Software vendor
ServiceNow
ServiceNow, an NYSE-listed company based in Santa Clara, California, sells AI agents that customers build in natural language in AI Agent Studio.
Function
- Ownership
- PublicSourceServiceNow, Inc. Form 10-K for the fiscal year ended December 31, 2025
ServiceNow, Inc. common stock (NOW) listed on the New York Stock Exchange, per its Form 10-K for 2025.
- Headquarters
- United StatesSourceServiceNow, Inc. Form 10-K for the fiscal year ended December 31, 2025
- SEC reporting
- YesSourceServiceNow, Inc. Form 10-K for the fiscal year ended December 31, 2025
- Audit firm
- No
- Website
- SourceAI Agents – ServiceNow (product page)
- Subprocessor list
- Not stated
- Contract terms
- Not stated
Offering · Platform module
ServiceNow AI Agents
ServiceNow describes its AI agents as acting autonomously across IT, customer service, HR and other functions, either out of the box or built in natural language in AI Agent Studio. SourceAI Agents – ServiceNow (product page)
Its AI Agent Orchestrator coordinates teams of agents, and Action Fabric connects third-party agents through the Agent2Agent protocol and a Model Context Protocol client, governed from what the page calls ServiceNow Super Intelligence Control Tower. SourceAI Agents – ServiceNow (product page)
Each agent runs under a set identity, either the invoking user or a dedicated AI user, with access control lists and role filtering limiting who can invoke it and what it can reach. SourceImplement access control in AI agents (Brazil release documentation)
ServiceNow uses customer data sharing to improve its Now LLMs, and customers can opt out in the Now Assist Admin console. SourcesUser data usage policy for Now Assist (Brazil release documentation)Opt out of data sharing for Now Assist (Brazil release documentation)
Where it fits
CoE analysis Fits organizations that run IT, HR or customer service workflows on ServiceNow and want agents to act on those records. SourceAI Agents – ServiceNow (product page)
CoE analysis Customer data sharing can be opted out of in the Now Assist Admin console, and an agent set to run as an AI user keeps its roles whoever invokes it, so the data-sharing setting and AI user roles both belong in the approval before go-live. SourcesUser data usage policy for Now Assist (Brazil release documentation)Opt out of data sharing for Now Assist (Brazil release documentation)Implement access control in AI agents (Brazil release documentation)
What it does
- Offering type
- Platform moduleSourcesAI Agents – ServiceNow (product page)ServiceNow, Inc. Form 10-K for the fiscal year ended December 31, 2025
- Workflow areas
- SourceAI Agents – ServiceNow (product page)
- Agent platform
- Process automation
- Customer size
- Not stated
AI
- Kind of AI
- SourcesAI Agents – ServiceNow (product page)ServiceNow, Inc. Form 10-K for the fiscal year ended December 31, 2025
- Agents and automated actions
- Built into the platform
- Suggests or acts
- ActsSourceAI Agents – ServiceNow (product page)
- Availability
- Not stated
The product page presents AI agents as embedded in the ServiceNow AI Platform; release status by feature is not stated on the pages read.
- Models named
- SourcesUser data usage policy for Now Assist (Brazil release documentation)ServiceNow, Inc. Form 10-K for the fiscal year ended December 31, 2025
- Microsoft (Azure OpenAI Service)
Control
- Who controls it
- Customer ITSourceImplement access control in AI agents (Brazil release documentation)
Commercial
- Pricing model
- Not stated
- Public price
- Not stated
Evidence
- Independent outcomes
- None found in the sources read.
- Sources read
- Vendor-stated 7
- Independent market evidence None read
- Independently verified outcomes None read
- Pricing None read
- Known limitations None read
- Coverage
- Directory record
- Last verified
Security and control
- Single sign-on
- Not stated
Not described on the pages read.
- User provisioning
- Not stated
Not described on the pages read.
- Role-based access
- SupportedSourceImplement access control in AI agents (Brazil release documentation)
Role-based access control lists set which roles can invoke each agent, agentic workflow and certain tools; agents run as the invoking user or a dedicated AI user, and role filtering limits the roles an agent inherits.
