Case Study: How Uber Identifies the Real Owners Behind Fleet Partners and Freight Operators
Last updated: October 22, 2026
Fleet fraud, shell intermediaries, and double-brokering. Uber needed to see who was actually behind the companies operating on its platform — not just the name on the application.
The Challenge
Uber’s business runs through intermediaries. In ride-hailing, fleet partners recruit and manage drivers on Uber’s behalf. In Uber Freight, carriers and brokers move billions of dollars in cargo. Both models create the same structural problem: Uber is one, two, or three steps removed from the people actually operating under its brand.
That gap has consequences, and they are documented, not hypothetical. In Poland and across Eastern Europe, investigations into Uber’s “partner” model found hundreds of intermediary companies operating in a legal gray area — firms that handled payroll, contracts, and driver relationships while the platform itself stayed a step removed from the employment questions those arrangements raised. In Ukraine, the Bureau of Economic Security opened a case against the management of Uber’s local partner company, alleging unpaid taxes of roughly UAH 40 million: fares collected from drivers were routed to the platform, and the commission-adjusted balance credited back to the partner company’s own accounts, obscuring where the money actually settled. Neither case required a sophisticated shell structure — just a company that was, on paper, hard to see through.
Uber Freight has the mirror-image problem at higher dollar values. Double-brokering fraud — where a fraudulent carrier intercepts a load tendered to a legitimate one, then disappears with the cargo or the payment — has become a defining risk in US freight. Industry fraud-tracking networks recorded a record-high quarter for freight fraud in early 2026, with roughly half of theft incidents traced to carriers that held legitimate operating authority and a clean prior history before being compromised. Change-of-ownership fraud, where a dormant or shell carrier entity changes hands and is immediately used to intercept loads, rose sharply year over year during the same period. Cargo theft losses across the US and Canada were estimated at roughly $725 million in 2025, up about 60% from the year before.
The root cause is the same in both divisions: Uber could not reliably see who actually owned and controlled the companies operating on its platform, at the moment it mattered and again every month after. Self-declared partner applications were easy to fake. Commercial databases provided surface-level company records but no ownership depth. And manual research across 70+ countries, each with its own registry format and disclosure regime, did not scale to a marketplace onboarding thousands of counterparties a year.
Why Marketplace KYB Is a Different Problem Than Single-Counterparty KYB
Most Know Your Business programs are built for a bank or a large enterprise screening a manageable number of counterparties: suppliers, borrowers, institutional clients. Uber’s problem is structurally different. It is not verifying one counterparty at a time under a relationship manager’s attention — it is running due diligence as a gate at the front of a two-sided marketplace, where thousands of fleet partners and freight carriers apply, get onboarded, and keep operating with minimal ongoing human review unless something trips a flag.
That difference changes what the due diligence program actually has to do well.
| Dimension | Single-Counterparty KYB | Platform-Scale KYB |
|---|---|---|
| Volume and cadence | Dozens to hundreds of entities, reviewed by analysts over days or weeks | Thousands of applications, decided in minutes or the applicant churns to a competitor |
| Network risk | Each counterparty assessed largely on its own | The same nominee director or registered address can recur across hundreds of unrelated applicants — a pattern only visible at portfolio scale |
| Risk after onboarding | Re-reviewed on a fixed periodic schedule | Ownership, directors, and registry status can change the week after onboarding, while the partner keeps transacting |
| Jurisdiction spread | Often concentrated in a handful of markets the institution already operates in | Partners and carriers register across dozens of countries with incompatible disclosure regimes, simultaneously |
| Who bears the consequence | Primarily the institution itself | Reputational and regulatory exposure rolls up to the platform brand, even though the partner is a separate legal entity |
Under the Financial Action Task Force’s Recommendation 17, a firm that relies on another party’s customer due diligence still has to satisfy itself that the relied-upon checks were adequate, and still carries ultimate responsibility for the relationship. A marketplace that treats partner vetting as a one-time form submission is, in effect, relying on the partner’s own self-declaration with no independent verification behind it — the weakest version of reliance, not a documented, risk-based one.
The Solution
Uber deployed Zavia.ai to answer the ownership question for every fleet partner, freight carrier, and corporate counterparty on its platform — at onboarding, and continuously afterward.
