Let’s say you’re working from home and your laptop freezes. It’s unresponsive, it’s behaving strangely and all signs point to a virus.

What do you do?

Most probably, you’d call your IT security team right away, if you have one. You’re scared something might have happened to your data, or that your work is lost. But what if it wasn’t a rogue virus, but a concentrated attack? Suddenly your call is one of hundreds waiting in line for assistance — hundreds of worried employees, insurance policyholders or consumers trying to stay calm while listening to elevator music interspersed with: “Your call is important to us. Please stay on the line.”

For IT, an intentional attack is no different from a natural disaster, in the sense that it can come from nowhere with catastrophic consequences. And as in a natural disaster, there are hundreds, if not thousands, or millions affected — all at the same time. Cyber incidents are not confined to regions, nor are they selective about what language you speak, which makes incident support challenging from a traditional call centre viewpoint.

So now you’re held up on the call, just like everyone else, waiting and sweating bullets. What’s next?

Digital agents, that’s what. A digital agent is an artificial intelligence (AI)-powered, chatbot-based support system trained to respond like a real person, and in cases of a coordinated cyber attack, it can make or break the response effort. With a bottlenecked, overloaded call centre, the digital agent can jump in and put up the first line of defence against cyber attacks, fielding and routing an enormous volume of calls to appropriate channels after correctly identifying and authenticating the person and issue, performing basic information-gathering, and proactively suggesting possible triage solutions.

A digital agent can be a voice-operated or text-based digital assistant, but it’s much, much more than a simple chatbot. For one, the digital agent learns in real time. This means that if it has received several instances of an issue, it learns the appropriate response, and leads with it the next time the issue comes up. Second, a digital agent is able to quantitatively measure and report on factors that can reduce the severity or frequency of cyber attacks. So it not only learns how to be more reactive, but it can also be proactive, taking steps and measures to protect against cyber attacks when it detects a risk. Not only that, but users themselves can successfully govern the use of the product to reduce cyber risk, creating a threefold defence system that learns and grows with the evolving cyber attack threat.

Why an integrated digital agent? To be successful at first contact resolution, and provide a good first line of cyber defence, there must be:

  • Personalisation. Interfaces can make sure the dialogue is relevant and returns correct answers. This is done through APIs and orchestration engines that can reach out to back-office systems such as policy management systems to authenticate by policy number, ID, email or other means.
  • Seamless escalation. Digital agents will initially not be expected to remediate a cyber attack, but to help identify it, and pass information along through a series of coordinated hand-offs between legal, technical and psychological remedies. The digital agent gets the ball rolling. Key to all effective digital agent implementations is the ability to move conversations and logs from the digital agent to live chat or a call, create and update case logs and continue conversations where they left off.
  • Good conversation design and pre-built libraries of intents. These can calm the user and provide a realistic and conversational customer journey. Simple dialogue and sentiment such as: “Yes, we see your policy number, and we will continue to work with you to resolve your issue” sets the tone for a good customer experience. Pre-built libraries with Q&As and decision-tree intents are valuable in any industry, but particularly valuable for cyber responses to help get going quickly. Since there are different conversations for a ransomware attack, phishing or a confidential breach, it pays to have the intents available and tuned for both voice and text digital agents.

The combination of these factors can set up a near-impenetrable wall to guard against nefarious attacks, or at least a wall that learns to rebuild itself quickly and efficiently. And with digital agents performing closer and closer to their human counterparts (many callers already can’t distinguish between a digital agent and a human one), we’re looking at a brighter, safer future for cyber security.