Fraud is being automated at industrial scale — and, for once, the defenders appear to be keeping pace. That is the core takeaway from Infobip’s Fraud and Security Trends 2026 report, which the global cloud communications company built from an analysis of billions of interactions across its network over the past year.
The headline number is a blunt one: threats detected across Infobip’s platform climbed 77% year over year, as scammers lean on generative AI to write, translate and personalize fraudulent messages faster than any human operation could. But the same report shows enterprises responding in kind. Adoption of AI-powered fraud detection rose 71% over the year, while use of pattern-based (signature) detection more than doubled, up 105% — a sign that companies are shifting toward adaptive security that can move as fast as the threats.
Phishing dominates, and the network layer is the new battleground
Phishing accounted for 49% of all malicious content Infobip blocked, with volumes up 94% year over year — nearly doubling. Just as notable is where the fight is moving. Use of network APIs, including SIM Swap checks and phone-number verification, jumped 91%, with the financial sector the heaviest adopter. In other words, banks and fintechs are increasingly verifying identity at the telecom layer, before a fraudulent transaction ever reaches an app.
The report leans on a customer example to make the point: UK fashion retailer NEXT uses Infobip’s AI- and machine-learning-based Signals product to spot and respond to abnormal spikes in messaging traffic, a common early warning of an attack in progress.
“Fraudsters are using AI to automate and scale their campaigns faster than ever, but AI-based protection is evolving at the same pace. Leading organizations no longer treat security as an afterthought — they are building it directly into their communications infrastructure.”
Matija Razem, Chief Commercial Telecom Officer, Infobip
A dose of skepticism is warranted, as with any vendor-authored trend report: Infobip sells the very detection tools whose adoption it is measuring, so rising “adoption” is also a rising revenue opportunity. Still, the directional signal is hard to argue with. As AI lowers the cost of running convincing scams at scale, the only realistic counter is defense that operates at the same speed — and the same layer — as the attack.