How Zavia.ai works differently for Uber:
- Fleet partner legitimacy scoring: Before onboarding a new fleet partner, Zavia.ai pulls the entity’s registration status, directors, and full ownership chain from the local government registry, then resolves the chain to the natural persons who ultimately own or control it — not just the name on the application. Partners that are recently formed, have no directors on record, or are registered at known corporate-service-provider addresses are flagged immediately. This is the same red-flag logic covered in Shell Company Red Flags: A Due Diligence Checklist Before You Onboard, applied at the scale of a marketplace intake queue rather than a single counterparty file.
- Freight carrier identity verification: For Uber Freight, Zavia.ai verifies that a carrier’s corporate identity matches its registry filing, and that the beneficial owners behind it are who the carrier claims. If a carrier presents itself as an established fleet but the registry shows a dormant company with no recent filings and a change of ownership in the prior weeks — the exact pattern behind the change-of-ownership fraud spike described above — Uber knows before the load is tendered rather than after it disappears.
- Beneficial-ownership screening tied to sanctions and PEP checks: Identifying the entity is only half the job. Each natural person resolved as a beneficial owner is screened against sanctions lists, PEP status, and adverse media, the same four-component verification stage described in UBO Verification Workflow: From Initial Check to Ongoing Monitoring. A fleet partner with a clean corporate filing but a sanctioned beneficial owner behind it is exactly the gap that entity-only screening misses — see Sanctions Screening and Beneficial Ownership for why a clean entity check alone does not catch a sanctioned owner.
- Intermediary chain mapping: Uber’s partner model creates intermediary chains — a fleet company that subcontracts to another fleet company that hires the drivers. Zavia.ai maps these chains by tracing ownership through each entity, identifying where a single individual or group controls multiple intermediary companies across markets. For the fragmented, multi-jurisdiction version of this problem, the same scoring logic used to flag when an ownership chain has crossed from ordinary structuring into enhanced-due-diligence territory is covered in Multi-Layered Corporate Structures: A Risk-Scoring Framework for Escalation Decisions.
- Real-time partner portfolio monitoring: When an active partner’s ownership changes, its registration lapses, or it gets struck off a registry, Zavia.ai sends an alert. Uber does not wait for a driver complaint, a missed load, or a fraud investigation to discover that a fleet partner has changed hands or gone dark.
What a single onboarding check actually runs through, stage by stage, looks like this — shown here as an illustrative walkthrough, not a specific real applicant:
The six-stage flow above describes how a check of this kind runs in general, to illustrate the mechanism. It is not a disclosure of a specific real applicant, decision, or outcome on Uber’s platform.
Cross-Border Complexity at 70+ Country Scale
A fleet partner onboarded in Poland, a freight carrier registered in Delaware, and a corporate counterparty incorporated in the UAE sit behind three completely different disclosure regimes — a public PSC-style register, a state-level filing with no beneficial-ownership disclosure at all, and a free-zone authority with its own rules. Running consistent due diligence across that spread is not a matter of applying one checklist everywhere; it is a matter of knowing, jurisdiction by jurisdiction, what data actually exists to check against, and treating the absence of a public registry as a risk signal in itself rather than a dead end.
This is also where the reliance question from FATF Recommendation 17 becomes concrete rather than theoretical. A marketplace cannot run its own boots-on-the-ground registry search in 70-plus countries for every applicant. What it can do is connect directly to the registries that are queryable, document what was checked and what was not, and apply enhanced scrutiny by default wherever the local regime does not support the same depth of check — rather than silently treating a jurisdiction with no public UBO register as equivalent to one that has full disclosure.
The Results
| Metric | Before Zavia.ai | After Zavia.ai |
| Fleet partner vetting | Self-declared applications plus basic database checks | Registry-verified ownership with AI-powered red-flag detection |
| Freight carrier identity | Carrier profiles accepted at face value | Cross-referenced against government registry filings before load tender |
| Beneficial-owner screening | Entity-level checks only, no sanctions/PEP layer tied to ownership | Every resolved UBO screened against sanctions, PEP, and adverse media |
| Intermediary chain visibility | None — Uber saw one layer deep | Full chain mapped to the natural person behind each entity |
| Fraud detection speed | Reactive — discovered after incidents | Proactive — flagged at onboarding or on ownership change |
Why It Matters
Uber’s regulatory risk scales with its partner network. Every fleet partner that turns out to be a shell company, every freight carrier that double-brokers a load, every intermediary that evades employment or tax law — these are liabilities that roll uphill to Uber, even though each one is, on paper, an independent company. The Ukrainian tax case and the Polish partner-company investigations are concrete examples of exactly that dynamic: the platform did not sign the contracts in question, but the exposure reached it anyway. Regulators in multiple markets are scrutinizing platform intermediary models more closely than they were even two years ago, and FMCSA’s own broker-and-freight-forwarder financial-responsibility rule, which took full effect in January 2026, is part of the same regulatory trend in freight specifically — tightening who is allowed to operate as an intermediary at all.
