Industry Spotlight: AB Handshake CEO Nadejda Papernaia On Leveraging AI to Battle Telecom Fraud

Fraud has been a part of the telecom landscape since, well, forever.  But the landscape has never shifted as quickly as it is now with the rise of artificial intelligence. Both threats and solutions have new, evolving tools available to them. With us today to talk about how things are changing and where they are going is Nadejda Papernaia, founder and CEO of AB Handshake. AB Handshake is moving to leverage predictive AI, collaboration, and regulation to shift the balance of power toward the defense of the network.

TR: What is your background, and what led to the founding of AB Handshake?

NP: I hold a master’s degree in physics, and this scientific and physical training defines the way I approach problems. Physics forces you to break down very complex systems, analyze how they behave, and solve them. That mindset has been invaluable for me in business, and particularly in telecom, because it is both technically complex and constantly evolving. Many, many years ago, I founded a telecom transit operator that became very successful and gave me deep insight into how international networks operate. It also showed me their weaknesses. This inspired me to found AB Handshake.

TR: What weaknesses did you see, and how did you approach fixing them?

NP: Fraud was everywhere, draining billions, and all the tools that we saw were reactive: like fighting fraud after the house was already burning. The idea was to flip the script. We created a system that validates every call in real time and then combines it with AI-driven detection. The result is proactive prevention rather than reactive defense. We think that fraud prevention is not just technical housekeeping anymore. It is becoming a pillar of digital trust. We started with defending our transit operator internally, providing anti-fraud to all the routes that we terminated. And then we saw that we could market it to other operators. So we spun off and reached out to the players in this field and started promoting our solution.

TR: What does the company look like today? Where do you sit in the ecosystem?

NP: We connect a lot of operators worldwide. We work with a lot of technology providers that sell our solution and market it. We’re also active members of GSMA, CFCA, i3forum, we actively contribute to ITU, and all other types of associations, which lets us ensure that our approach is aligned with international standards and widely operable. We are a neutral, trusted validation layer for voice and SMS. Our role is to protect operators from fraud, but also give them commercial upside. As an example, many operators now use our technology not just to offer fraud prevention, but also to use it as a service to their customers. It turns what used to be a cost center into a new revenue stream for them. They partner with us and they install our solutions into whatever they provide to their customers, saying “Powered by AB Handshake.” It is much more easier for them to sell it because they already have a trusted base of customers.

Also, beyond operators, we work directly with national regulators. We have a national anti-fraud platform that gives regulators a tool to monitor and block fraud at the countrywide level. It empowers governments to enforce industry-wide protection and also closes the gap that fraudsters exploit when operators fight fraud in isolation.

TR: How is the rise of AI changing the telecom fraud prevention space?

NP: It is changing the rules of the game. Fraudsters adapt faster to any static rule system. So basically, traditional fraud detection is like playing whack-a-mole. You only hit what you already know about. AI, by contrast, adapts dynamically, and spots anomalies in real time across vast and huge datasets. At AB Handshake, we adopted AI long before it was fashionable. We understood early that adaptive learning was the only way forward. But AI alone isn’t enough. That’s why we combine it with our real-time validation. Those two solutions reach very high accuracy and fraud detection together.

TR: How does your real time validation solution work?

NP: Say an AT&T customer is placing a call to a customer of a Spanish operator. On its way, the call goes through intermediate transit operators that can potentially spoof it and insert all types of fraud in it. So in parallel on another out-of-call channel, AT&T tells the Spanish operator they are sending a call from number A to number B and asks if they see this call. If the Spanish authority says, “No, I don’t see the call with this number,” then both know the call is spoofed. It’s done in a separate, encrypted channel, and it’s so quick that even before the call reaches its destination, the other operator already can be blocked– or not blocked. But AI fights very many different types of fraud that can’t be caught by validation, so it’s always the best thing to combine those two.

TR: What other types of fraud are common these days that AI can help detect and prevent?

