Telesign’s Verify API goes all in with AI and ML to secure omnichannel growth

Omnichannel selling is now table stakes for any organization hoping to grow at scale, underscoring how critical it is to protect customer identities and fight fraud. Protecting identities and omnichannel revenue is a growing, complex challenge for every online merchant and e-commerce business. 

Adversaries that range from rogue attackers to nation-states are relying on generative AI to take their tradecraft to the next level to defraud online retailers and their customers at scale. 

The primary driver motivating adversaries to up their attack game is how fast omnichannel and e-commerce sales are growing. Global retail e-commerce sales reached an estimated $5.8 trillion in 2023 and are projected to grow at a 39% growth rate, surpassing $8 trillion by 2027. Omnichannel sales are among the fastest-growing areas of retail e-commerce, accelerated by AI-based shopping personalization. 

Merchant losses from online payment fraud are expected to exceed $362 billion globally between 2023 and 2028. Global business-to-consumer (B2C) e-commerce fraud losses are expected to grow by more than a 40% compound annual growth rate (CAGR) from 2023 to 2028.

Omnichannels are under siege 

The more successful a fraud attack is, the more it damages a brand. E-commerce fraud decimates brands, goodwill and trust, driving customers away to competitors. It’s on CIOs and CISOs to get e-commerce fraud detection and response right. Telesign found that 94% of customers hold businesses accountable and believe they must be responsible for protecting their digital privacy. 

Sift found that cybercriminals and fraudsters rely on AI and cutting-edge automation techniques that democratize access, resulting in new fraud-as-a-service offers. One of the most visible and highly subscribed is FraudGPT. Fraud schemes are becoming so pervasive that 24% of those surveyed report having seen offers to participate in account takeover schemes online. Sift also found that 73% of consumers believe the brand is accountable for ATO attacks and responsible for protecting account credentials. 

The Telesign Trust Index found that 44% of data breach victims tell friends and family not to associate with a brand that’s been breached, and 43% quit associating with the brand. Making matters worse for the brands experiencing breaches, 30% of customer victims share the incident on social media, further amplifying the event. 

Fighting online fraud with smarter APIs 

Adversaries’ tradecraft is getting stepwise boosts in performance from the combination of gen AI, ransomware-as-a-service and fraud kits, including FraudGPT, all of which are available across the dark web. Hardening Application Programming Interfaces (APIs) through the combination of AI and machine learning (ML) can reduce fraud by securing identities and transactions being shared across multiple verification channels. 

Taking an AI and ML-based approach to add contextual intelligence to APIs while consolidating omnichannel verification traffic in a single API streamlines transactions and cuts fraud risks. Telesign saw the need for AI-enabled APIs that could consolidate verification channels early and began working with their customers on the idea. Within a few short months, what began as a customer-driven concept became Verify API. 

In a recent interview with VentureBeat, Telesign CEO Christophe Van de Weyer explained how the company is using AI and ML to deliver its Intelligence API. “Machine learning has the power to constantly learn how fraudsters behave. It can study typical user behaviors to create baselines and build risk models. At Telesign, our Intelligence API, which can be paired with Verify API, uses ML to analyze phone numbers, email addresses, IP addresses and more.” 

When asked how Telesign is capitalizing on its expertise in using telephone numbers to verify identities in combination with AI/ML, Van de Weyer explained, “It helps identify red flags based on phone number activity and patterns. It looks for anomalous behavior by analyzing call velocity and duration, including looking at call patterns and usage to help flag risky numbers. This process informs the risk recommendation and score that Intelligence API provides, which can be used by a customer to better understand when to step up their authentication processes.

Why the Telesign Verify API defines the future of omnichannel verification 

Brendon O’Donovan, vice president of GTM Strategy at Telesign, explained in a recent interview with VentureBeat how Verify API capitalizes on Telesign’s machine learning expertise. “The Verify API works with our machine learning AI products, and we are continually working to find new turnkey ways to allow you to turn it on and recommend do you want to check the risk of this phone number before you send the OTP (one-time password).”  

Telesign’s Verify API is noteworthy in being the first omnichannel API that is embedded in the company’s broad base of AI/ ML algorithms to reduce the risk of fraud, increase identity security and reduce verification-related costs.

Here are the key areas that define how the Verify API is defining the future of omnichannel verification and e-commerce: 

• Integrates with seven leading user verification channels and chooses the optimal channel based on cost, experience and reliability by country or region: SMS, Silent Verification, Push, Email, WhatsApp, Viber and RCS (Rich Communication Services) into a unified API. With a single integration, Telesign’s Verify API enables businesses to effortlessly scale new authentication channels with minimal development resources.

• AL and ML algorithm support for all APIs ensures real-time responses to any omnichannel transactions and identity verification: AI and machine learning are integral components of Telesign’s new Verify API, as they enhance the platform’s ability to combat fraud and provide secure verification.

• Risk Assessment and Fraud Detection designed into the new API library: Telesign’s Verify API works in conjunction with machine learning AI products like Intelligence to deliver a phone number reputation score. This score is derived from machine learning algorithms that analyze global traffic patterns, phone data attributes, and comprehensive phone number intelligence. By evaluating these factors, the system can recommend a fraud score risk and identify unusual patterns that may indicate fraudulent activity.

• Anomaly Detection Based on Device Attributes: AI in the Verify API can detect anomalies based on the attributes of a device. For example, suppose a device that the phone number was moved to a new device recently or changed sim cards. In that case, the system can flag this as high risk and take appropriate action, such as not sending an OTP or requiring additional verification steps.

• Friction and Communication Channel Determination: Businesses use AI to determine the appropriate level of friction for a given transaction based on the assessed risk. This means that for higher-risk transactions, the organization may introduce additional verification steps. Conversely, for lower-risk interactions, the process can be streamlined to improve the user experience.

• Integration with Existing Fraud Models: Telesign’s API can integrate with a company’s internal fraud models, leveraging AI to improve protection against synthetic identity fraud, IRSF attacks, promotion abuse and more. This integration allows for a more comprehensive defense against various types of fraud.

Verify API’s vision is to deliver cost-effective, secure multichannel messaging at scale 

Throughout his interview with VentureBeat, de Weyer emphasized the dual nature of having AI/ML as a core part of the Verify API architecture. Reducing fraud risk and protecting identities is one, and reducing cost-per-message charges is the other. 

De Weyer told VentureBeat during last week’s interview that “SMS messaging is ubiquitous and relatively low-friction to use, but companies that spend a lot of them can face rising, and often unpredictable, cost-per-message charges, depending on the country (or countries) in which they are located. With Verify API, customers can choose the primary and fallback channels by termination destination to help control costs, improve customer experiences, and deliver messages on their preferred channel in a given country or region.” 

When asked how these innovations are delivered via the Verify API service customers, Van de Weyer said, “Customers can use Verify API to customize their verification and authentication across multiple channels, such as WhatsApp, Push, or email, to achieve a more stable cost-structure and a more secure experience. SMS is still a common fallback if and when a preferred channel fails. The overall goal is a more reliable cost structure, better customer experiences, and less fraud, regardless of where a customer is located geographically.”