Q: What sets Trully apart from existing solutions in the market, especially in terms of financial fraud prevention and identity theft? 

EB: Trully addresses the inadequacy of traditional fraud prevention methods used with emails, phone numbers, and fingerprint devices, in which we observed recurring cases of identity theft that significantly impacted various industries. The need for a unique identifier led us to focus on the face as the key to synthesizing digital identities. What sets us apart is our ability to work with various tools that identify both faces and documents, generating a list of identities with positive and negative behaviors. Through a collaborative network, we analyze the behavior of these digital identities, identify patterns, and provide a differentiated solution for identity validation in the financial industry and other sectors.

FP: The financial industry continues to rely on traditional methods like comparing official IDs with a person's face. These methods are vulnerable to false IDs. In contrast, Trully uses information gathered during the digital customer onboarding process to thoroughly analyze the data obtained from photos and IDs. Unlike other companies conducting simple verifications, we delve deep into data to gain a holistic understanding of the customer, preventing identity theft through collective intelligence.

Q: Of all the digital interactions analyzed by Trully, what percentage turn out to be false or fraudulent? 

FP: We observed that fraudulent digital interactions increased by 54% in Mexico during 2021. This percentage has significantly increased with digitization, as digital processes have facilitated the rise in fraud. We observed cases where the same person obtained up to 150 credits with the same financial institution simultaneously, resulting in substantial losses for the institution. Our work with various companies reveals that between 3% and 5% of interactions contain previously unidentified fraud even if they had a KYC provider in place. This figure is three times higher than what institutions typically estimate in their customer portfolios. Fraud often hides within delinquent portfolios, making detection challenging and leading companies to erroneously attribute it to credit risk instead of recognizing it as fraud. Fraud is on the rise, driven by advances in AI and the increasing sophistication in creating synthetic identities.

Q: How does Trully’s collective intelligence network work, and what advantages does it offer to companies using this specific solution?

EB: The concept of collective intelligence is not new, but applying it to fraud prevention is an innovation. In cybersecurity, tools like Threat Intelligence Sharing Systems share indicators of compromise, such as IP addresses or malware samples. Trully applies this concept to fraud prevention, using faces as indicators, as a facial hash is unalterable. The fundamental advantage it offers is its ability to mutually enrich and protect collectively. By sharing information about fraudsters’ methods and attacks, we create a robust collaboration network. Trully anonymizes data so no company knows who was affected by a particular fraudster. Each organization contributes to this collective intelligence network, strengthening Trully’s network and allowing the effective identification and blocking of fraudsters.

Q: How is Trully using artificial intelligence and other emerging technologies to prevent identity theft and financial fraud?

FP: Trully employs over 15 internal models for various methods and purposes. In line with privacy concerns, we provide a platform that allows companies to securely share data. For example, we use AI in photo processing to automatically extract over 25,000 facial features in milliseconds through internally trained neural networks. This translates to a vector of visual features, a non-reversible numerical set preserving customer privacy and used for comparisons through big data. We compare a face against millions of digital identities in seconds, leveraging intensive computing, big data, and AI. Additionally, we have identification models that, through machine learning, determine the authenticity of an ID. Finally, we implement emerging neural graphs, which map interconnections between data to quickly assess how many faces are associated with an email. This AI-based technique is used to rapidly go through large amounts of data and is supported by graph-trained models to have an effective classification of digital identities.

Q: What challenges has Trully faced in dealing with data privacy regulations in Mexico?

FP: The main challenge has been education, especially for clients. The request to share data for proof of concept was initially met with concern. However, there is no regulatory risk as we offer a secure method to share data that protects the end customer's privacy and our process. Therefore, education has been crucial to make companies understand that they need collective solutions. We constantly explain to our clients that collaboration is essential for fraud prevention because an individual network is not enough if a company has already fallen victim to fraud. Although educating people can be difficult, Trully’s growing reputation supports our ability to provide secure and effective solutions.

Q: What recent updates and improvements will Trully introduce in the near term?

EB: We are launching the "Magic Link" feature, which allows companies to streamline the customer origination process from their initial application to the final decision or recommendation. Through a customizable link, the user can complete the entire data collection process needed for credit evaluation. This process performs email validation, phone number, IP address, location, face, and document validation through a single integrated and simplified link. This functionality allows companies looking to determine the legitimacy of an applicant to do so efficiently and quickly, without adding a single line of code or undergoing long and difficult integrations. Trully makes the integration process incredibly easy.

Q: What challenges has Trully faced in the Mexican market regarding financial fraud identification?

FP: The challenges have changed in the last year due to the malicious adoption of AI, but we remain focused on constant innovation. We are observing that fraudsters make numerous attempts instead of only one or two. This presents a significant challenge as fraudsters seek creative ways to bypass detection systems. For example, we had a case in which the same face was detected 150 times applying to the same company. In another company, the same face was observed in 14 attempts using different identities.

EB: This scenario underscores the issue Trully addresses: identifying when the same individuals use multiple identities against the same company. This also affects the original victim whose identity was stolen, potentially making them unable to access other financial instruments.

Q: Besides solutions involving biometric data, what other available measures could be effective in identifying and preventing potential financial fraud?

FP: Biometric solutions, though valuable, are not the sole answer to identifying and preventing financial fraud. Criminals are constantly looking for new ways to bypass security systems, and no single measure can be entirely effective. Companies must adopt holistic, 360° approaches to address this ever-evolving challenge. Companies need to be aware that they cannot rely solely on redundant security measures, as fraudsters will always find different ways to evade barriers.

Constant innovation and adaptation are essential. Trully believes in creating connections between companies to report and share information about fraud attempts, as collaboration allows for a more robust and efficient system for fraud identification. Decision-making based on quality labels and the full exploitation of large amounts of data will be crucial in the near future. Companies must be connected in a system where they can share and report information, enabling the retraining of AI models to enable a much more accurate probabilistic assessment of fraudulent behavior.

EB: In cybersecurity, if an IP address is blocked by a firewall, the criminal simply changes the IP address and continues attacking. The same happens in fraud prevention: getting a new phone number, email, or IP address is quick and accessible, so traditional strategies based on data like emails and IP addresses are insufficient to prevent fraud. Trully’s approach focuses on the effective exploitation of data, backed by our expertise and knowledge in cybersecurity.