NP: All types: wangiri, vishing, phishing, etc. There is one type called PBX hacking, in which a fraudster transits calls through a customer’s local PBX. We have seen scenarios where the company gets a bill at the end of the month with like 200,000 worth of calls that were tunneled through that PBX. What AI is doing now is real-time detection of unusual traffic patterns. It uses adaptive risk scoring to automatically adjust thresholds as fraud evolves. And it also uses predictive blocking: stopping fraud before it’s happening. But I think that the next wave will be even more exciting than this: topology discovery. Right now, the traditional fraud scheme has to be identified by an expert. Only then can a fraud prevention solution say, “Oh, this is wangiri. Block it.” But with AI, we will be able to identify new clusters of abnormal behavior, and flag them as an emerging fraud. It’s like antivirus software, which can spot new malware before humans even have a name for it. That could be the future. Another thing is cross-channel integration. Fraudsters can start with a flash call, then send a phishing estimate, then place a spoofed voice call. AI also allows intelligence to flow across all channels instantly and to change its watch list worldwide. The third thing that is becoming very valuable and important now is AI-driven operator systems, helping fraud teams that are flooded with alerts across different systems.

TR: How can AI help with those decisions?

NP: It is always a balance. Operators don’t know what to do with every alert because they want to save revenue. They don’t want to block things, but they also don’t want to lose something to fraud. AI can help prioritize them and make business recommendations about what to block. It can run what if scenarios and tell you the smartest move, whether it is to block it or put it through. For this we have something that’s already working, but it’s not a whole ecosystem yet.

TR: Is the other side also taking advantage of generative AI to come up with new threats?

NP: Yes, Fraudsters are already experimenting with generative AI. They use it for deep fake voices, spoofed calls, and very highly convincing scams. It raises the stakes significantly. But in practice, I think that defenders like AB Handshake hold a stronger position. I think predictive AI, especially when paired with real-time validation, can evolve faster than fraudsters can scale.

TR: Why would predictive AI be able to evolve faster than generative AI can scale?

NP: Because before they even invent something they have to try to experiment with it, and then we can detect it then to see if it is phishing or maybe some new type of fraud or new scheme or new algorithm. It’s easier to watch than to create. Generative AI will make scams more convincing, but predictive AI will make defense more precise. In the long run, prevention scales better than deception. The balance will lean toward defenders, provided that we invest in collaborative AI-powered ecosystems.

TR: Is the industry making those investments the way they should right now, or do we need to do more?

NP: All operators try to implement some type of fraud solution because it’s billions of losses, direct losses. Fraud also stops people from picking up the phone, which means operational losses too. If we add the national regulators to it, it will be even more effective. They can set up the rules; they can monitor things. Sometimes there are people in the companies who benefit from this fraud because they collaborate with fraudsters and send traffic and receive traffic for some remuneration. If the national regulator can monitor things, it can make it much harder for people inside an operator to continue to support fraudsters.

TR: Beyond the investments needed in predictive AI, what challenges lie ahead for fraud prevention further down the road?

NP: The most obvious challenge is scale, because telecom fraud drains nearly $40 billion annually. But the deeper issue is the fragmentation. Too many operators and vendors fight fraud in silos, and that creates gaps that fraudsters exploit. That’s the reason we built this national anti-fraud platform. When regulators and operators work together in a uniform framework, fraudsters lose.

TR: Why do you think operators often fight in silence?

NP: Because everybody has their own solutions. Everybody protects their own network. There’s always a gap that you can use between two operators. I think the future of fraud prevention is collaboration and also fewer borders. Fraud is global and our defense must be global too. The question is whether the industry and the regulators will align quickly enough.  At AB Handshake, our mission is to provide the tools and the ecosystems so that operators, partners, and regulators can fight fraud together. That’s the way it should be done, in my opinion.

TR: Thank you for talking with Telecom Ramblings!

If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!

Categories: Artificial Intelligence · Security · Unified Communications · VoIP