The company that cannot show it verified its partners’ ownership, on a documented and repeatable basis, is the company that pays the fine, absorbs the reputational hit, or both. For a related real-world example of ownership verification applied under a different compliance driver — Germany’s supply chain due diligence law rather than a marketplace fraud problem — see Case Study: How SAP Embeds UBO Discovery Into Supply Chain Due Diligence for LkSG Compliance.
Bottom line:
Zavia.ai gives Uber the ability to see through its intermediary model — verifying fleet partner legitimacy, catching fraudulent freight carriers, and mapping the ownership chains behind the companies operating under Uber’s brand across 70+ countries.
Frequently Asked Questions
Why is a marketplace like Uber exposed to its partners’ ownership, if the partner is a separate legal entity?
Legal separation does not eliminate reputational or regulatory exposure. Regulators, journalists, and drivers or shippers hold the platform brand responsible for what happens under it, regardless of the corporate structure behind a given fleet partner or carrier. The Ukrainian and Polish cases referenced above are examples of exactly this: the platform did not sign the disputed contracts, but the scrutiny reached it anyway.
What is double-brokering, and why is it a beneficial-ownership problem rather than just a fraud problem?
Double-brokering is when a fraudulent carrier intercepts a load that was tendered to a legitimate one, then diverts the cargo or payment. It is a beneficial-ownership problem because a large share of these incidents involve carrier entities with valid operating authority and a clean prior history that were recently acquired or repurposed — an identity-and-ownership check at the moment a load is tendered, not just a one-time authority check, is what catches a carrier that changed hands days earlier.
How is platform-scale KYB different from the KYB a bank runs on a single corporate client?
A bank typically reviews a manageable number of counterparties with analyst time behind each one, re-checked on a fixed schedule. A marketplace onboards thousands of applicants that need a decision in minutes, has to watch for the same nominee or address recurring across many unrelated applicants, and needs ownership changes to trigger re-verification automatically rather than waiting for the next periodic review.
What does FATF Recommendation 17 have to do with a ride-hailing or freight marketplace?
Recommendation 17 governs when a firm may rely on due diligence performed by another party. It requires the relying firm to satisfy itself the underlying checks were adequate and keeps ultimate responsibility with the relying firm. A marketplace that accepts a partner’s self-declared ownership with no independent verification is, in effect, relying on an unverified source — the weakest and least defensible form of reliance under that framework.
Why does a recently formed company with no directors on record get flagged automatically?
A company that is only weeks old, has no directors listed in the registry, and is registered at a known corporate-service-provider address is consistent with a shell entity set up specifically to pass an onboarding check rather than operate as a going concern. None of those three facts alone proves fraud, but together they are a well-established red-flag pattern, covered in detail in Zavia’s shell-company due diligence checklist, and warrant manual review before onboarding rather than automatic approval.
Does verifying the entity’s registration status also verify its beneficial owners?
No. Entity verification confirms the company legally exists and is in good standing; it says nothing about who controls it. A fully legitimate, active, properly filed company can still have a sanctioned or high-risk individual behind it. Catching that requires resolving the ownership chain to natural persons and screening each one separately, which is why entity checks and beneficial-ownership screening are run as distinct, complementary steps rather than treated as the same check.
Why does onboarding verification need to continue after a partner is approved?
Ownership, registration status, and risk exposure are not fixed at the moment of approval. A partner can change hands, lapse its registration, or have an owner newly added to a sanctions list months into an active relationship. Continuous, trigger-based monitoring closes that gap; a one-time check at onboarding only answers the question as of that day.
For the broader beneficial-ownership mechanics referenced throughout this case study — thresholds, control tests, and the direct-versus-indirect ownership math — see What Is a UBO? Ultimate Beneficial Owner Guide for 2026.
