StrongestLayer launches tools to stop payment fraud and hijacked AI assistants

Tools to freeze fraudulent payments and hidden AI commands

StrongestLayer released two tools this week to stop email-based threats that most security systems ignore. Intent Workflows pauses payment requests, payroll updates, and executive account changes until someone verifies them. It zeroes in on actions that can’t be undone—like changing bank details or wiring large sums—and lets companies set rules, such as flagging every payment over $5,000 or matching requests against past transactions.

Alan LeFort, StrongestLayer’s co-founder and CEO, said these attacks don’t need malware or suspicious links. They masquerade as routine vendor emails asking for bank updates, tricking even trained employees.

Why traditional security misses the costliest attacks

Most email security tools scan for malware or phishing links, but that leaves a critical gap. StrongestLayer analyzed 32,052 attacks that bypassed defenses at 45 organizations in the second quarter. Just 3.6% of those attacks involved payment or bank-change requests—but they accounted for more than 98% of the financial exposure.

Even worse, 94% of these fraudulent requests passed sender authentication checks, meaning they looked legitimate to standard security tools. Instead of asking whether an email looks malicious, StrongestLayer asks what it’s trying to do. For example, an email requesting a bank update is automatically held for verification, even if nothing else about it seems off.

This approach doesn’t just catch fraud. It also cuts down on false alarms, sparing security teams from chasing emails that might look odd but pose no real threat.

Blocking hidden AI commands before they execute

The second tool, AI Instruction Intercept, tackles a newer threat: hidden commands in emails designed to exploit AI assistants like Microsoft 365 Copilot. Attackers embed prompts that order the assistant to leak data, execute unauthorized actions, or ignore security protocols—all without the user noticing.

Joshua Bass, StrongestLayer’s co-founder and chief product officer, said email is the easiest entry point for these attacks. "All an attacker needs is a valid address," he said. "Once the email lands, the AI assistant does the rest."

The tool scans each email against a company’s business context, using a reasoning model to spot suspicious instructions. Emails containing hidden commands are quarantined by default, with a clear explanation of what the instruction was attempting to do.

Law firm Orrick enforces manual checks for high-risk payments

Global law firm Orrick, Herrington & Sutcliffe uses the tools as both a customer and design partner. Eric Sanchez, the firm’s chief information security officer, said relying on employees to catch every fraudulent email was ineffective. "Even trained staff can miss subtle cues," he said. "Now, every irreversible payment request is flagged for manual verification by a designated owner, removing the guesswork."

Tools integrate with existing email security

Both tools are available now for Microsoft 365 and Google Workspace. They run alongside existing email security tools, so companies don’t have to overhaul their systems.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

Businesses can now stop fraudulent payments and hidden AI commands before they trigger irreversible actions—without replacing their current email security tools.

Key takeaways

• Intent Workflows halts high-risk payment requests until they are verified, targeting the 3.6% of attacks that drive nearly all financial exposure.
• AI Instruction Intercept blocks hidden AI prompts in emails, preventing attackers from hijacking assistants like Microsoft 365 Copilot.
• Traditional security tools miss these attacks because they focus on malware, not the intent behind an email.
• Law firm Orrick uses the tools to enforce manual verification for every irreversible payment.

Related on Lazyfounder

• Microsoft Corp.
• Google LLC
• More technology news
• Dell expands cyber resilience portfolio to cover identity, AI, and quantum-safe storage