Telecom networks have spent decades playing defense against spam and scams—blocking bad numbers, blacklisting URLs, and reacting after damage is done. Tanla Platforms Limited believes that model is finally breaking down. And if the numbers coming out of Indonesia are even directionally accurate, the industry may be on the verge of a reset.
At an impact showcase hosted by Indosat Ooredoo Hutchison (IOH), Tanla revealed early results from Wisely AI, its AI-native communications security platform. After just six months of live deployment, the system has analyzed more than 11 billion communications across 100 million users, identifying over 2 billion spam and scam interactions in real time. Tanla estimates the intervention helped prevent as much as $500 billion in potential financial losses—a headline-grabbing figure that underscores the scale of the problem Wisely AI is designed to address.
Whether or not one accepts the full valuation of “losses prevented,” the takeaway is hard to ignore: AI-driven fraud prevention is moving from pilot projects to national-scale infrastructure.
AI at the Core of the Network, Not the Perimeter
Wisely AI isn’t a bolt-on filter or an app-level feature. Tanla positions it as an “AI nerve center” embedded directly into the telecom network, analyzing traffic as it moves—not after the fact. That architectural choice matters.
Traditional anti-spam systems often rely on static rules, manual intervention, or post-delivery reporting. Wisely AI instead evaluates sender behavior, content patterns, call-to-action signals, and URLs in milliseconds, before messages or calls reach subscribers. Tanla claims 99% model efficacy, with decisions made fast enough to avoid latency or service degradation.
Indosat has integrated Wisely AI across both its core network and handset-level applications, extending protection beyond SMS into voice, messaging apps, and now VoIP channels. The company also announced new features aimed at making its network effectively “scam-free,” a bold claim in an era when fraudsters rapidly shift tactics across channels.
For subscribers, the change is largely invisible—and that may be the point. According to Indosat, more than 95% of customers report feeling safer, with independent research showing improved customer experience scores.
Why Telcos Are Rethinking Scam Prevention Now
Spam and scam traffic isn’t new, but its economics have changed. Fraud has become industrialized, fueled by automation, cheap cloud infrastructure, and AI-generated content. Voice phishing (vishing), SMS scams, and malicious links are now coordinated across channels, often adapting faster than rule-based systems can respond.
Regulators are also tightening expectations. Governments across Southeast Asia, Europe, and the Middle East are pushing telecom operators to take a more active role in protecting consumers—not just providing connectivity. That creates pressure to deploy systems that work at scale, in real time, and across multiple communication modes.
Wisely AI appears designed with that regulatory and operational reality in mind. By operating at the network layer, it gives telcos a way to demonstrate proactive protection rather than reactive cleanup.
As Indosat CEO Vikram Sinha put it during the event:
“AI technology in our network helps filter threats before they can impact our customers.”
That framing—before impact—is increasingly where telcos want to be.
Big Numbers, Big Claims—and Why They Matter
The most striking figure Tanla shared was the $500 billion in estimated losses prevented. While such estimates naturally depend on assumptions, they highlight the sheer scale of scam activity moving through modern telecom networks.
Even a fraction of that number represents enormous economic and reputational risk. For operators, unchecked fraud doesn’t just hurt users—it erodes trust in the network itself. And trust, once lost, is expensive to regain.
Equally notable is the detection of 2 million malicious senders and CTAs, including URLs. That suggests Wisely AI is not only blocking known bad actors but continuously identifying new ones—an essential capability when attackers rotate identities faster than traditional blacklists can keep up.
The platform’s ability to process billions of interactions across 100 million users also signals maturity. Many AI security tools work well in controlled pilots but struggle under real-world load. National-scale deployment is where theoretical performance meets operational reality.
How Wisely AI Compares to Existing Approaches
Most telecom anti-fraud systems today fall into three categories:
- Rule-based filters, which are fast but brittle.
- Signature-based detection, effective against known threats but weak against novel attacks.
- Post-event analytics, useful for reporting but too slow to prevent harm.
Wisely AI aims to replace or augment all three by combining behavioral analysis, real-time inference, and continuous learning at the network level. That puts it closer to how hyperscalers approach security—monitoring patterns across massive datasets rather than relying on predefined rules.
Competitors in the space, including mobile OS-level spam detection and third-party messaging firewalls, typically operate at narrower layers of the stack. Tanla’s bet is that telcos want a unified, AI-first control plane rather than a patchwork of tools.
A Strategic Win for Tanla—and a Signal to the Market
For Tanla Platforms, Indosat’s deployment serves as a high-profile proof point. Founder and CEO Uday Reddy described Wisely AI as “first of its kind” at this scale, and emphasized growing interest from operators across Southeast Asia and EMEA.
That timing makes sense. Many emerging markets face higher exposure to scam traffic due to rapid smartphone adoption, diverse messaging ecosystems, and uneven digital literacy. Telcos in these regions are under pressure to differentiate on trust, not just price or coverage.
If Wisely AI continues to show measurable impact, it could shift how telecom operators evaluate security investments—from compliance-driven spending to competitive advantage.
What Comes Next
The next test for Wisely AI will be durability. Fraudsters adapt quickly, and AI models require constant tuning to avoid drift, bias, or blind spots. Transparency around false positives, user impact, and long-term performance will matter as deployments expand.
There’s also a broader implication: as AI becomes embedded deeper into national communication infrastructure, questions around governance, explainability, and oversight will follow. Telcos and vendors alike will need to balance automation with accountability.
For now, Tanla and Indosat are setting a benchmark—one that suggests AI-native platforms may soon become table stakes for telecom security, not optional upgrades.
If the industry has been waiting for evidence that AI can secure communications at national scale, Wisely AI’s early results are likely to fuel that conversation well beyond Indonesia.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
