SentinelOne
Also known as: Purple AI, Purple AI Athena, Purple AI Agentic Investigation, Singularity Platform, Singularity Hyperautomation, Purple AI MCP Server, Prompt Security
Purple AI runs zero-click investigations to a verdict with an auditable evidence chain, then either triggers policy-driven response or prompts an analyst. Hyperautomation adds 130+ templates and approval gates. Plans start at $179.99 per endpoint per year, and the agentic analyst is in Enterprise.
SentinelOne is an endpoint and security operations platform whose agentic analyst is Purple AI, embedded in the Singularity Platform. Purple AI Agentic Investigation runs zero-click, autonomously initiated investigations from alert to verdict, and each verdict is backed by a complete, auditable evidence chain. It combines Anthropic's Claude, OpenAI's GPT and SentinelOne's own Ultraviolet models. Purple AI also auto-triages alerts, supports hunting through investigation notebooks, and answers security questions across the customer's OCSF-normalized native and third-party data.
Depending on governance settings, a verdict either triggers an automated, policy-driven response or prompts an analyst with recommended actions. Singularity Hyperautomation, the no-code workflow layer, brings 160+ integrations, 130+ workflow templates, triggers on high-severity alerts, exposures and console activity, human approval gates through messaging apps or email, and a log of every action. The Purple AI MCP Server extends platform data into customers' own agents, and REST and GraphQL APIs cover most console functions.
AI work is paid for in Singularity Credits, a unified currency across the platform. The platform runs as SaaS on AWS or as an on-premises virtual appliance. On premises, self hosted, air gapped and sovereign offerings have been announced for endpoint, Prompt Security and the AI data pipeline, and SentinelOne does not say whether Purple AI itself runs on premises. SentinelOne holds SOC 2 Type 2, ISO 27001 and FedRAMP High.
Vendor details
Canonical URL
https://www.sentinelone.com
Category
Security / SOC agent
Subcategory
Agentic security operations and endpoint platform
Funding status
SentinelOne is a public company (NYSE: S). Its co-founder and CEO is Tomer Weingarten, and it owns Prompt Security.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Singularity Hyperautomation provides 160+ prebuilt integrations across security and IT tools with thousands of actions, and Purple AI works over OCSF-normalized native and third-party data. The Purple AI MCP Server extends platform data into customer-built agents, and REST (OpenAPI) and GraphQL APIs cover most console functions.
In practice
An analyst opens the console to find investigations already completed, each with a verdict and a full evidence chain, rather than a queue of alerts to triage from scratch.
A team keeps its existing security tools and adds agentic investigation on top. Purple AI reasons over native SentinelOne telemetry and integrated third party sources, normalized to OCSF.
A security engineering team uses the Purple AI MCP Server to pull live platform intelligence into an agent it built itself, rather than working only inside the vendor's console.
Sources & related URLs
Research sources
Agentic Index coverage score
10.5 / 14 capabilities · 75%
| Integrations & Tool Calling | Full |
|---|---|
|
Singularity Hyperautomation offers 160+ prebuilt integrations across security and IT tools with thousands of actions. Purple AI works over OCSF-normalized native and third-party data, and the Purple AI MCP Server extends platform data into customers' own agents. Sourcesentinelone.com/platform/singularity-hyperautomationread 2026-09-28 |
|
| Workflow Orchestration | Full |
|
Investigations run from alert to verdict in Purple AI, and customers build their own no-code Hyperautomation workflows, with governed LLM actions, human-in-the-loop steps and reusable snippets, that act on those verdicts. Sourcesentinelone.com/platform/singularity-hyperautomationread 2026-09-28 |
|
| Knowledge Grounding & RAG | Full |
|
Analysts ask Purple AI security questions across the customer's OCSF-normalized native and third-party data in the Singularity Platform, without writing queries. That is retrieval over a maintained store of the customer's own security data. Sourcesentinelone.com/platform/purpleread 2026-09-28 |
|
| Human Oversight & Guardrails | Full |
|
With the Agentic Investigation release, verdicts either trigger automated, policy-driven responses or prompt an analyst with recommended actions under configurable governance. Hyperautomation adds human approval gates for high-impact actions, with approvals through messaging apps or email. Sourcesentinelone.com/platform/singularity-hyperautomationread 2026-09-28 |
|
| Security, Identity & Governance | Full |
|
Certifications include SOC 2 Type 2, ISO/IEC 27001:2022, 27017, 27018 and FedRAMP High, listed on the Trust Center at trust.sentinelone.com. Access controls include SSO and SAML, MFA and role-based access control. Sourcetrust.sentinelone.comread 2026-09-28 |
|
| Observability & Auditability | Full |
|
Every verdict carries a complete, auditable evidence chain, and investigation notebooks document the work. Automated actions are logged for compliance, and every Hyperautomation action is logged and reviewable. Sourcesentinelone.com/platform/purpleread 2026-09-28 |
|
| Memory & State Persistence | Not documented |
|
Investigation notebooks document an analyst's work, and the platform's data store holds the customer's security data, but neither is agent memory. There is no agent memory with its own scope and lifetime. Sourcesentinelone.com/platform/purpleread 2026-09-28 |
|
| Deployment & Data Residency | Full |
