Hi All
Organizations adopting GenAI are increasingly seeking conversational experiences, intelligent automation, and multi‑agent interoperability within their digital ecosystems. ServiceNow now enables this through capabilities such as AI Agent triggering within Virtual Agent flows, embedding VA topics as tools, and invoking AI Agents through the AI Connector for deeper contextual responses. With Quick MCP Server, customers can securely expose ServiceNow skills as tools for external AI systems, while also allowing ServiceNow to act as an MCP client to consume third‑party tools. Additionally, A2A (Agent‑to‑Agent) empowers secure collaboration between ServiceNow AI Agents and external agent frameworks like SAP Joule or AgentForce. Together, these capabilities enable a robust, flexible multi‑agent architecture suitable for enterprise-grade automation, orchestration, and AI-driven workflows.
🔹 Example Use Cases
- Trigger AI Agents during live VA conversations for contextual resolutions.
- Allow external AI agents (Claude, GPT, SAP Joule, AgentForce) to execute ServiceNow actions via MCP tools.
- Consume external MCP servers for data enrichment (financial datasets, DevOps platforms, knowledge sources).
- Automate cross‑platform tasks using A2A, such as multi‑agent case routing or triage.
- Build multi-agent workflows where ServiceNow AI Agents collaborate with enterprise systems to complete end‑to‑end processes.
(1) TRIGGER AI Agent from Now Assist SNVA:
1.Ensure your AI Agent use case is published and not in draft.
Go to AI Agent Designer → Use Case → Make sure it is active and published.
2. Virtual Agent Assistant
Navigate to: Conversational Interfaces > Virtual Agent > Assistants
Open your active assistant (e.g., "Now Assist for Virtual Agent").
Under Source, ensure it includes: AI Agent use cases
3.AI Agent Use Case Channels
Make sure the AI Agent use case is available for the Virtual Agent channel:
Go to the AI Agent Designer-->Open your Use Case-->Go to the Channels tab-->Ensure Virtual Agent is selected as an available channel
4. Intent Detection
5. Test in Virtual Agent Client
Go to Virtual Agent > Test Console Enter an utterance that matches the AI Agent intent
====================================================================================
(2) ADD Virtual Agent Topic in AI Agent as a Tool:- Add a Virtual Agent topic to an AI agent in AI Agent Studio so that you can use conversations to get additional information from the user.
- Navigate to All > AI Agent Studio > Create and manage > AI agents.
- Open the AI agent that you want to add a conversational topic to and navigate to the Add tools and information section.
- In the Add tool drop-down list, select Conversational topic.
==================================================================================
(3) How to Invoke AI Agent inside VA Topic: - enables integration of generative AI into custom chat flows, allowing for improved, contextual responses. An AI agent performs a task that you pass to it inside the AI Connector utility in Virtual Agent Designer. Make Sure AI Connector Utility is Installed.
Open VA Designer and Select AI Connector NODE
To use the AI agent--->select the Skill type drop-down list in the AI Connector utility NODE
then select the AI agent's type from the Select AI agent drop-down.
=================================================================================
(4) Expose A Tool (SKILL) Using ServiceNow Quick MCP Server:
- Install & Enable the Quick MCP Server on Your ServiceNow Instance
- Ensure MCP Server is enabled (Zurich Patch 4+ / Yokohama Patch 11+).
If MCP Server is not visible, ServiceNow notes that until mid‑Jan 2026, an Incident must be logged for “CNS - Application Delivery Controller” to enable MCP Server.
Sample Quick MCP - https://<instance>.service-now.com/sncapps/mcp-server/mcp/sn_mcp_server_default
2 . Publish Skills as MCP Tools
Now Assist skills OOTB or Custom can be used as tools. But ALL Now Assist Skills at this moment can not be exposed as Tool.
- Navigate to All > MCP Server Console.
- From the Configuration tab, select Tools.
- Select New --Fill the fields and Save.
3. Configure Authentication
Connect to an MCP server from an MCP client by creating an OAuth inbound integration and configuring a client with the server details. Configure authentication: OAuth 2.1, API Key, or Authless (OAuth recommended).
STEPS Here in URL:- Connect • Zurich Enable AI • Docs | ServiceNow
4. CALL ServiceNow Exposed Tools from External AI Agents using Quick MCP
Configure an MCP client to connect to an MCP server and prompt the server for information or to perform an action.
-> Configure an MCP client with the following details as required by the client.
erver details Value
| Server URL | https://<server-instance>/sncapps/mcp-server/mcp/<server> To connect to the preconfigured Quickstart Server, use https://<server-instance>/sncapps/mcp-server/mcp/sn_mcp_server_default. |
| Host | <server-instance>.service-now.com |
| Base URL | /sncapps/mcp-server |
| Scope | mcp_server |
| Authentication type | OAuth 2.0 |
| Identity Provider | Generic OAuth 2 |
| Authorization URL | https://<server-instance>.service-now.com/oauth_auth.do |
| Token URL | https://<server-instance>.service-now.com/oauth_token.do |
| Token Revocation URL | https://<server-instance>.service-now.com/oauth_revoke.do |
| Refresh URL | https://<server-instance>.service-now.com/oauth_auth.do |
| Redirect URL | https://<server-instance>.service-now.com/oauth/callback |
| Client ID | The client ID from the OAuth inbound integration on the instance. |
| Client secret | The client secret from the OAuth inbound integration on the instance. |
After configuring these details, the client calls the server with the Authorization: Bearer <token> header. If the token is validated by the server, the client receives the list of tools available for use.
