AI & ML | The Ultimate Defense Against Threats to Enterprise Messaging Ecosystems

Spotlight on AI & ML | TelecomDrive.com

In today's digital age, enterprise messaging has emerged as a vital communication channel for businesses to engage with their customers in the most cost-effective way possible.

In today's highly competitive markets, messaging provides businesses with a channel to drive the lifetime value of customers through highly interactive and engaging communications tailored to each individual's unique identity and demands. Similarly, the growth of messaging has allowed operators to create new sources of revenue besides rising up the value chain in the messaging economy. Since the messaging opportunity is so critical for operators, as well as the enterprise segment, there is a growing interest in AI & Machine Learning to ensure the continued growth and health of the overall messaging ecosystem.

However, this has also attracted a significant number of spam calls and messages, causing inconvenience and frustrations for individuals. According to a recent extensive survey from LocalCircles, about 92% respondents in India said they get unwanted/ pesky calls despite being registering on the Do Not Disturb (DND) list. To address this issue, the Telecom Regulatory Authority of India (TRAI) has also recently directed telcos to deploy artificial intelligence (AI) and machine learning (ML) technologies to detect and prevent spam calls and text messages from unregistered telemarketers.

The Rising Demand for Enterprise Messaging

The increasing mobile penetration is one of the most important factors driving messaging popularity. According to TRAI, India has over 114.1 crore mobile subscribers. Also, there are over 600 mn smartphone users in India which is also growing at a rapid pace.

Secondly, SMS has made B2C communications easier. With short messaging services, businesses can reach out to anyone with a mobile phone. SMS is also a high ROI messaging channel, where SMS open rates are measured in seconds. Studies have shown that four out of every five customers will read an SMS within 30 seconds, which is faster than any other medium. When we compare this number to email open rates, it is easy to see why SMS has become so important for enterprise communications today. Third, the advancement of analytics, combined with customers' willingness to share their data in exchange for better service, has made it easier for enterprises to understand the impact and ROI of each messaging platform, as well as fine-tune it to different customer behaviors and requirements.

Opportunity for Telecom Operators

Operators are facing a decline in their voice and SMS business, making it necessary for them to explore new sources of revenue. One promising avenue for operators is A2P (application-to-person) communication, which ensures steady revenues, especially with the rapid expansion of the app ecosystem. However, before operators can fully capitalize on the A2P opportunity, they must address the challenge of Grey Routes.

To comprehend this issue, it is important to distinguish between P2P (person-to-person) messages, which involve the exchange of SMS messages between individuals, and A2P (application-to-person) messages, which involve the exchange of SMS between an application and an individual. The problem arises when an A2P message is disguised as a P2P message, primarily to avoid A2P termination charges or to conceal the sender's identity for spamming purposes. Various methods are employed to mask an A2P message, including GT spoofing and the use of SIM farms. GT spoofing involves altering the message's global title to hide its true identity. SIM farms, on the other hand, involve accumulating a large number of SIM cards and using them to send out A2P messages while pretending to be P2P messages.

When enterprises or aggregators try to send commercial messages via illegitimate or zero-rated routes, it is known as grey routes. Grey Route compromises the ability of the operator to monetise the messaging opportunity leading to operator losses running into billions.

 In addition to financial losses, Grey Routes have a negative impact on an operator's ability to ensure high-quality traffic flow on its networks. The lack of mechanisms to differentiate between legitimate and unauthorized traffic prevents the operator from prioritizing message delivery effectively. As a consequence, congestion in traffic can occur, resulting in delayed message delivery, particularly in critical sectors such as banking. For instance, a customer expecting immediate notifications for every ATM withdrawal may experience slower message delivery. In such cases, it is the enterprise that bears the responsibility for managing dissatisfied customers. Moreover, if the messaging system is exploited for spamming purposes, it undermines the operator's reputation and credibility.

The Way Forward

Operators have traditionally relied on rules-based SMS firewalls to protect their networks from misuse. These firewalls utilize techniques such as blacklisted numbers, keyword searches, and URL destinations to categorize messages. However, determined scammers have become adept at circumventing these traditional detection and prevention methods, which are based on deterministic rules, limited pattern searches, and blacklists. Another drawback of these deterministic platforms is their lack of 100% accuracy, resulting in potential blocking of legitimate traffic that happens to meet the platform's criteria. This can negatively impact the customer experience, causing them to miss out on promotions and leading to an overall poor experience. Therefore, to address these issues and improve the messaging ecosystem, it is necessary to adopt a more nuanced approach to the problem.

In this context, the use of AI capabilities offers a more comprehensive approach. By harnessing advancements in natural language processing, an AI-based SMS firewall can automatically classify messages into different categories. Unlike conventional platforms that rely on limited pattern searches, the AI platform utilizes extensive training data from millions of similar messages. It analyzes the words within messages using pattern matching techniques and considers the context in which those words are used to accurately predict the message category. Once messages are categorized, operators can implement policy controls at a more granular level, effectively protecting subscribers from spam and fraud, preventing revenue losses, and reducing operational efforts. This approach also helps minimize subscriber churn within the network, resulting in improved overall efficiency and customer satisfaction.

The article is published in the June 2023 issue of Disruptive Telecoms.