|
SentinelOne runs as SaaS on AWS or as an on premises virtual appliance. On premises, self hosted, air gapped and sovereign offerings have been announced for endpoint, Prompt Security and the AI data pipeline, and FedRAMP High is supported, so customers have options to run it in their own environment. SentinelOne does not say whether Purple AI itself runs on premises. Sourcesentinelone.com/press/sentinelone-brings-ai-security-to-on-premise-regulated-sovereign-self-hosted-and-airgapped-environmentsread 2026-09-28 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
There are 130+ Hyperautomation workflow templates for common SOC use cases that customers adopt, alongside Purple AI's built-in auto triage, agentic investigation and hunting. Sourcesentinelone.com/platform/singularity-hyperautomationread 2026-09-28 |
|
| Triggers & Channel Coverage | Full |
|
Purple AI starts zero-click investigations on its own, and Hyperautomation workflows are triggered by high-severity alerts, detected exposures and console activity, so work starts without a person initiating it. Sourcesentinelone.com/platform/singularity-hyperautomationread 2026-09-28 |
|
| Model Flexibility & Routing | Partial |
|
Agentic Investigation combines Anthropic's Claude, OpenAI's GPT and SentinelOne's own Ultraviolet models. SentinelOne picks among those providers itself, and customers cannot choose a model. Sourceinvestors.sentinelone.com/press-releases/news-details/2026/sentinelone-opens-purple-ai-agentic-investigation-to-all-customers-bringing-frontier-ai-directly-into-the-soc/default.aspxread 2026-09-28 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
RESTful APIs in OpenAPI (Swagger) sit alongside GraphQL APIs, and most UI functions are exposed through a customer-facing API. The API reference sits in the console's API Hub, behind a customer login. The Purple AI MCP Server extends platform data into customers' own agents. Sourcesentinelone.com/faqread 2026-09-28 |
|
| Testing, Debugging & Optimization | Not documented |
|
The Community Verdict in Purple AI is trained on investigations from SentinelOne's own Wayfinder managed service and refined by the analyst community. Customers get no harness to evaluate Purple AI or their workflows. Sourcesentinelone.com/platform/purpleread 2026-09-28 |
|
| Browser & Computer Use | Not documented |
|
Actions in Purple AI and Hyperautomation run through platform data and integrations, with no browser or computer use. Sourcesentinelone.com/platform/purpleread 2026-09-28 |
|
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
Singularity Complete at $179.99 per endpoint per year (about $15 a month), and Commercial at $229.99. Enterprise, which includes the agentic AI SOC analyst, is quoted.
Per endpoint per year for the platform packages, plus Singularity Credits for AI work. The agentic AI SOC analyst is an add on for Complete and Commercial and is included in Enterprise.
What is public
Annual per endpoint list prices for Singularity Complete ($179.99) and Commercial ($229.99), the Enterprise inclusions (agentic AI SOC analyst), Singularity Credits as the AI currency, and a complimentary credit allotment to trial Agentic Investigation.
Billing mechanics
Platform packages are priced per endpoint per year and bought through authorized partners, who agree the final price. AI powered work draws on Singularity Credits.
Cost watchouts
The agentic AI SOC analyst is included in the quoted Enterprise package and is an add on for Complete and Commercial, and AI work draws credits, so the listed per endpoint prices do not cover agentic investigation volume.
Variable cost rationale
Two things are billed. Platform packages carry a fixed price per endpoint per year, so that part grows only with the number of endpoints. AI work, including investigations the agentic analyst starts on its own, draws Singularity Credits, so AI spend grows with alert volume, not with endpoint count. Credits come as an allocation, but neither its size nor the price of a credit is published, and the credits used per investigation are not published either. No cap or admin control on credit use is published.
Additional watchouts
Zero click investigations start without anyone launching them, and each one draws credits, so turning them on broadly raises credit use with alert volume. Neither the credit allocation nor the credits per investigation is published.
Sales call required
Mixed (some tiers require a call)
Free / trial
Complimentary Singularity Credits allotment to trial Agentic Investigation, no payment method required
Lowest paid plan
Singularity Complete at $179.99 per endpoint per year
Commercial notes
Public company (NYSE: S).
Key ambiguities
The price of a credit, credits per investigation and the Enterprise package price are not published.
Support SLA / resale
Sold through authorized partners, who agree the final price. Training services and guided onboarding and deployment advisory are offered.
Missing data
Credit rate and allocation, credits per investigation, and Enterprise package price.
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 SentinelOne
The closest documented capability profiles to SentinelOne among security and SOC agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- BlinkOps10.5 / 14Adds documented Memory & State Persistence
- Drata10.5 / 14Adds documented Testing, Debugging & Optimization
- Palo Alto Networks10.5 / 14Fuller documented coverage on Model Flexibility & Routing
- Swimlane10.5 / 14Fuller documented coverage on Model Flexibility & Routing
- Vanta9.5 / 14A lighter documented profile than SentinelOne
- Zeron10.5 / 14Adds documented Testing, Debugging & Optimization
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded