A finance employee at Arup’s Hong Kong office lost approximately HK$200 million (US$25.6 million) after joining a video conference in January 2024 that appeared to show the company’s UK-based CFO — a meeting Hong Kong police later said had been created from publicly available video and audio and was in fact a pre-recorded deepfake.
The Arup deepfake fraud, January 2024
The incident at engineering firm Arup began with an email ostensibly from the UK-based CFO asking for a series of transactions. The employee, initially suspicious, was convinced by a video conference that included what looked and sounded like the CFO and colleagues. Hong Kong police concluded the meeting was a deepfake built from publicly available material; the subsequent transfers totalled roughly HK$200 million (US$25.6 million). The case illustrates how synthetic media can substitute for verification and override an employee’s caution.
Marks & Spencer cyberattack, April 2025
In April 2025, British retailer Marks & Spencer (M&S) suffered a major cyberattack that disrupted operations for weeks: online orders were suspended, contactless payments and Click & Collect services were affected, and customer data was stolen. M&S initially estimated the incident would reduce its 2025/26 operating profit by around £300 million before mitigation. The attack was linked to the Scattered Spider cybercriminal group — a group the reporting ties to social engineering — and M&S later confirmed that attackers gained entry by deceiving employees at a third-party contractor rather than exploiting a direct technical flaw in M&S systems.

Audit-ready is a season. It shouldn't be.
Evidence in spreadsheets, controls drifting between audits, frameworks multiplying on flat headcount. Nubivance runs continuous compliance on Rapid7 Cyber GRC - SOC 2, HIPAA, ISO 27001, PCI, CMMC.
End the scrambleMicrosoft: impersonating IT support through Teams
Microsoft reported a campaign in which threat actors impersonated IT support staff using Microsoft Teams. Attackers contacted employees through external Teams messages or calls and used convincing pretexts — for example, urgent technical issues — to persuade staff to hand over remote access through screen sharing or legitimate support tools such as Quick Assist. Once connected, attackers used PowerShell to install malicious software, establish persistent access and gather information, and Microsoft observed attempts to move laterally toward critical infrastructure including domain controllers and certificate authorities.
Trust replacing verification: common techniques and hard data
These three incidents share a through line: attackers substitute signals of trust for genuine verification. Deepfakes recreate familiar faces and voices; helpdesk impersonation exploits routine processes; third-party relationships widen the attack surface. The underlying risk is not merely technological novelty but the human and procedural reliance on trust. Verizon’s Data Breach Investigation Report is cited in the reporting to show how frequently credential compromise plays a role in breaches: stolen credentials were involved in 44.7% of breaches. That statistic sits alongside the incidents above to make a plain point — attackers rarely need new zero-days when they can persuade legitimate users or outsourced staff to grant access.
What this means for technologists, procurement leaders, and end users
- Technologists and security teams: The story directs attention to embedding reliable verification into service desks and automated workflows. The report highlights measures such as phishing-resistant multifactor authentication, least-privilege access, and monitoring for suspicious account activity. It also names specific tooling approaches — for example, using more robust verification factors and integrating them into support procedures — to prevent social engineering from converting a routine call into a compromise.
- Procurement leaders and third-party risk managers: The M&S case makes explicit that a provider’s verification practices can be the weak link. Organizations with strong internal controls may still be exposed if contractors use weaker identity checks; the reporting stresses that third-party processes deserve the same scrutiny as direct suppliers.
- End users and helpdesk agents: The incidents underline the limits of awareness training. The reporting notes that even well-trained staff can be deceived by convincing impersonation, and argues organizations should give employees dependable verification tools and clear escalation paths rather than leaving judgement calls to individuals in the moment.
Specops Software — the sponsor of the reporting — is presented with concrete product recommendations intended to address those gaps: Specops Secure Service Desk allows agents to verify callers with more than 15 MFA factors (including options that do not require a phone) and integrates with identity platforms such as Duo and Okta; the Specops Authentication API is described as a way to add workforce authentication and identity verification into internal applications, chatbots and automated workflows. The reporting also recommends requiring independent confirmation for sensitive financial and privileged actions and building verification checks into self-service portals and automated processes.
The record in these cases is stark: synthetic media, social engineering and weak third-party verification have produced large, measurable losses and operational disruption. The practical implication is clear from the incidents themselves — verification must be built into the routine tools and processes employees use every day, and into the systems and controls of the contractors who support them.




