The Invisible Shield: How Fluz Protects Billions in Transactions Without Anyone Noticing

At Fluz, we operate one of the largest incentives and rewards solutions, with nearly $10 billion in transactions since 2018. At that scale, a reactive posture was never an option. Attacks on reward and incentive platforms have become more frequent, more sophisticated, and harder to detect. The instruments involved, whether digital gift cards, prepaid cards, or disbursements, share one common vulnerability: they move fast and they move value.

So we built something designed around that reality: our Invisible Shield.

What the Invisible Shield Is

The Invisible Shield is a multi-layer fraud prevention architecture with one governing objective: create a completely frictionless experience for legitimate users while building an impenetrable wall behind the scenes.

It operates across three integrated pillars, each addressing a distinct stage of risk:

Know Your Business (KYB): who are we letting into the platform?
Know Your Customer (KYC): who is actually transacting?
Fraud Monitoring: is this specific transaction legitimate, right now?

These aren't independent checkboxes. Each layer feeds the next, and all three run in real time.

Layer 1: Know Your Business

Before a single transaction is ever processed, Fluz needs to know who it's doing business with.

New partners complete a fully embedded KYB flow without ever leaving the platform, with no redirects or external portals. The process validates business identity, ownership structure, and control upfront, supporting AML requirements and keeping bad actors out from the start.

The stack is deliberately layered across best-in-class vendors, each handling a distinct slice of the problem:

MidDesk verifies the business entity and monitors it over time
Vouched confirms the identities of the people behind the business
Inscribe.ai detects forged or altered documents
Greenlight.ai handles sanctions and OFAC screening for higher-risk cases
Fingerprint.js flags suspicious device behavior during onboarding

No single check determines the outcome. Every signal feeds into a unified internal trust model, producing a combined view of business legitimacy, identity integrity, document authenticity, sanctions risk, and device behavior.

Layer 2: Know Your Customer

Once a business is verified, individual users transacting within the platform go through the same embedded, real-time approach.

Known details are pre-filled where available. Sensitive fields are handled securely within the experience. But the critical difference from standard KYC is timing.

Most KYC flows are submit-and-wait: the user fills out a form, waits days for review, and gets pulled back in when something's missing. Ours runs verification in real time as data is entered. If anything additional is required, it surfaces immediately so the user can complete it on the spot, in one session, without ever being sent away and brought back.

The result is a process that's thorough without being punishing.

Layer 3: Fraud Monitoring

This is where the architecture gets serious.

Every transaction, before it completes, passes through a multi-layer stack combining Riskified, Oscilar, Idology, Plaid, and Prove with internal trust models continuously refined by a dedicated data science team. The system evaluates transaction velocity, purchasing patterns, device characteristics, account history, and behavioral anomalies, all simultaneously, at the moment of purchase.

The AI-driven decisioning layer, powered by Oscilar, is trained specifically to detect account takeover attempts, automated purchasing attacks, and abnormal spend behavior. The goal is not simply to flag suspicious activity and queue for review. The system is designed to intervene before a risky transaction completes.

Because value moves the instant a transaction is approved, post-transaction intervention isn't an option. The entire system is architected around that constraint: every purchase is treated as a live risk event that must be cleared in real time.

Supporting the technology is a compliance, risk, and fraud team that represents nearly 15% of the Fluz organization, working in direct conjunction with data science to keep detection models sharp as fraud patterns evolve.

The Results

The Invisible Shield operates across hundreds of thousands of users and merchants and billions of dollars in transactions. The numbers reflect what's possible when you build for the problem specifically.

Fraud Rate: Fluz's year-to-date fraud rate for 2026 stands at 0.05%, half the industry's best-in-class benchmark of 0.1% per the Federal Reserve Payments Fraud Survey.

Approval Rate: Fluz's current approval rate of 94% sits squarely within the ideal industry benchmark range of 90% to 98%+, meaning stronger security isn't coming at the cost of legitimate conversions.

Blocked Fraud: Over the past three years, more than 11,000 fraudsters have been blocked, representing nearly $2.8 million in prevented losses.

Why This Matters

As incentive and reward platforms scale, fraud prevention stops being a back-office function and becomes a core growth requirement. Fraud erodes trust, kills engagement, and can quietly destroy program economics before the damage is visible.

Generic fraud tooling wasn't built for this category. The Invisible Shield was. Three layers, one unified trust model, zero friction for the people who belong in the system.

That's what purpose-built infrastructure looks like.