logic. In the MSP context, agents are essential for endpoint visibility, proactive issue resolution, and executing AI-assisted or automated workflows.

Recognized Synonyms:
• Monitoring Agent
• Management Agent
• Endpoint Agent

Relationships:
• Agent isInstalledOn Device
• Agent isManagedBy Product (e.g., N-central, Cove)
• Agent executes Script or Task
• Agent reportsTo Platform or OrchestrationService
• Agent monitors Performance, Security, and PolicyCompliance
• Agent enables Action or Automation

Prompt Examples:
• “Check if the N-central agent is online on device ‘HR-LT-003’.”
• “List all agents running outdated versions across the ‘Finance’ tenant.”
• “Trigger the backup task via the local agent on all Windows servers.”

Agent Execution:
When a prompt refers to an “Agent,” the AI agent will:
• Identify the installed agent on the specified device(s)
• Check health, version, and last check-in time
• Determine the agent’s capabilities (e.g., scripting, monitoring, backup)
• Initiate agent-driven tasks such as running scripts, collecting metrics, or applying policies
• Return results or errors from agent-initiated actions

AI Agent

Canonical Term: AI Agent

Definition:
An AI Agent is a software-based autonomous entity that interprets natural language prompts and executes multistep reasoning and actions using defined tools, APIs, and structured knowledge. AI Agents in the N-able context interact with the platform, devices, users, and other systems to perform tasks such as triage, reporting, automation, or decision support. AI Agents can chain tools, call scripts, and escalate issues, often acting on behalf of a technician or service provider.

Recognized Synonyms:
• Intelligent Agent
• Proactive Assistant
• Automation Agent
• GenAI Agent (in LLM-driven systems)

Relationships:
• AI Agent interprets Prompt
• AI Agent calls Action or Tool
• AI Agent queries DataSource or KnowledgeBase
• AI Agent interactsWith User, Device, or Platform
• AI Agent isEnabledBy Model Context Protocol (via MCP)
• AI Agent produces Result or SuggestedResolution

Prompt Examples:
• “AI Agent, why is device ‘ENG-LT-22’ running slowly this week?”
• “Summarize patch compliance across all tenants and suggest remediation for non-compliant devices.”
• “What backup risks exist for customers using Microsoft 365, and how should we address them?”

Agent Execution:
When a prompt is directed to an AI Agent, it will:
• Interpret the natural language request and identify intent
• Use structured terminology (from the MSP MCP Terminology Standard) to disambiguate terms like “asset” vs. “device”
• Call tools (e.g., APIs, scripts, workflows) to gather data or take action
• Reason over context (e.g., device health, policy violations, previous incidents)
• Return human-readable results or initiate further actions
• Optionally escalate, log, or interact with other AI Agents

Artificial Intelligence (AI) External Standards Committee

Terms and Conditions for Dictionary of Definitions

Welcome to the Dictionary of Definitions (the “Dictionary”) developed and maintained by the N-able AI External Standards Committee (the “Committee”). These Terms and Conditions (“Terms”) govern your access to and use of the Dictionary, including all content, definitions, and related services provided through this community. By accessing, using, or contributing to the Dictionary, you agree to be bound by these Terms.

1. Purpose
   This Agreement governs the use of the Dictionary, which is intended to be a collaborative resource that provides clear, consistent definitions for terms related to AI community standards, ethics, and governance. It aims to support transparency, accountability, and shared understanding across organizations and communities with regard to the use of AI.

2. Open Access and Use
   The Dictionary is provided under an open-source model as follows:

Permitted Use: You may access, use, and share the definitions for educational, research, policy-making, or organizational purposes, provided proper attribution is given.

Attribution: When using or referencing content from the Library, you must credit the source as follows: “Source: Community Standards Library of Definitions, https://developer.n-able.com/n-central/page/msp-and-it-pro-nlp-ai-mcp-terminology-standard .”

Prohibited Use: You may not use the content for commercial purposes without prior written permission of N-able. You may not alter definitions in a way that misrepresents their original intent or context.

Users may contribute new definitions or suggest edits to existing ones. By submitting content, you grant the Library a non-exclusive, royalty-free, worldwide license to use, reproduce, and distribute your contributions.

All submissions are subject to review by the Committee to ensure accuracy, neutrality, and alignment with the Library’s mission.

3. No Warranty
   The Dictionary is provided “as is” and “as available,” without warranty of any kind. N-able and the Committee expressly disclaim all warranties, whether express, implied, statutory, or otherwise, including but not limited to warranties of merchantability, fitness for a particular purpose, accuracy, or non-infringement.

4. Limitation of Liability
   In no event shall N-able, the Committee, or any of its contributors be liable for any direct, indirect, incidental, special, exemplary, or consequential damages (including, but not limited to, procurement of substitute goods or services; loss of use, data, or profits; or business interruption) arising in any way out of the use or inability to use the Dictionary, even if advised of the possibility of such damages.

5. No Endorsement or Obligation
   Use of the Dictionary does not imply endorsement by N-able or the Committee. The Committee is under no obligation to update, maintain, or support the Dictionary or to respond to feedback or requests.

6. Intellectual Property
   All content in the Library, unless otherwise noted, is the intellectual property of the Library or its contributors and is protected by applicable copyright and intellectual property laws. All contributions to the Dictionary are subject to open use and redistribution. Contributors agree that their submissions may be freely used, modified, and distributed by others under these terms.

7. Governing Law
   This Agreement shall be governed by and construed in accordance with the laws of the State of Massachusetts without regard to its conflict of law principles.

8. Modifications to Terms

We reserve the right to update or modify these Terms at any time. Changes will be posted on this page with an updated effective date. Continued use of the Library after changes constitutes acceptance of the new Terms.

9. Contact

For questions, feedback, or permissions, please contact us at:

[email protected]