Post Thumbnail

Beyond the Phishing Email: How Attackers Weaponize Trusted Workflows 

Kasey Cromer, Netlok | November 11, 2026 

Executive Summary 

The phishing email is not dead. But the assumption that employees can protect themselves by recognizing a suspicious message from an unknown sender is. The attacks documented in 2026 do not rely on that recognition failure. They route malicious content through familiar business workflows, trusted colleagues, legitimate software, and browser interfaces that look exactly like the real thing. The credential harvested at the end of a calendar booking workflow, a fake browser sign-in window, or a remote support session is just as useful to an attacker as one stolen through a crude spoofed email. What changed is how attackers get there. 

Microsoft detected approximately 8.3 billion email-based phishing threats in the first quarter of 2026 and approximately 7.6 billion in the second. [1][2] In that second quarter, attacks designed to steal login credentials accounted for 94 to 96 percent of all malicious email payloads each month. [2] The target has not changed. The delivery chain has. 

The shift matters for enterprise security leaders because the controls most organizations have invested in address the attack as it was, not as it is. KnowBe4’s 2026 phishing analysis found that 41.3 percent of phishing emails bypassed secure email gateways, reaching 48.3 percent in July. [3] When attackers route their lure through a trusted colleague, the message that arrives in the target’s inbox does not look like a phishing email because, technically, it is not. 

The Trusted Request Is Now the Attack Surface 

Fortra’s September 2026 research documents a campaign that illustrates how completely attackers have reframed the email threat. [4] An attacker poses as a prospective customer and contacts a non-sales employee at a target organization, asking to be connected with the sales team. The non-sales employee, performing a routine and helpful action, forwards a meeting-booking link to the appropriate sales contact. The sales representative receives what appears to be a credible business opportunity passed along by a known colleague. 

The booking workflow itself appears legitimate. The recipient selects an available time, completes the initial scheduling steps, and is then presented with a Microsoft 365 work-or-school sign-in prompt, framed as necessary to synchronize the calendar booking. The credentials entered at that prompt go directly to the attacker. 

No step in this chain requires the target to do anything unusual. The request came from a colleague. The booking workflow resembled real scheduling tools. The sign-in prompt resembled the Microsoft 365 interface the target uses every day. The attacker does not have to impersonate every trusted employee. They persuade one real employee to carry the lure across the trust boundary. 

Fortra separately documented that malicious calendar files processed automatically by Outlook can create events and reminders that persist independently of the original email. [5] Deleting the email or moving it to junk did not remove the calendar entry. Microsoft reported that this attachment type increased 277 percent in June 2026, though that figure reflects a single month within one category, not an annual growth rate. [2] Security teams that remove phishing emails without checking for calendar artifacts may leave an active lure in place after the message is gone. 

When the Browser Window Lies 

One of the more precisely engineered techniques documented in 2026 involves what researchers call browser-in-the-browser attacks: an attacker renders a fake browser window inside a phishing page, complete with a title bar, padlock icon, and address bar showing a convincing URL — all visual elements, not real browser controls. [6] 

Malwarebytes documented a 2026 campaign in which victims received a personalized YouTube copyright enforcement warning, complete with real channel branding. When the recipient clicked to respond, a fake Chrome sign-in window appeared on screen, visually indistinguishable from the genuine Google interface. The address bar within it displayed what appeared to be accounts.google.com — it was not. The entire window was rendered content on the attacker’s page. The real browser address bar remained visible at the top of the screen, but the designed visual focus directed attention to the simulated interface. [6] 

This technique is significant for enterprise security leaders because it directly undermines the visual inspection habits security training has historically reinforced. Users have been taught to check the address bar and look for the padlock. Browser-in-the-browser attacks copy those elements. The evidence users have been taught to trust becomes part of the deception. 

Legitimate Software as the Payload 

Fortra’s October 2026 research documented a 475 percent increase in phishing attacks incorporating remote monitoring and management tools during the first nine months of 2026 compared to the full prior year. [7] The campaigns targeted North American financial institutions. Victims received a call, email, or text message warning of an account problem and directing them to what appeared to be their bank’s website. A prominently placed live chat feature on that site triggered a download of a legitimate remote monitoring and management application, most commonly AnyDesk. 

An attacker then guided the victim through installation and requested the connection key that enables remote access. With it, the attacker had full control of the victim’s device. Fortra’s research identifies credential theft, ransomware deployment, and resale of access as documented subsequent actions. [7] 

The enterprise implication extends beyond financial services. Remote monitoring and management tools are software IT departments use daily. An attacker who persuades an employee to install one and share the connection key has achieved persistent remote access without exploiting a single software vulnerability. Fortra recommends that organizations allowlist the tools their IT departments support and block unauthorized installations of others. [7] 

GenAI Is Lowering the Cost of Production 

Attackers have adopted generative AI primarily as a site-building tool. Zscaler’s 2026 Phishing and Initial Access Report identified more than 413,000 AI-generated phishing sites, approximately 9 percent of which were classified as malicious. Documented uses included AI tools producing convincing replicas of government portals and credential-harvesting login pages. [8] 

KnowBe4’s October 2026 research separately reported a 228 percent increase in attacks impersonating AI brands between April and August 2026. [3] Security leaders evaluating controls should distinguish between AI-generated content and high-volume automated scripting: Microsoft’s second-quarter 2026 report documents a business email fraud campaign that reached more than 67,000 users across more than 42,000 organizations in under three hours using scripted message generation, not generative AI. [2] The scale of automation does not require AI to be significant. 