-> From MCP Client we can see the List of Tools exposed.
-> Enter a prompt for the information you need or for the tool to perform an action on the instance. Example Prompts on MCP Client like - "Summarize all cases closed this week."
==============================================================================
(5) Consume External Tool Using ServiceNow acting AS a MCP Client:
ServiceNow acts as a Model Context Protocol (MCP) client when its internal AI agents and applications, such as Now Assist AI Agents or Virtual Agent topics, need to access data and perform actions within external systems that expose their capabilities via an MCP server.
Example MCP Servers exposed: https://github.com/modelcontextprotocol/servers , https://learn.microsoft.com/api/mcp ,https://mcp.financialdatasets.ai/mcp
How to
(A) Install MCP Client Application on ServiceNow Instance.
Plugin Name-SN_AIA & sn_aia.enable_mcp_toolsystem property is set to true to enable the external MCP tool experience.
(B) Add External MCP Server in AI Agent Studio.
Add MCP Server with any of below Authentication Method- OAUTH 2.1 or API Key. Test the Authentication is working then only tools list will come as per next step.
(C) Add MCP Server Tools (Tools Exposed by External System)
- Navigate to All > AI Agent Studio > Create and manage > AI agents.
- Open the AI agent to which you want to add an MCP tool to and navigate to the Add tools and information section.
- In the Add tool drop-down list, select MCP server tool.
- fill the form with
(D) Test your AI Agent
Add tool in ServiceNow AI Agent and Do the Test in PLAYGROUND.
Note: MCP Client roles and Properties in this URL- (https://www.servicenow.com/docs/r/intelligent-experiences/mcp-reference.html)
(6) A2A Expose ServiceNow AI Agent
ServiceNow Agent-to-Agent (A2A) feature allow autonomous AI agents within the ServiceNow platform to securely interact, collaborate, and share data with third-party AI agents. Some Details.
- Now Assist AI Agents: Serves as the primary interface for A2A interactions.
- AI Agent Studio: Used to configure and activate external A2A agents.
- AI Agent Fabric: Connects ServiceNow with third-party tools.
Pre-Requisites:
- Platform version:
- Yokohama Patch 11+ (January 2026)
- Zurich Patch 4+ (December 2025)
- Now Assist AI Agents 6.0.x+ (December 2025 release)
- Settings Done on Instance
- Turn on external agents (AI Agent Studio > Settings
- Allow ServiceNow to access External AI agents (sn_aia.external_agents.enabled: true)
- Allow third party to access ServiceNow AI Agents ( sn_aia.internal_agents.enabled_external: true)
How to Test from POSTMAN (Calling ServiceNow AI Agent from POSTMAN):
List of A2A APIs :
(GET) Agent Card V1: /api/sn_aia/a2a/id/{agent_id}/well_known/agent_json
(GET) Agent Card V2: /api/sn_aia/a2a/v2/agent_card/id/{agent_id}
(POST) Agent Execution V1: /api/sn_aia/a2a/v1/agent/id/{agent_id}
(POST) Agent Execution V2: /api/sn_aia/a2a/v2/agent/id/{agent_id}
Sample URL for AI Agent Card Discovery: https://YOURINSTANCE.service-now.com/api/sn_aia/a2a/id/5159a3e6fb422a101e62fbb202efdcf1/well_known/agent_json
(A) Replace ID with SYS_ID & SYS_ID Of AI Agent . This Can be Fetched from sn_aia_agent Table.
(B) Create OAUTH API End Point for External Client .
(C) Construct URL and Parameters for API Call (Agent Card Discovery and Details GET Method)
- In Postman, construct a GET request, like this:
- Method: GET
- URL: https://<your instance name>.service-now.com/api/sn_aia/a2a/id/<your AI Agent ID>/well_known/agent_json
- Auth Type: OAuth 2.0
- Add authorization data to: Request Headers
- Token: (empty until you pull a token)
- Header Prefix: Bearer
- Token Name: <enter a name for your token>
- Grant type: Authorization Code
- Callback URL: https://<your instance name>.service-now.com/login.do
- Authorize using browser: false, unchecked
- Auth URL: https://<your instance name>.service-now.com/oauth_auth.do
- Access Token URL: https://<your instance name>.service-now.com/oauth_token.do
- Client ID: <your Client ID value from the Application Registry record in ServiceNow>
- Client Secret: <your Client Secret value from the Application Registry record in ServiceNow>
- Scope: (leave blank, response should select a2aauthscope by default)
- State: <enter some random value, but keep it unique between each request send>
- Client Authentication: Send as Basic Auth header
- Refresh Token URL: (will be populated automatically by other fields)
Response as below your Will Get:
=================================================================
{
"protocolVersion": "0.3.0",
"name": "Custom Test AI Agent",
"description": "Searches for Listed Incidents for a Given Month.",
"url": "https://YOURINSTANCE.service-now.com/api/sn_aia/a2a/v1/agent/id/5178a3e6fb44a010e96fbb872efdcf8",
"preferredTransport": "JSONRPC",
"additionalInterfaces": [],
"provider": {
"organization": "ServiceNow",
"url": "https://YOURINSTANCE.service-now.com/"
},
"iconUrl": "",
"version": "1.0.0",
"documentationUrl": "",
"capabilities": {
"streaming": false,
"pushNotifications": true
}
}
=================================================================
Here i end my Article. Hope This Helps.
Regards
RP
