SailPoint
Also known as: sailpoint.com, SailPoint Harbor Pilot, SailPoint Agentic Fabric, SailPoint Human Fabric, SailPoint Identity Security Cloud, SailPoint IdentityIQ, IdentityNow
SailPoint sells identity governance through Identity Security Cloud and IdentityIQ, and puts agents to work on both sides of it. Harbor Pilot's agents search documentation and tenant data, build workflows from a description and submit access requests for users, while MCP servers let outside agents create and cancel access requests and read identity risk. Agentic Fabric discovers the AI agents a company runs, assigns owners, puts them through approvals and certifications, and blocks or deactivates them when a threat is confirmed.
SailPoint, based in Austin, Texas, is one of the largest identity governance vendors: Identity Security Cloud and IdentityIQ decide who gets access to which applications and data, run access certifications and separation of duties checks, and connect to thousands of enterprise systems. More than half of the Fortune 500 are customers, Currys among them, and 95% of customers renew.
Thoma Bravo took the company private in 2022 and listed it again on Nasdaq in February 2025. Its platform now splits into Human Fabric for people, the next version of Identity Security Cloud, and Agentic Fabric for AI agents and other nonhuman identities, sold together as SailPoint Identity Security, with technology from its Entro acquisition folded in for secrets and credentials.
Harbor Pilot is the agent layer inside Identity Security Cloud, launched in April 2025 and used by about half of SailPoint's customers. It shows its plan as it works, and its agents cover four jobs: a documentation agent that answers with links to sources, an admin search agent that answers questions about the tenant such as identities not certified recently, a workflows agent that builds new workflows from a description, and an access request agent that finds the right application and submits the request into SailPoint's approval process for the user.
Two MCP servers open the same ground to outside agents: Access Requests lets Claude, Cursor and other clients list requestable items and create, view and cancel requests, stamping each cancellation as made by AI on the user's behalf, and Identity Intelligence returns a person's profile, access, risk and certification history read only. Agentic Fabric, generally available since August 2026, governs the agents customers deploy.
Endpoint and browser sensors, cloud and SaaS connectors for Microsoft 365 Copilot, Amazon Bedrock, Vertex AI, Agentforce, Snowflake Cortex and others, and SIEM and XDR telemetry feed a registry of agents, MCP servers and credentials; each agent gets an owner, joins certification campaigns and requests access through approvals, and the Protect tier watches behavior, redacts credentials and personal data from prompts before they reach outside models, and blocks or deactivates agents when a threat is confirmed.
Pricing is quoted. Agentic Fabric sells standalone in Standard, Essentials and Advanced tiers with a Protect add on, or inside the Agentic Standard, Agentic Business and Agentic Business Plus suites, and a free discovery tool trial shows a company its agents first. SailPoint holds SOC 1 and SOC 2 Type II, C5, ISO 27001, 27017, 27018 and 27701 and CSA STAR Level 2, and Identity Security Cloud carries FedRAMP Moderate and GovRAMP authorization on AWS GovCloud. Tenants run on AWS in a region the customer picks at signup, Harbor Pilot runs only where its AWS Bedrock model is offered, and IdentityIQ still installs on premises.
It fits large enterprises already governing human access with SailPoint that now need the same owners, approvals and certifications for AI agents, and identity teams that want routine requests and workflow builds handled in conversation. It fits less well for buyers wanting an agent to resolve SOC alerts across many tools, teams outside SailPoint's platform, or anyone expecting a list price.
Vendor details
Canonical URL
https://www.sailpoint.com
Category
Security / SOC agent
Subcategory
Identity governance platform with Harbor Pilot agents for identity teams, MCP servers for access requests, and Agentic Fabric to govern customers' AI agents
Funding status
Public on Nasdaq (SAIL) since its February 2025 listing; Thoma Bravo took the company private in 2022 and remains its largest shareholder.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Identity Security Cloud and IdentityIQ connect to thousands of applications, directories and cloud platforms. Agentic Fabric connectors discover agents in Microsoft 365 Copilot, Microsoft Foundry, Amazon Bedrock, Google Vertex AI, Databricks, Salesforce Agentforce, ServiceNow AI Platform, Snowflake Cortex AI and Wiz, plus endpoint and browser sensors and SIEM and XDR telemetry. MCP servers connect Claude, Cursor and other MCP clients.
In practice
A new analyst asks Harbor Pilot for read access to the finance data warehouse. The access request agent finds the right application and entitlement, fills the request form and submits it into the approval chain, where the analyst's manager and the data owner sign off.