- Administrator controls for AI
- Not stated
Per-agent access control lists, run-as identity and role filtering; data-sharing opt-out and data masking for Now Assist. Whether each AI feature can be switched off on its own is not stated on the pages read.
- Audit logs
- Not stated
Not described on the pages read.
- Certifications
- Not stated
- Data retention
- SourceUser data usage policy for Now Assist (Brazil release documentation)
Prompts and outputs: Input and output data are transient and not cached or stored on the ServiceNow compute hubs; the data used to generate a response is deleted from the hub after the response is generated. Retention in the customer instance is not addressed.
- Data residency
- Not stated
AI workloads go over TLS 1.2 to one of three centralized ServiceNow compute hubs; ServiceNow may use third-party endpoints such as Azure OpenAI, and Azure-hosted GPUs for capacity bursting, within its network boundary. Hub locations are not named.
Azure OpenAI Service named as a possible third-party endpoint.
- Training on customer data
- Opt outSourcesUser data usage policy for Now Assist (Brazil release documentation)Opt out of data sharing for Now Assist (Brazil release documentation)
Scope: Vendor models
Where stated: Documentation
Customer data sharing is used to improve ServiceNow's Now LLMs, and customers can opt out from the Now Assist Admin console Settings page. Data sharing is never enabled for GCC or self-hosted instances. Data processed by third-party endpoints is not subject to use or access by those providers.
AI features
As the vendor or firm describes them. Notes in small print record what the sources leave open.
AI agents and AI Agent Studio
Out-of-the-box agents, or custom agents built in natural language with guardrails in AI Agent Studio; AI Agent Advisor tests use cases against real data before going live.
- Kind
- Agents and automated actions
- Suggests or acts
- Acts
- Who controls it
- Customer IT
- Availability
- Not stated
Control: ACLs set which roles can invoke each agent; the run-as identity and role filtering set what it can do.
SourcesAI Agents – ServiceNow (product page)Implement access control in AI agents (Brazil release documentation)AI Agent Orchestrator and Action Fabric
Coordinates teams of agents; connects ServiceNow and third-party agents through Agent2Agent and MCP.
- Kind
- Agents and automated actions
- Suggests or acts
- Acts
- Who controls it
- Customer IT
- Availability
- Not stated
Control: Managed and monitored from Super Intelligence Control Tower.
SourceAI Agents – ServiceNow (product page)Now Assist
AI tools that operate autonomously with human oversight and predefined guardrails, using ServiceNow's or third-party models.
- Kind
- Built into the platform
- Suggests or acts
- Suggests and acts
- Who controls it
- Customer IT
- Availability
- Not stated
Control: Organizations choose the tools and models; data-sharing opt-out in the Now Assist Admin console (data sharing is never enabled for GCC or self-hosted instances).
SourcesServiceNow, Inc. Form 10-K for the fiscal year ended December 31, 2025User data usage policy for Now Assist (Brazil release documentation)Opt out of data sharing for Now Assist (Brazil release documentation)
Questions to ask
For this offering
- Have we opted out of ServiceNow's customer data-sharing program, what is the default for our instance, and what data was shared before any opt-out?
- Which of our agents run as AI users, which roles do those users hold, and who approves changes to them?
- Which compute hub will process our AI workloads, and when does ServiceNow route them to Azure-hosted capacity?
- How is AI agent use priced and metered in our contract, given that no public pricing was found?
- Is Now Assist for Data Privacy masking active for our agents, and which fields does it cover?
From the diligence question set
- DQ-06 Does each feature suggest or act? Can it create, post, approve or change records, and where is that limit enforced?
- DQ-07 Who can change each feature's configuration, prompts, thresholds, data connections or model: your company, our administrators or individual users? Can we restrict and log those changes?
- DQ-12 Does the feature learn from our data, including corrections and overrides, and how does it detect and handle changes in our business?
- DQ-16 Is our data, including inputs, outputs, documents and corrections, used to train or improve any model, yours or a third party's? Is that off by default, and is the commitment in the contract?
- DQ-18 Can a feature reach data or take actions beyond the permissions of the user it works for, and how is it protected against instructions hidden in the documents or data it reads?
- DQ-19 How are the AI features priced: included in the subscription, sold as an add-on or higher tier, or metered by use? What is the billing unit?