Unicode Trick Fuels Million-Email Phishing Surge

Microsoft is flagging what it describes as a high-volume phishing operation that leans on a quiet, adaptable trick: invisible Unicode characters inserted into financial keywords. The characters render as nothing to the human eye, but they break up words like "funding" so detections looking for a literal string never see it.

Invisible Disruption

The attack uses ASCII Smuggling, a method in which non-rendering Unicode characters conceal content inside text that appears ordinary. Human interfaces skip these code points, so an email recipient sees natural language. But an email filter or large language model ingesting the raw bytes may treat the hostile content as real text, opening the way for prompt injection if the system cannot reliably separate intended instructions from embedded text.

Microsoft's security research team reports that instead of hiding commands from people while exposing them to AI models, the attackers here injected the characters into common financial lure words to split them apart, preventing email filters from parsing a contiguous keyword.

Financing the Deception

The most abused range is the Unicode Tags block, U+E0000 to U+E007F, according to Microsoft. This block carries a shadow copy of printable ASCII characters — U+E0041 mirrors "A," U+E0061 mirrors "a" — and was designed for language tagging before becoming largely deprecated.

For instance, a finance-related lure term such as "funding" becomes "fun⟨U+E0020⟩ding" in the email. The recipient still reads "funding," and any parsing pipeline that normalizes away the invisible characters sees the same word. But a detector searching for the literal string or a regex that doesn't account for interleaved invisible code points will not find a contiguous match.

"To a recipient, and to parsing pipelines that drop or normalize these characters, the word still reads as funding," Microsoft explained. "To a detector matching the literal string funding, or a regex that does not account for interleaved invisible code points, the byte sequence no longer contains the contiguous keyword."

Millions of Messages

The campaign entered a high-volume phase that lasted roughly three months before dropping sharply after May 15, 2026. It followed a weekly cadence, nearly going silent on weekends and resuming in full swing on Mondays.

Volume reached an estimated 1 to 2.37 million messages on weekdays, peaking on February 26, 2026.

• 1 to 2.37 million weekday messages
• Peak on February 26, 2026
• Roughly three-month high-volume phase
• First emerged early February 2026

Attack patterns built on this approach are said to have first appeared in early February 2026, per Microsoft.

Beyond the Keyword

Microsoft ties the activity to a broader phishing campaign that weaponized the ActiveCampaign marketing and automation platform to distribute thousands of AI-generated phishing emails targeting Small Business Administration (SBA) loan applicants. Fortra's Intelligence and Research Experts (FIRE) team disclosed details of that operation in September 2025, noting the focus on collecting detailed business and financial information, likely to enable highly targeted spear-phishing in future attacks.

"The campaign's sophistication and uniqueness lies in the ability to mass-produce convincing, tailored websites that adapt to different illegitimate or impersonated domains," Fortra noted at the time. "Threat actors are able to scale sophisticated phishing by using ActiveCampaign's AI-powered marketing automation features to vary the design, content, and flow, ultimately creating more convincing phishing campaigns, quicker."

Borrowed Reputation

The campaign uses hundreds of disposable, finance-themed sender domains with lures mimicking business loan, line-of-credit, and advance-funding phishing patterns typical of fraud or credential-harvesting schemes. The top 10 sender domains by most hits include guardiangrowthfunding[.]com, digitalcapitalboost[.]com, thebusinessloanexpress[.]com, yourlocfunding[.]com, advancefundingboost[.]com, guardiancapitalway[.]com, harboradvancefunding[.]com, unitedfundingwave[.]com, directcapitalboost[.]com, and onlinedirectfinance[.]com.

Those finance-themed domains relay through ActiveCampaign, routing every outbound link through ActiveCampaign's own click-tracking domains ("acemlnd[.]com" and "activehosted[.]com").

ActiveCampaign said it has tested its content-moderation systems with messages containing invisible Unicode characters and that such emails receive the same moderation verdict as their unobfuscated equivalents. It also treats heavy use of the technique as a "suspicious signal."

Lookalike Tactics, New Scale

Using invisible or look-alike characters is not new in phishing and homoglyph attacks. What stands out here is the choice of the Unicode Tags block and the campaign's scale, generating multi-million messages daily.

"As with any shared sending service, attacker abuse of customer accounts or workflows can complicate reputation-based filtering," Microsoft said. "By originating from a reputable marketing platform with established IP reputation and authentication, the activity may appear more similar to legitimate marketing traffic and can complicate reputation-based filtering."

Why It Matters

This campaign shows evasion techniques crossing over from the AI security world into ordinary spam pipelines, which could mean defenders need to broaden what they watch for. The reliance on a reputable platform's infrastructure suggests that even trusted senders can be used to deliver attacks, potentially lowering the bar for what passes initial checks. For organizations, it hints that keyword-based detection alone may no longer be enough, and that filtering pipelines may need to account for interleaved invisible characters to stay effective against similar ploys.