Catch IT service desk vishing attacks with AI

Artificial intelligence can help save your company’s — and your customers’ — sensitive data, as well as your critical systems and networks, from weaponized voice calls. 

Many of these are calls to your IT service desk. They all start innocently enough. The caller, claiming to be an employee, explains that he is unable to access his account and has an urgent need to get in. Observing protocol the service desk agent replies that she would be happy to help — once the caller provides his name, company email and employee ID number. 

That’s when the game starts, as the caller tries to circumvent the company’s identification processes, even as the agent explains that the standards are in place for his protection, too: 

“Can’t you just give me access? I’m in a hurry.”

“Come on, this is wasting my time.”

“I’m giving a presentation in five minutes. You’re putting my job at risk. You don’t want to be responsible for that, do you?”

The service desk agent has been on the receiving end of a vishing attack, short for “voice phishing” attack. Similar to email-based phishers, vishing attackers impersonate a legitimate party to take advantage of someone’s trust — and even apply pressure tactics — to get in the door to steal valuable data, install malware or launch a ransomware attack. 

Vishing actually has been practiced on consumers for some time. You yourself may have answered calls from people purporting to be government officials or financial institution representatives who try to convince you to share your Social Security or credit card number with them on some pretext. Now vishing attacks on businesses are growing: One report states that vishing incidents accounted for 11% of all cyber incidents in 2025 — making it the second-most common attack vector following traditional phishing methods.

In many enterprises, service desk agents with access to confidential corporate information are easy marks for vishers. Service desk agents are instructed to provide good service, and are evaluated on their ability to help other employees, often in a quick, time-sensitive manner. It probably shouldn’t come as a surprise, then, that it was reported that a cybercrime collective is aiming to recruit women for voice phishing campaigns targeted at help desks. 

Unfortunately, service desk agents often haven’t been trained to recognize and respond to vishing attacks — which, as technologies such as AI mature, are becoming even more sophisticated in targeting these workers. Already, cyber criminals can harness AI-powered tools, such as Microsoft Copilot, and draw from social media sources to automatically pull together profiles of employees – information they can use to help them try to convince a service desk agent that they are legitimate callers.   

Vishing protection: To fight vishing, enlist AI

How to detect vishing attcks by training service desk agents to identify and block them is important. But even more compelling is embracing artificial intelligence to help them identify fraudulent vishing calls in real time, safeguarding their organizations’ confidential data and IT systems. 

Vishing attack prevention is possible. How to detect vishing with a new AI-powered fraud voice phishing detection system available from DXC Technology, for example, is one way. It monitors calls to service desks to detect known fraud tactics. If the AI detects a suspicious pattern, it flashes an alert with details on the service desk agent’s screen. After acknowledging that, the agent can ask more identity-related questions. Since this all takes place in real time, the agent also can message a manager to join in on the suspicious call. The call can be automatically terminated if the caller is unable to satisfactorily respond to a certain number of further queries pre-determined to have provided sufficient information for the AI to make the right judgment call.  

After the call, the system updates a contact trace record. This contains information the organization can use for further investigation, saving a tremendous amount of time on the process. 

The AI system also learns from each incident, adding new patterns to its inventory for future observation. That’s important because online criminals frequently update their tactics and ruses. Across DXC’s customer base, we receive more than 36 million service desk contacts a year — including an estimated 800 to 1,000 vishing attempts – giving our AI system a lot of data to learn from. We also offer our customers access to fraud detection training across all of our 60 support centers globally.

DXC’s AI-powered anti-vishing cybersecurity system is available to all existing customers using DXC Service Desk. The AI Fraud Detection also works with our Agentic Support Experience powered by DXC OASIS to provide protection, whether the contact is to a live or agentic AI agent.

• Support Services
• Cybersecurity