Who Attackers Are Targeting 

In one ransomware-group victim cohort analyzed by Zscaler in its 2026 Ransomware Report, 62 percent of victims held manager-level titles or above. [9] Approximately 75 percent of that cohort worked in finance, sales, operations, human resources, and marketing. These are the employees most likely to receive credible-looking business requests from external parties and most likely to have access to financial systems, operational data, and sensitive communications. 

The calendar phishing technique documented by Fortra is calibrated for exactly this population. A sales representative receiving a forwarded meeting booking link from a colleague is performing the daily work of their role. The attack does not ask them to do anything unusual. It asks them to do something routine. 

Photolok by Netlok: Authentication That Addresses the Credential Target 

The attacks documented above share a final objective: a valid credential the target organization would not knowingly approve. The calendar phishing campaign harvests a Microsoft 365 username and password. The browser-in-the-browser attack captures a Google account credential. The remote access tool campaign establishes device control that may subsequently be used to collect additional credentials. In each case, the attacker’s goal is to obtain something that logs in as a legitimate user. 

Photolok does not govern email filtering, block malicious calendar attachments, or restrict remote software installations. Those are controls for different layers of the security architecture. What Photolok addresses is the authentication event itself: the moment a credential is presented and access is either granted or denied. 

Photo-based authentication. Photolok replaces passwords with photo-based authentication in which each photo carries an encrypted code that changes with every login, is tied to the user’s registered device, and requires a server-side access code to complete. Even if an attacker captures those photos through a convincing fake login page, the encrypted code has already expired and the device requirement cannot be met from the attacker’s machine. What the attacker captures cannot be used to log in. [10] 

1 Time Photo. Users can configure up to five single-use photos for authentication. When at least one 1 Time Photo is configured, it appears on the second login panel by itself. Once used, that photo is no longer available, and the encrypted codes tied to it have already expired, making it inherently resistant to replay attempts. For employees who authenticate from locations outside the office, including those completing a sign-in from a public or shared environment after receiving a credible-looking calendar booking request, the 1 Time Photo provides an authentication layer that observation or interception cannot defeat. [10] 

Duress Photo. Duress Photos are silent alarms, randomly selected for display during login. Users can configure up to two. When selectin a Duress Photo, the login process proceeds normally, so an attacker who has obtained physical or social access does not see a failure. At the same time, selecting a Duress Photo triggers a real-time alert to the security operations team. The attacker sees a successful authentication. The security team sees a distress signal. [10] 

The Bottom Line 

The attacks documented in 2026 do not ask employees to click on something that looks dangerous. They ask employees to complete a calendar booking, respond to a copyright notice, or accept help from a support representative. Each of those actions feels appropriate in the moment because the attacker has constructed a chain in which every individual step is familiar. 

KnowBe4’s 2026 phishing simulation benchmark found a baseline engagement rate of 33.2 percent, rising to 39.5 percent for enterprises with more than 10,000 employees. [11] After twelve months of continuous testing and training, the global average dropped to 4.2 percent. Training changes behavior over time, but it does not eliminate susceptibility, and it does not address the scenario in which the message arrives from a trusted colleague rather than an unknown external sender. 

For enterprise security leaders, the question is not whether employees will recognize every attack. The documented evidence from 2026 suggests they will not, and the attacks are specifically engineered to avoid triggering that recognition. The question is whether the credential an attacker successfully harvests at the end of a trusted workflow can actually be used to authenticate. 

Request Your Personalized Demo 

Request a personalized Photolok demonstration from Netlok. https://netlok.com/contact-sales/ 

About the Author 

Kasey Cromer is Director of Customer Experience at Netlok. 

Sources 

[1] Microsoft Security. Email Threat Landscape: Q1 2026 Trends and Insights. April 30, 2026. https://www.microsoft.com/en-us/security/blog/2026/04/30/email-threat-landscape-q1-2026-trends-and-insights 

[2] Microsoft Security. Email Threat Landscape: Q2 2026 Trends and Insights. July 23, 2026. https://www.microsoft.com/en-us/security/blog/2026/07/23/email-threat-landscape-q2-2026-trends-and-insights 

[3] KnowBe4. New Threat Trend Data: When Bad Actors Use Reality Against Us. October 29, 2026. https://blog.knowbe4.com/new-threat-trend-data-when-bad-actors-use-reality-against-us 

[4] Fortra. CalPhishing Through the Trust Chain. September 28, 2026. https://www.fortra.com/blog/calphishing-through-trust-chain 

[5] Fortra. New Calendar Invite Phishing Campaign: ICS Abuse and Post-Delivery Persistence. May 14, 2026. https://www.fortra.com/blog/new-calendar-invite-phishing-campaign-ics-abuse-and-post-delivery-persistence 

[6] Malwarebytes. Fake YouTube Copyright Notices Can Steal Your Google Login. April 15, 2026. https://www.malwarebytes.com/blog/threat-intel/2026/04/fake-youtube-copyright-notices-can-steal-your-google-login 

[7] Fortra. Increase in Credential Phishing via Remote Management Tools. October 8, 2026. https://www.fortra.com/blog/increase-credential-phishing-remote-management-tools 

[8] Zscaler ThreatLabz. One-Click Compromise: ThreatLabz 2026 Phishing and Initial Access Report. June 10, 2026. https://www.zscaler.com/blogs/security-research/one-click-compromise-threatlabz-2026-phishing-and-initial-access-report 

[9] Zscaler. CXO Monthly Roundup: September 2026. October 19, 2026. https://www.zscaler.com/blogs/cxo-insights/zscaler-cxo-monthly-roundup-september-2026 

[10] Netlok. How Photolok Works. https://netlok.com/ 

[11] KnowBe4. 2026 Phishing Industry Benchmarking Report. July 7, 2026. https://blog.knowbe4.com/2026-phishing-industry-benchmarking-report 

More Articles