AegisAI Perspectives · 084

Best email security for SaaS companies: how to choose

The best email security for a SaaS company judges what each email asks for, catches the lures aimed at developers and removes bad mail after delivery.

Best email security for SaaS companies: how to choose

The best email security for a SaaS company judges what each email asks for, because the lures that reach SaaS teams come from real platforms with SPF and DKIM passing20. AegisAI reads mail through the Microsoft 365 or Google Workspace API after delivery, with no change to mail routing (MX records)14.

At an AI company, one credential that matters is an API key for a sending platform; a Workspace mailbox and the approval behind an invoice matter too21. Third-party involvement in breaches has doubled to 30%9, and every integration a SaaS company ships or buys is one more third party.

What you can do by the end

  • Name the four attack types that reach SaaS inboxes, starting with lures aimed at developers.
  • Set the app consent controls in Microsoft 365 and Google Workspace before any product trial.
  • Run a side by side trial on your own mail and track three numbers before you sign.

Which email attacks hit SaaS companies hardest?

One compromise began with an email asking an open source maintainer to update his two-factor authentication credentials8. After that account was taken, malicious versions of at least 18 widely used npm packages were uploaded to the registry. Those packages have over 2 billion weekly downloads.

Since October 2024, one group has used social engineering, in particular voice phishing, to get into organizations' Salesforce accounts2. Authorizing a malicious connected app bypasses many traditional defenses such as MFA, password resets and login monitoring. A second group exploited compromised OAuth tokens for an AI chatbot app integrated with Salesforce.

LangChain saw fake emails impersonating CEO Harrison Chase asking for gift cards24. It also saw fraudulent contract attachments and spear phishing aimed at engineers with production system access. Business email compromise losses reported to the FBI reached $3,046,598,558 in 20251.

At Writer, one lure aimed at a developer was a sending platform API alert with SPF and DKIM passing, linking to a sign-in clone first seen the day it was used.

Why do sender checks miss these attacks?

Sender checks prove where a message came from. A DMARC policy of reject tells receivers to refuse mail that fails the DMARC check13. A lure sent through a genuine platform, or from a hijacked vendor mailbox, passes that check.

Eric Freeman, CISO at Writer, described the trade on the Writer customer page: "I was tired of choosing which way to be wrong."22

How should a SaaS company choose email security?

On Google Workspace or Microsoft 365, an API connection keeps mail flow as it is and lets a trial run beside the current stack.

A verdict only helps if the message leaves every inbox it reached, including invites that land on calendars16.

Each verdict should name the signal that decided it17.

How do you lock down app consent in Microsoft 365 and Google Workspace?

  1. In Microsoft Entra ID, open the user consent settings. By default, all users can consent to applications for permissions that do not require administrator consent10.

  2. Allow user consent only for apps from a verified publisher, as Microsoft recommends.

  3. In Google Workspace, set each app's access to Trusted, Limited, Specific Google data, or Blocked12.

  4. Review every third-party integration and, for each application, rotate its keys and tokens, as the FBI advises5.

  5. Repeat the review before any product trial, so the trial starts from a clean list.

Compare the four approaches by criterion

CriterionAPI-basedGatewayBuilt-in platform controlsManaged service
Lure from a genuine platformReads what the message asks forSender checks passSender checks passDepends on the tools the provider runs
OAuth consent and app grantsReads the mail that asks for the grantSees the message onceConsent settings per appDepends on the access you grant
Removal after deliveryRemoves from mailboxes over the APIHolds mail before deliveryAdmin search and purgeDepends on the access you grant
Mail routingNo MX changeMX record changeInside the mail platformSet by the tools the provider runs
Rules to writeAsk the providerAsk the providerPolicy settingsRun by the provider

What does a developer-targeted lure look like?

Illustrative example, figures are fictional. A backend engineer at a 120-person SaaS company gets an email from a real code hosting platform. The link leads to a copy of the sign-in page on a domain registered that morning. SPF and DKIM pass.

The Google Workspace guide and the page on BEC protection for Microsoft 365 cover the platform side in more depth.

How AegisAI handles this

AegisAI reads mail through the Microsoft 365 or Google Workspace API after delivery, with no change to mail routing (MX records). It starts read-only and runs in monitoring mode first; removal needs write access your admin grants and can revoke15. The median verdict takes 1.9 seconds, and a convicted message is removed from every mailbox it reached within seconds. Every verdict comes with its reason in plain language, so your team can check it and share it.

What did SaaS companies find in their own inboxes?

Two SaaS customers tell their stories on the customers page.

In its first week connected to AegisAI, Writer found more than 69 malicious emails sitting in employee inboxes with no configuration or threat hunting by the team. Its results table lists a +26% increase in recall over its existing stack. The full story is on the Writer customer page.

Ryan, LangChain's Head of IT, set it up in under five minutes, and the team did not write a single rule23.

How to measure it in 30 days

Track three numbers. The first is malicious emails found in inboxes in week one: in its first week, Writer found more than 69 malicious emails sitting in employee inboxes. The second is recall against your current stack: Writer reports a +26% increase in recall over its existing stack. The third is minutes from sign-up to protection: Ryan, LangChain's Head of IT, set it up in under five minutes, and the team did not write a single rule.

  • Review every third-party app with access to mail, files or your CRM, and rotate the keys you cannot account for.
  • Set app consent to verified publishers or admin approval in your mail platform.
  • Connect the product read-only, in monitoring mode, beside your current stack.
  • Count the malicious emails it finds that were already sitting in inboxes.
  • Pull every lure aimed at developers and check what each one asked for.
  • Turn on removal and time each message from delivery to removal.
  • Count false alarms on real customer and prospect mail and review each one.

Questions to ask a provider on the trial call:

  • Does the trial need an MX change, or can it start read-only through the API?
  • Is a convicted message removed from every mailbox it reached, invites included?
  • Does each verdict name the signal that decided it?
  • What did it find in week one at a company your size?
  • How long does setup take, and how many rules does it need?

The guide on how to evaluate email security vendors lists more trial questions. See your results in five days: connect a tenant, run AegisAI in monitoring mode, and compare it with what your current stack delivered25. The full Writer and LangChain stories are on the customers page.

Questions people ask

What is the biggest email threat to a SaaS company?

At Writer, a developer got an API alert with SPF and DKIM passing that linked to a sign-in clone. Business email compromise losses reported to the FBI reached $3,046,598,558 in 2025.

Is DMARC enough for a SaaS company?

No. A reject policy tells receivers to refuse mail that fails the DMARC check.

How do I stop OAuth consent phishing?

Microsoft recommends allowing user consent only for apps from a verified publisher. In Google Workspace, set each app to Trusted, Limited, Specific Google data, or Blocked.

Do SaaS companies need a gateway or an API-based product?

On Google Workspace or Microsoft 365, an API connection keeps mail routing as it is and needs no MX change.

How fast can a SaaS company deploy email security?

LangChain's Head of IT set it up in under five minutes.

Sources

  1. AegisAI customer story, Writer
  2. AegisAI inbound email security page
  3. Verizon, 2025 Data Breach Investigations Report release
  4. Help Net Security on the npm package compromise
  5. FBI FLASH on two groups targeting Salesforce
  6. AegisAI customer story, LangChain
  7. FBI IC3, 2025 Annual Report
  8. IETF, RFC 7489 (DMARC)
  9. Microsoft Learn, configure user consent
  10. Google Workspace Admin Help, control which apps access data
  11. AegisAI homepage

Recent posts

Every email, investigated.