Our next story in the series on how real customers use the Zumigo platform to solve real problems: a bank that doesn’t call our signals selectively at all — it calls a full, coordinated set of them, every time, in real time, and has scaled that pattern into one of the largest signal programs on our platform.
Our customer in this story is a large U.S. bank offering retail and digital banking to millions of account holders. A bank’s exposure is spread across the whole account relationship — logins, account changes, money movement, card activity — any of which can be the moment an account takeover (ATO) attempt shows up. No single check covers all of that surface, and this customer’s answer is to not rely on one.
In short: this customer calls a coordinated, multi-pronged stack every time, live against the network, rather than leaning on any single check or a stored record — account status, call handling, IMEI change, and SIM change, fired together in the same transaction. Each signal is watching a different way an attacker could get in; together, in real time, they cover the whole front. And the protection that pattern provides has meant the customer has dramatically increased its usage year over year.
Most integrations we see start narrow — a single signal to answer a single question — and grow into a stack as the customer discovers what combining signals catches that no individual one does. This customer’s pattern looks different: practically all of its volume already calls the same combination, every time, rather than escalating to it case by case.
That’s a meaningfully different design choice. Rather than asking “which single check answers this specific question,” the system always asks the fuller question: has anything about this phone, this SIM, or the device it’s paired with changed recently, in combination with what the account itself looks like right now? A SIM swap alone might be explainable. A SIM swap and an IMEI change and an account-info anomaly, all in the same transaction, is a very different signal — and monthly volume on this pattern has kept climbing, with adoption spreading across more of the customer’s products as the signals prove their value.
Fraud doesn’t attack a phone number one way — it attacks the relationship between a person, their SIM, their device, and their account, and each of those is a separate door an attacker can try:
No individual signal here is new. What’s distinctive is that this customer treats them as one question asked four ways at once. A fraud attempt that looks clean on any single signal still has three other doors to get past in the same instant, and it has to clear all of them simultaneously, in real time — not at separate points in a review queue.
These aren’t static identity records pulled from a database and cached — they’re live reads of the state of a phone and its carrier relationship at the moment of the check. A stolen or SIM-swapped phone doesn’t look wrong in a static database; it looks exactly like the legitimate customer’s phone, because the database doesn’t know anything happened. It only looks wrong the moment you check the live state of the network — which is precisely when this customer checks it: not at onboarding, not on a schedule, but at the instant of the transaction itself, every time.
Real-time answers “is this true right now”; the four-signal stack answers “right now, according to every angle that matters.” Either alone leaves a gap. This customer runs both at once, which is why the combination is the story. And, Zumigo’s real-time mobile intelligence with live carrier integration provides the answers each time.
Fraud evolves, and defenses have to keep pace with it. The fastest-growing accounts we see tend to stop asking “which one check do we need here” and start asking “why would we ever leave a door unchecked.” A single signal is a data point about one thing; four signals, checked live against the network on every transaction, is closer to a perimeter. And, to be sure, there are many more real-time signals that Zumigo can provide to prevent identity fraud across the customer journey.
The case for multi-pronged, real-time coverage isn’t “trust us, it works” — it’s structural: a fraud attempt has to beat all four checks, all live, all at once, instead of just the one check a narrower integration would have run. As fraud itself gets faster and increasingly AI-driven, real-time contextual signals keep your defenses from falling a step behind — and it’s worth a conversation if yours isn’t there yet.
Madhu Vudali is VP, Product Management at Zumigo. Comments or questions? Connect on LinkedIn: @madhuvudali