Abnormal AI sells behavioral email and identity security software that autonomously detects threats, removes malicious messages, locks compromised accounts, and triages employee-reported email.
Abnormal AI develops a behavioral security platform for email, identity, and connected cloud applications. Its email products integrate with Microsoft 365 and Google Workspace to model normal communication and identity behavior, detect anomalous activity, remove malicious messages, lock compromised accounts, and automate employee-reported email triage. As of July 2026, the company reported more than 4,500 customers, including more than 25% of the Fortune 500.
Founded in 2018, Abnormal AI is a privately held company headquartered in Las Vegas, Nevada. LinkedIn lists its company size as 1,001 to 5,000 employees. The legal name changed from Abnormal Security Corporation to Abnormal AI, Inc. in April 2025. Its August 2024 Series D raised $250 million at a $5.1 billion valuation after the company reported crossing $200 million in annual recurring revenue.
Documented from public materials · these are the products the AMS scores
Inbound Email Security
Shipping since 2019
Inbound Email Security uses identity, relationship, behavioral, and content signals to detect business email compromise, impersonation, phishing, malware, and payload-free attacks. It can automatically remove malicious messages before users interact with them and supports investigation and bulk remediation across Microsoft 365 and Google Workspace.
Account Takeover Protection
Shipping since 2021
Account Takeover Protection learns normal sign-in, device, location, and email behavior for each user. It correlates identity and email signals to detect compromised accounts and can automatically sign out active sessions, block access, and force password resets, with an administrator option for manual review.
AI Security Mailbox
AI Security Mailbox
AI Security Mailbox autonomously classifies employee-reported messages, removes related malicious email across the organization, and sends policy-aligned explanations to reporters. The product supports conversational follow-up and connects to existing report buttons, abuse mailboxes, Google Workspace alerts, and third-party workflows.
2024 - 2026 ·gold = a monetization move· grey = product or capital
Abnormal announced Control Center, Email DLP Rules, and AI Phishing Coach upgrades. Control Center added custom AI models and rules, while Email DLP Rules introduced an AI Triage Agent for reviewing outbound policy matches. Some capabilities became generally available on August 31, 2026, while others entered early access. Source: Abnormal AI.
Hampton Roads Sanitation District approved a $75,590 one-year Abnormal subscription through CDW Government, with two renewal options and an estimated cumulative value of $226,770. The award covered 1,050 Inbound Email Security licenses bundled with core account protection, AI Security Mailbox, and Email Productivity, at a documented 4% discount from retail pricing. Source: HRSD commission agenda.
Pricing not public. Abnormal does not publish a standard rate card. The company has documented a per-user mailbox pricing model whose rate varies with the platform features purchased. Customers can procure subscriptions directly, through channel partners, or through a private Azure Marketplace offer.
Packaging is modular and can be bundled. The March 24, 2026 HRSD award combined 1,050 Inbound Email Security licenses with core account protection, AI Security Mailbox, and Email Productivity for $75,590 for one year, with two renewal options and a $226,770 estimated cumulative value. The agency documented a 4% discount from retail pricing. Public-contract estimate: dividing the first-year award by 1,050 licenses implies approximately $72 per licensed mailbox per year for that specific bundle. This is a derived estimate from the March 2026 contract, not a public list price, and it includes multiple modules and support.
Abnormal's cloud terms allow annual and multi-year subscriptions. Unless an order states otherwise, invoices are due within 30 days, later years of a multi-year term are billed on the anniversary date, and fees are non-refundable and non-cancellable. Usage exceeding an order's baseline quantity by more than 5% can be charged at the contracted per-unit rate on a prorated basis. No public evidence shows a separate meter for agent actions, investigations, messages processed, automated responses, or remediations.
Analyst layer · placement follows the documented capability above
Independence. The provisional L rating reflects documented workflows that can operate with under 20% routine human involvement. Inbound Email Security automatically removes malicious messages, AI Security Mailbox classifies reports and remediates related campaigns, and Account Takeover Protection can terminate sessions and force credential resets. Administrators retain configuration, exception-handling, and manual-review options, but routine detection and response can execute autonomously.
Job width. The provisional M rating reflects an end-to-end workflow inside the security function. The products cover behavioral detection, message or identity investigation, remediation, compromised-account containment, employee response, and audit visibility. The work spans more than one isolated task but remains centered on email and security operations rather than coordinating a cross-functional business process.
Output versus cost. The provisional Inflecting rating is based on documented customer and vendor evidence of material labor reduction. Abnormal reported in June 2026 that customers had reduced reported-email review time by up to 95%, based on internal company data. If comparable work volume is assumed, a 95% time reduction implies an estimated 20 times increase in analyst throughput, which falls within the 10 to 100 times Inflecting band. Separately, Mercury Engineering reported saving 90 security-team hours per month through automated investigation and remediation.
Monetization pattern. Abnormal monetizes through quote-based subscriptions tied primarily to licensed user mailboxes and selected platform modules. Public contracts show annual and multi-year commitments, channel procurement, bundled products, renewal options, and negotiated discounts. The autonomous capabilities are packaged inside products such as Inbound Email Security, Account Takeover Protection, and AI Security Mailbox rather than sold through an agent-specific consumption or outcome meter.
| Edition | AMS score | Metric on record | Change |
|---|---|---|---|
| Aug 2026 | L · M · Infl | Provisional AMS score calculated from public product documentation, September 2026 | Baseline, first edition |
Future editions add a row whenever the score or the metric moves, with the evidence that moved it.
About this data
Everything on this page is a synthesis of public information: published reporting, vendor documentation, and procurement data that anyone can access. We link to those sources throughout. Nothing on this page draws on private or confidential information, and the positions are Monetizely's analysis of that public record, refreshed monthly.