Abnormal AI
Behavioural AI email security, now stretching to identity and AI
Overview
Abnormal AI, which renamed itself from Abnormal Security in April 2025, uses behavioural AI to catch business email compromise, phishing and account takeover that rule-based gateways miss. It connects to Microsoft 365 and Google Workspace by API and in 2026 added identity security and an AI Security suite. Pricing is quote-only.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 14 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
Use cases
What you can produce with Abnormal AI
- Business email compromise, vendor fraud, phishing and account takeover caught and remediated in Microsoft 365 or Google Workspace
- API deployment with no MX record changes
- Identity Threat Protection that flags risky accounts and compromised sessions after sign-in
- Outbound email DLP rules reviewed by an AI triage agent
- Adaptive phishing-simulation training through AI Phishing Coach
- AI Governance discovery of sanctioned and unsanctioned AI tools and agents
Abnormal remains the reference point for API-based, behavioural email security: it deploys into Microsoft 365 or Google Workspace without touching mail flow, and it is built to catch the socially engineered messages that signature filters pass. The 2026 releases push it into identity threat protection, outbound DLP and AI governance, which makes it a broader platform bet than the old Abnormal Security name suggests.
The costs are opacity and scale: nothing is priced publicly, so smaller ASEAN organisations will need a sales conversation to learn whether it fits. IMDA accreditation from July 2026 makes it easier for Singapore agencies to buy. Shortlist it when email-borne fraud is the main risk and the budget is enterprise-sized.
What we found
Abnormal AI is the company this page used to list as Abnormal Security. On 16 April 2025 it announced a return to the name it had used when it started out in 2018, changing its legal name from Abnormal Security Corporation to Abnormal AI, Inc. and moving to the abnormal.ai domain; abnormalsecurity.com now redirects there. Chief executive Evan Reiser presented the change as a move beyond email towards a wider behavioural AI platform.
The core product is still email. Abnormal connects to Microsoft 365 or Google Workspace by API rather than through MX record changes, learns what normal communication looks like for each person, vendor and workflow, and pulls messages that break the pattern: business email compromise, vendor fraud, phishing and account takeover. 2026 brought a run of additions. July added Identity Threat Protection, AI Governance and Infiltration Prevention. August added custom detection rules, outbound email DLP checked by an AI triage agent, and adaptive phishing-simulation training. September assembled an AI Security suite, including AI Cloud Security investigations run on OpenAI Daybreak models, after Abnormal joined OpenAI's Daybreak cyber partner programme.
Most of the commercial picture comes from the company. Its last priced round was a US$250 million Series D in August 2024 at a US$5.1 billion valuation, led by Wellington Management, which took total funding to about US$546 million; no later round has been announced. In January 2025 it hired Lisa Banks, formerly of ServiceNow, as chief financial officer to prepare for a public listing, with no date set. Customer counts vary by page: more than 4,500 in its August 2026 releases, 5,000 on its homepage today.
For ASEAN buyers, Singapore's IMDA accredited Abnormal from 1 July 2026, opening the government Greenlane procurement track, and SoftBank began reselling it in Japan in September. It holds FedRAMP Moderate authorisation and is pursuing High. No prices are published; everything is quoted through sales.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
Notable facts
- Abnormal Security detects BEC attacks with 99.75% accuracy — compared to 60% for traditional secure email gateways — because it analyses behavioural anomalies rather than malicious links.
- The platform processes over 45 billion threat signals per month to build its behavioural models for each protected organisation.
- Business email compromise attacks cost global organisations over $55 billion from 2013 to 2023 according to the FBI — the exact threat Abnormal was designed to stop.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including Abnormal AI's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Abnormal AI unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Abnormal AI directly →
Spotted something out of date? Suggest an update →
More in Security & Safety
