Veza
Also known as: Veza Technologies, Veza from ServiceNow, Veza Access Agents, Veza Access Graph, veza.com
Veza, now part of ServiceNow, maps who and what can take which action on which data across SaaS, cloud and on premises systems, including AI agents and machine identities. Its Access Graph feeds governance that provisions and removes access as people join, move and leave, finds agents across Bedrock, Copilot Studio, Agentforce and Claude Code and strips their excess permissions. Access Agents answer identity questions in plain language and guide reviewers through access reviews.
Veza sells identity security built on one question: who and what can take which action on which data. Founded in 2020, it raised a $108 million Series D in April 2025 at an $808 million valuation, and ServiceNow completed its acquisition on March 2, 2026. Veza still sells under its own name, now presented as Veza from ServiceNow, while ServiceNow folds its Access Graph into a wider agent identity control plane. Customers include Sallie Mae, Blackstone, Wynn Resorts, Snowflake, Choice Hotels, Genesys, Expedia and Intuit, and Instacart found tens of privilege escalation issues within a day of turning it on.
The Access Graph unifies identities, groups, roles, policies, permissions and resources across more than 325 integrations, from Entra ID, Okta and Active Directory to AWS, Salesforce, Snowflake and GitHub, and custom applications join through the Open Authorization API.
Posture management, monitoring and access intelligence sit on that graph, along with identity governance: access reviews, access requests with just in time access that expires on its own, separation of duties and lifecycle management.
Lifecycle management provisions birthright access when someone joins, adjusts it when they move and removes it, local accounts included, when they leave, across thousands of SCIM applications, with dry runs and safety limits that stop large unintended changes.
AI agents get the same treatment as people. AI Agent Security discovers agents and their connections across Bedrock and AgentCore, Vertex AI, Azure AI Foundry, Copilot Studio, Agentforce, the OpenAI Agent Platform, Claude Code and agents registered with Entra Agent ID, including links to some 2,000 public MCP servers.
Each agent is tied to a human owner, excess permissions and risky connections can be removed, and alerts fire when an agent's owner leaves the company. Access Agents add three assistants of Veza's own: a prompt agent that answers identity risk questions, an access search agent that draws permission paths from one request, and an access review agent that flags risky identities and explains decisions in plain language.
Veza sells through demos and a free trial, with no public price list. It holds SOC 2 and ISO 27001 certifications and signs US, UK and EEA data processing agreements, while its trust center and product documentation sit behind ServiceNow and customer sign in. The platform fits large enterprises in finance, hospitality, technology and the public sector that need to govern human, machine and agent access in one place, and especially those already running ServiceNow. Teams after a lightweight tool for a handful of agents, or a published list of hosting regions, will find a heavier enterprise platform.
Vendor details
Canonical URL
https://veza.com/
Category
Security / SOC agent
Subcategory
Identity security and governance for human, machine and AI agent access
Company status
acquired
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
More than 325 integrations, including Entra ID, Active Directory, Okta, AWS, Google Cloud, Salesforce, Snowflake, GitHub, Workday and thousands of SCIM apps, with agent discovery across Bedrock, Vertex AI, Azure AI Foundry, Copilot Studio, Agentforce, the OpenAI Agent Platform and Claude Code.
In practice
A security team finds a Copilot Studio agent with write access to a finance SharePoint site and removes the permission it never uses.
When an engineer leaves, Veza removes her Okta, GitHub and local AWS accounts the same day and logs each change.
A reviewer asks the access review agent why a contractor holds Snowflake admin rights before revoking them.
Sources & related URLs
Related / legacy domains
Research sources
Agentic Index coverage score
8.5 / 14 capabilities · 61%
| Integrations & Tool Calling | Full |
|---|---|
|
Lifecycle management creates accounts, grants birthright entitlements and removes access, local accounts included, across Entra ID, Active Directory, Exchange, AWS, Google Cloud, Okta, Salesforce, Snowflake, GitHub, Atlassian Cloud, Coupa and thousands of SCIM applications. Custom and legacy apps take changes through the OAA Write framework and custom REST actions, and AI Agent Security removes risky agent connections and excess permissions. Sourceveza.com/product/lifecycle-managementread 2026-11-20 |
|
| Workflow Orchestration | Partial |
|
Joiner, mover and leaver policies run multistep changes across many systems at once, granting a new role's access while removing the old one, and access requests move from a catalog choice through approval to automated provisioning. Dashboards can start workflows in ServiceNow or Jira. These are policy driven sequences rather than a runtime that branches across or coordinates several agents. Sourceveza.com/product/lifecycle-managementread 2026-11-20 |
|
| Knowledge Grounding & RAG | Full |
|
The Access Graph keeps a current map of identities, groups, roles, policies, permissions and resources across people, machines and third parties, from apps to data and on premises to SaaS. Access Agents and Access Search query that graph in plain language, with attribute, conditional and negated filters, so an answer about who can reach an S3 bucket rests on the customer's own entitlements. Sourceveza.com/product/access-airead 2026-11-20 |
|
| Human Oversight & Guardrails | Full |
|
Access requests pass through approval before automated provisioning, and access reviews keep the decision with a reviewer while the review agent flags risky identities and explains each case. Lifecycle changes can run as a dry run against production data first, and safety limits halt large unintended user changes caused by unexpected upstream data. Role recommendations simulate the full impact of access before anyone grants it. Sourceveza.com/product/access-requestsread 2026-11-20 |