An identity admin describes a leaver process in a sentence: disable accounts, remove group memberships and notify the manager. The workflows agent builds the new workflow, which the admin reviews and turns on.
A security team finds 400 AI agents across Copilot Studio, Bedrock and developer laptops in Agentic Fabric's registry. Each gets an owner and joins the quarterly certification, and when one starts calling an unfamiliar server at night the Protect tier flags it and the kill switch deactivates its access.
Sources & related URLs
Related / legacy domains
Research sources
Agentic Index coverage score
9.0 / 14 capabilities · 64%
| Integrations & Tool Calling | Partial |
|---|---|
|
Harbor Pilot acts inside Identity Security Cloud: it searches the tenant, submits access requests and creates workflows there, and the workflows it builds then reach connected applications through SailPoint's own connectors. Agentic Fabric's connectors to Microsoft 365 Copilot, Bedrock, Vertex AI, Agentforce, Snowflake Cortex and Wiz discover and govern agents rather than letting Harbor Pilot act in those systems on a user's request. Sourcesailpoint.com/products/harbor-pilotread 2026-11-18 |
|
| Workflow Orchestration | Full |
|
Harbor Pilot coordinates a set of agents behind one prompt and shows its logic and plan for the next actions as it works. Its workflows agent turns a plain language description into a workflow in Identity Security Cloud's engine, with triggers, actions and operators, so a multistep identity process such as a leaver flow can be designed in conversation and then run by the platform. Sourcedocumentation.sailpoint.com/saas/help/ai/harbor_pilot/index.htmlread 2026-11-18 |
|
| Knowledge Grounding & RAG | Partial |
|
The documentation agent searches SailPoint's product documentation in plain language and links to the sources it used, and the admin search agent answers questions about the live tenant, such as which identities have not been certified recently. Neither searches the customer's own policies or documents, and when an object is not found the agent may return roles or entitlements instead. Sourcesailpoint.com/products/harbor-pilotread 2026-11-18 |
|
| Human Oversight & Guardrails | Full |
|
Every access the agent asks for is a request that enters SailPoint's approval process rather than a grant, its forms must be completed and saved before submission, and the workflows agent can create new workflows but cannot edit existing ones. Cancellations through MCP carry an automatic note that AI made them on the user's behalf, and Agentic Fabric puts customers' agents through approvals, certifications and a one click kill switch. Sourcedeveloper.sailpoint.com/docs/extensibility/mcp-available-toolsread 2026-11-18 |
|
| Security, Identity & Governance | Full |
|
SailPoint holds SOC 1 and SOC 2 Type II reports, a C5 Type II audit, ISO 27001, 27017, 27018 and 27701 and CSA STAR Level 2, and Identity Security Cloud carries FedRAMP Moderate and GovRAMP authorization. Admins switch Harbor Pilot on only after accepting AI terms, decide separately whether end users get it, and grant each MCP server its own OAuth scope. Sourcesailpoint.com/why-us/trustread 2026-11-18 |
|
| Observability & Auditability | Full |
|
Actions taken through the MCP server, creating or cancelling access requests, generate audit events that appear in Identity Security Cloud search, and Harbor Pilot shows its reasoning and planned steps as it works. Agentic Fabric builds behavior baselines for customers' agents from EDR and SIEM telemetry, gathers findings into a risk center and produces framework aligned reports for audits. Sourcedocumentation.sailpoint.com/saas/help/ai/mcp.htmlread 2026-11-18 |
|
| Memory & State Persistence | Partial |
|
Context lasts for the current conversation, which a user resets through Menu and New Conversation, and Harbor Pilot keeps no profile of a user's preferences between sessions. Lasting state belongs to the identity platform itself: requests, workflows and the Identity Graph that links each AI agent to its owner, credentials, entitlements and the data it reaches. Sourcedocumentation.sailpoint.com/saas/help/ai/harbor_pilot/index.htmlread 2026-11-18 |
|
| Deployment & Data Residency | Full |
|
Tenants run on AWS in the hosting region the customer elects when the service starts, the FedRAMP instance runs separately on AWS GovCloud, and IdentityIQ still installs on premises. Harbor Pilot runs only in AWS regions where its Bedrock model is offered, which rules out the UAE region, and tenants with sovereignty requirements use tenant specific MCP addresses. Sourcesailpoint.com/legal/sub-processorsread 2026-11-18 |
|
| Prebuilt Agents / Templates / Packs | Full |
|
