Abnormal AI
Verified Oct 16, 2026 ActiveBehavioural AI platform that catches email fraud, account takeover and rogue AI agents by learning what normal looks like
Listed on EffectHub ·
Pricing
- Pricing model
- Contact sales
No public price on the site: every call to action is a demo request, and the email, identity and AI security modules are quoted by sales as enterprise contracts.
Use tool ↗About Abnormal AI
Abnormal AI models the normal behaviour of every person, vendor and identity in a company and flags the deviations that signature filters miss: business email compromise, vendor invoice fraud, account takeover and, since 2025, misuse of AI tools and agents. It connects by API to Microsoft 365, Google Workspace, identity providers such as Okta and Entra, cloud platforms and AI providers, so there is no mail gateway to insert and no MX change. The platform is sold in modules for email, identity and AI security on top of a shared behaviour engine, with an AI-run security mailbox and phishing coach for employees. Founded in 2018, the company returned to its original Abnormal AI name in April 2025 and sells only through sales-led demos.
- Platforms
- Web · API
- API
- Yes
- Last verified
- Oct 16, 2026
Verified sources
- Home page presents the Behavioral Security Platform with email, identity and AI security modules on a shared behaviour engine, integrating via API with Microsoft 365, Google Workspace, Entra, Okta, AWS, GCP, Azure, Anthropic, OpenAI and Copilot, activating with no agents and no MX change; only call to action is Request a Demo; copyright 2026 Abnormal AI, Inc. (15 October 2026)abnormal.ai ↗
- About page: founded in 2018 by Evan Reiser, more than 30% of the Fortune 500 as customers (15 October 2026)abnormal.ai ↗
- Press release dated 16 April 2025: Abnormal Security rebranded as Abnormal AI, legal name changed from Abnormal Security Corporation to Abnormal AI, Inc., new domain abnormal.ai, a return to the name used at inception in 2018 (read 15 October 2026)abnormal.ai ↗
Is Abnormal AI legit?
We don’t label tools safe or unsafe. Below are the records we checked ourselves, each with the date we read it.
- Domain registered
- Dec 15, 2017 (8 years ago)Registry record (RDAP), registrar MarkMonitor Inc., looked up Sep 15, 2026
- TLS certificate
- Issued by DigiCert Inc, valid until Nov 8, 2026Seen in our own connection to the site on Oct 16, 2026
- Last checked
- Oct 16, 2026The site answered
- On the market since
- 2018The vendor's own material, read by us
- Operated by
- Abnormal AI, Inc., United StatesOur research, card verified Oct 15, 2026
Pros & cons of Abnormal AI
Abnormal is the behavioural-baseline approach to email security done at enterprise scale: it reads the API feeds of your mail, identity and cloud tenants, learns who talks to whom and how, and blocks the socially engineered message that passes every content filter. It is a sales-led enterprise product with no price list and a widening scope into identity and AI-agent governance, so it suits security teams with budget and a Microsoft 365 or Google Workspace estate, not a ten-seat shop looking for a spam filter.
Pros
- API deployment with no MX change, so it runs alongside an existing gateway and can be switched on in minutes
- Behavioural detection targets the payload-free attacks (BEC, vendor fraud, lateral phishing) that content filters miss
- Identity and AI security modules extend the same behaviour engine beyond the inbox
- Named legal entity with a public rebrand history and Fortune 500 references, which simplifies procurement
Cons
- No public pricing, trial or self-service tier — evaluation starts with a sales call
- Headline accuracy claims (46x fewer misses, 21x faster) come from the vendor's own testing, flagged as illustrative in its footnotes
- It needs broad API access to mail, identity and cloud tenants, which some security and privacy teams will resist
- The platform is widening fast into identity and AI governance, so module scope and pricing can shift between renewals
Best for
- Enterprises on Microsoft 365 or Google Workspace that already have a gateway and still lose money to invoice and payroll fraud
- Security teams that want account-takeover and identity signals in the same console as email
- Organisations rolling out Copilot, ChatGPT or Anthropic models that need usage guardrails on the same platform
Not for
- Small businesses that need a published per-mailbox price and self-service setup
- Teams that want an on-premises or gateway deployment rather than API access to their tenant
Watch out: The company's legal name changed from Abnormal Security Corporation to Abnormal AI, Inc. in April 2025; older contracts, reviews and security questionnaires may carry the old name and domain.
Alternatives to Abnormal AI
Abnormal AI head-to-head
Side-by-side fact sheets against the competitors our research or their makers name — pricing, platforms, API and both verdicts on one page.