|
| Security, Identity & Governance | Partial |
|
Veza holds SOC 2 and ISO 27001 certifications, states GDPR and CCPA compliance and limits its own staff's access by least privilege with monitored logs. The roles, permissions and single sign on that govern who can act inside the Veza console sit in product documentation behind customer sign in, and the trust center now routes to a ServiceNow support article that also requires sign in. Sourceveza.com/company/trust-and-securityread 2026-11-20 |
|
| Observability & Auditability | Full |
|
Every provisioning and deprovisioning event is logged together with the policy changes behind it, so a reviewer can trace why an account was created or removed. AI Agent Security adds visual audit trails of agent access and audit ready answers on who can use an agent and what it can reach, aligned with SOX, NIST and OWASP LLMSecOps monitoring guidance. Sourceveza.com/product/ai-agent-securityread 2026-11-20 |
|
| Memory & State Persistence | Not documented |
|
Access Agents work from the live Access Graph on each question, and review history, request records and provisioning logs belong to the governance data model rather than to a memory the assistants consult. Veza names no store of past conversations, saved preferences or learned facts that the prompt, search or review agents carry from one session to the next, with a scope or a way to clear it. Sourceveza.com/product/veza-access-agentsread 2026-11-20 |
|
| Deployment & Data Residency | Not documented |
|
Veza runs as a hosted service that reaches cloud, SaaS and on premises systems through integrations. Hosting regions, a choice of region and any customer environment option are unknown: the trust center sits behind a ServiceNow sign in, the product documentation requires a customer login, and the US, UK and EEA data processing agreements name no hosting location or sub processor list. Sourceveza.com/legalread 2026-11-20 |
|
| Prebuilt Agents / Templates / Packs | Full |
|
Access Agents ship as three named assistants that each do a separate job. The prompt agent answers identity risk questions such as which AWS users can reach S3, the access search agent draws permission paths across clouds and SaaS from one request, and the access review agent flags high risk identities and explains reviews so approvals stop being rubber stamped. Sourceveza.com/product/veza-access-agentsread 2026-11-20 |
|
| Triggers & Channel Coverage | Full |
|
Changes in authoritative sources such as Workday, SAP HCM, Oracle HCM, BambooHR, HiBob or Okta set off joiner, mover and leaver work without anyone filing a ticket. An agent's owner leaving the company raises an alert, dashboards can launch workflows in ServiceNow or Jira, and people start access requests from a self service catalog in Access Hub. Continuous monitoring flags new risk as permissions shift. Sourceveza.com/product/lifecycle-managementread 2026-11-20 |
|
| Model Flexibility & Routing | Not documented |
|
Access AI brings generative features to every Veza product and Access Agents reason over the Access Graph, but the models behind them go unnamed and customers get no choice of provider. The AI platforms Veza lists, among them Bedrock, Vertex AI, Azure OpenAI and the OpenAI Agent Platform, are estates it discovers and secures rather than models it runs. Sourceveza.com/product/access-airead 2026-11-20 |
|
| APIs / SDKs / MCP Extensibility | Full |
|
The Open Authorization API pushes identity, permission and resource metadata from custom or in house applications into Veza through a JSON schema and a REST push API, with a Python library, a .NET package, quick start templates and sample connectors on a public developer site. The wider platform API reference sits in product documentation that requires a customer login. Sourcedeveloper.veza.comread 2026-11-20 |
|
| Testing, Debugging & Optimization | Partial |
|
Lifecycle changes can run as a dry run that shows which users a policy would change before it reaches production, and role recommendations simulate the impact of access before it is granted. Those checks test identity policies and entitlements. The prompt, search and review agents have no test cases, scoring or release gate that measures the quality of their answers. Sourceveza.com/product/lifecycle-managementread 2026-11-20 |
|
| Browser / Computer-use | Not documented |
|
Every change Veza makes runs through an API, SCIM or a custom REST action, and legacy systems join through the OAA framework or flat files rather than through screen automation. The Access Hub is a web console that people use. No agent in the product drives a browser, desktop or remote machine to change access in an application that lacks an interface for it. Sourceveza.com/integrations/open-authorization-apiread 2026-11-20 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Quote only; free trial and demo
Enterprise SaaS subscription by product
What is public
Veza sells its identity security platform through sales, with a free trial and demos for products such as AI Agent Security and Access Agents. A SaaS end user customer agreement governs the subscription.
Billing mechanics
The order sets which Veza products, integrations and identities the subscription covers.
Sales call required
Yes, required for paid access
Free / trial
Free trial
Commercial notes
Veza has been part of ServiceNow since March 2026 and still sells under its own name. Access AI is offered through an early access request.
Missing data
Every price comes from a sales quote, so the fee and billing unit appear only in the order.
Related vendors
- 7AI — Agentic SOC from the Cybereason founders: domain specialized agents…
- Abnormal AI — Behavioral AI email security with three named autonomous agents: a…
- AirMDR — AI native MDR whose AI analyst investigates every alert and writes…
- Airrived — Agentic OS that unifies SOC, GRC, IAM, vulnerability management, IT,…
- Andesite — Human-AI SOC platform where customers build their own agents and…
- Anomali — Anomali combines a security data lake, ThreatStream intelligence and…