Book a demo →

Media library

AegisAI podcasts, talks and explainers

What Cy Khormaee, Ryan Luo and the AegisAI team have learned building an AI lab for email security.

12 recordings3 press articlesLatest: 5 Nov 2026

Press coverage

Press kit. AegisAI was founded by Cy Khormaee and Ryan Luo, former Google security leads. $49M raised: a $13M seed led by Accel and Foundation Capital and a $36M Series A led by Battery Ventures in July 2026.

Questions these episodes answer

How do calendar invite worms spread?

One email compromises an account, then the attack sends itself to every contact in it: investors to portfolio companies, companies to customers and vendors. It spreads across work and personal accounts alike.

Source: Trust hijacking episode, 10:00 ↗

Why do compliant companies still get breached?

Compliance is the baseline, not the finish line. MGM, Capital One and British Airways were all compliant when they were breached, and slow compliance processes can stop teams adopting the defenses they need.

Source: Trust hijacking episode, 30:09 ↗

Can security awareness training stop AI phishing?

Not on its own. Annual training cannot keep pace with attacks at machine scale, so better detection, in layers from email to endpoint, has to carry the load.

Source: Trust hijacking episode, 12:38 ↗

What is trust hijacking?

Attackers compromise whoever they can reach at scale, then use those trusted accounts (a friend, a vendor, a family member) to reach their real target. Your security is only as strong as your least secured contact.

Source: Trust hijacking episode, 5:55 ↗

Is MFA enough to stop phishing?

No. Relays, man-in-the-middle attacks and stolen device codes get around MFA. The answer is defense in depth that does not leave the decision to the person reading the email.

Source: Trust hijacking episode, 7:10 ↗

How do you catch phishing sent from a trusted company?

In one case SPF, DKIM and DMARC all passed. AI agents caught it by noticing the email did not match how the real sender's emails are built, then opening its links in a sandbox.

Source: Trust hijacking episode, 25:25 ↗

  1. 01Connect
  2. 02Investigate
  3. 03Protect
  4. 04Extend

Every email, investigated.