Four ready agents ship in Harbor Pilot, each for its own job: documentation answers, admin search over the tenant, workflow building and access requests. Customers turn them on from settings rather than build them, and about half of SailPoint's customers were using them within months of launch. Agentic Fabric adds ready connectors for the major agent platforms and puts agents straight into existing certification campaigns. Sourcesailpoint.com/products/harbor-pilotread 2026-11-18 |
|
| Triggers & Channel Coverage | Partial |
|
Harbor Pilot wakes only when a person prompts it in the Identity Security Cloud interface or an outside agent calls the MCP servers. Event driven work runs elsewhere: identity workflows fire on platform triggers, and in Agentic Fabric a validated threat sets off workflows that block an endpoint agent or remove its access by policy, without any of Harbor Pilot's agents starting on their own. Sourcesailpoint.com/blog/whats-new-sailpoint-agentic-fabric-garead 2026-11-18 |
|
| Model Flexibility & Routing | Not documented |
|
Harbor Pilot runs on a single AWS Bedrock model that SailPoint selects and keeps unnamed, and customers have no setting to choose another model or bring their own. Agentic Fabric's prompt scanning sits in front of the outside models customers' agents use, which governs those calls rather than giving Harbor Pilot a choice of model. Tenants in regions without that model, such as the UAE, go without Harbor Pilot. Sourcedocumentation.sailpoint.com/saas/help/ai/harbor_pilot/index.htmlread 2026-11-18 |
|
| APIs / SDKs / MCP Extensibility | Full |
|
Two MCP servers sit behind a global address that finds the tenant at sign in, use OAuth with PKCE and a scope per server, and name their tools: list-requestable, create-access-request, view-access-requests and cancel-access-request for access requests, and get-identity-intelligence and get-identity-resource for read only identity investigation. Beneath them sit the full Identity Security Cloud REST API, SDKs and a CLI. Sourcedeveloper.sailpoint.com/docs/extensibility/mcp-available-toolsread 2026-11-18 |
|
| Testing, Debugging & Optimization | Not documented |
|
Thumbs up and down ratings feed SailPoint's own tuning of Harbor Pilot, and its answers carry a reminder that they can be wrong and need checking. Customers get no test sets, scored evaluations or replay of agent runs, and the Agentic Fabric Test Drive is a demonstration environment rather than a way to test the customer's own agents. Quality checks fall to the people who approve requests and review new workflows. Sourcedocumentation.sailpoint.com/saas/help/ai/harbor_pilot/index.htmlread 2026-11-18 |
|
| Browser / Computer-use | Not documented |
|
No agent drives a browser or operates an application screen on a user's behalf. Agentic Fabric's browser and endpoint sensors watch for AI tools people use, including unsanctioned ones, which is discovery of other agents rather than a computer use ability of SailPoint's own. Access changes reach target systems through connectors and APIs rather than through screen automation. Sourcesailpoint.com/blog/whats-new-sailpoint-agentic-fabric-garead 2026-11-18 |
|
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; suites and Agentic Fabric tiers sold through sales
Enterprise subscriptions by suite and identity count
What is public
SailPoint sells by quote. Agentic Fabric comes standalone in Standard, Essentials and Advanced tiers, with a Protect add on for Advanced, or inside the Agentic Standard, Agentic Business and Agentic Business Plus suites, with a Non-Human Identity Protect add on for Business Plus. A free discovery tool trial maps a company's agents.
Billing mechanics
Each tier and suite adds capability on top of the last, from visibility and ownership through governance to risk insight and SOC response, and the suites cover human and nonhuman identities together.
Cost watchouts
Governing AI agents needs Agentic Fabric, sold as its own tiers or inside the Agentic suites, and the Protect capabilities come only as add ons to the Advanced tier or the Business Plus suite.
Variable cost rationale
Contracts are quoted subscriptions rather than metered agent usage.
Additional watchouts
No list prices appear for any suite, tier or add on, and Harbor Pilot's inclusion by suite is not stated.
Sales call required
Yes, required for paid access
Free / trial
Free Agentic Fabric discovery tool trial; Test Drive demo environment
Lowest paid plan
None; every suite and tier is quoted
Commercial notes
Agentic Fabric is sold to new customers on its own as well as to existing Identity Security Cloud and IdentityIQ customers.
Missing data
SailPoint gives no prices for its suites, Agentic Fabric tiers or add ons, and does not say which suites include Harbor Pilot.
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…
- Anvilogic — Agentic security operations platform that works on top of existing…
Alternatives to SailPoint
Other products that do the work SailPoint does, and what each one is.
- BlinkOpsAgentic security operations platform
- NetskopeCloud security platform
- Dropzone AIAI SOC analyst
- Proofpoint
- SprintoCompliance platform
- D3 Security (Morpheus)