Edmonton-based Trust Science has acquired Toronto-based Lenders API to create what they are calling "Canada's first industry-wide defence system against coordinated lending fraud." The combined platform targets three specific fraud schemes that alternative business lenders know intimately: loan stacking, bust-out fraud, and synthetic identity fraud. Financial terms were not disclosed.1
The acquisition matters less for what it is today (a Canadian deal between a 12-year-old AI risk company and a 4-person startup) and more for what it signals: the lending industry is finally building infrastructure to address coordinated fraud across institutional boundaries. For MCA funders, factoring companies, and equipment finance lenders who lose millions annually to borrowers playing multiple lenders against each other, this is the most operationally relevant development in fraud prevention this year.
What Trust Science and Lenders API Actually Built
Lenders API, founded in 2023 by Tal Schwartz (CEO) and his father Gary Schwartz, built a consortium model that solves the coordination problem at the center of lending fraud. Here is how it works: participating lenders share anonymized credit and fraud data in real time. As Schwartz explained, "Borrower data comes in; we anonymize it at source; we notify other lenders anonymously."1
The anonymization is the critical design choice. Competing lenders will not share proprietary portfolio data with each other. They will, however, share fraud signals if they cannot be traced back to specific borrowers or deal terms. Lenders API anonymizes data before it leaves the originating lender's system, creating a layer of fraud intelligence that no single lender could build alone.
Trust Science brings AI-powered risk and decisioning technology to the combination. The integration layers consortium-based fraud intelligence on top of AI-driven credit decisioning, giving lenders both the signals (is this borrower simultaneously active elsewhere?) and the analysis (what does that pattern mean for default probability?).1 2
The Three Fraud Types This Addresses
Loan Stacking
A merchant takes cash advances from three to five funders simultaneously against the same revenue stream, knowing the combined daily payments are unsustainable. Each funder sees only their own deal. By the time the first default hits, the merchant has collected $200K to $500K across multiple advances and the recovery is pennies on the dollar.
Loan stacking is the single largest fraud vector in MCA and revenue-based financing. The industry has known this for years. The problem is structural: funders compete with each other, so they have no mechanism to see what a borrower is doing at other shops. Services like DataMerch and informal industry blacklists address the problem partially, but they are reactive (flagging after defaults) rather than proactive (flagging at application).3
A consortium that flags simultaneous applications in real time, before funding, would fundamentally change the economics of stacking fraud. The math is straightforward: $500 to $1,000 per questionable deal verification versus $50,000 or more in losses per fraudulent stacked advance.1
For a deeper look at how loan stacking exploits the "first position" assumption in MCA, see our analysis: The "First Position" Myth: How Loan Stacking Happens.
Bust-Out Fraud
An established business builds credit history with one or more lenders over 12 to 18 months, making payments reliably and building trust. Then it secures loans, equipment leases, or advances from multiple institutions simultaneously and defaults on all of them at once. The credit history was the setup. The simultaneous drawdown was the play.
Bust-out fraud is harder to detect than stacking because the borrower has a legitimate operating history. Individual lenders see a good customer suddenly going bad. A consortium sees the pattern: the same entity drawing down across five institutions in the same 30-day window. That pattern is invisible to any single lender but obvious when the data is pooled.
The First Brands fraud, one of the largest factoring frauds in history at $2.3 billion, demonstrated how sophisticated lenders (UBS, Jefferies, Raistone) can fail to detect coordinated fraud when each one trusts their own due diligence without cross-referencing. When multiple lenders independently verify the same borrower but never check what the borrower is doing elsewhere, the verification is incomplete by design.4
Read the full breakdown: First Brands Alleged Multibillion-Dollar Fraud.
Synthetic Identity Fraud
Fabricated identities combining real and fake data (a real Social Security number with a fake name and address) are used to open accounts and secure financing. FinCEN has identified synthetic identity fraud as the fastest-growing fraud type in financial services. The challenge: each individual data point may pass verification. The identity itself is the fabrication.
A consortium approach helps because synthetic identities often hit multiple lenders simultaneously (the fraud economics require it). Seeing the same data pattern applying across institutions, even when the names differ slightly, creates a detection signal that single-lender systems miss.
Why This Matters for US Alternative Lenders
The immediate limitation is geographic: Trust Science and Lenders API operate in Canada. No US launch has been confirmed, and the regulatory environment differs (US data privacy laws, state-level licensing, antitrust considerations for cross-lender data sharing). Scaling from Canada's fintech market to the US, which is roughly 10 times the volume in alternative business lending, is a significant operational challenge.1
The strategic signal is more important than the current product availability. The consortium model addresses a gap that no US tool currently fills adequately. DataMerch provides post-default blacklists. LexisNexis and Experian offer some cross-lender visibility through credit bureau data. But nothing in the US market provides real-time, anonymized, pre-funding fraud intelligence across competing lenders.
If Trust Science successfully enters the US market (or if a US competitor builds a similar system), early adopters gain an information advantage that compounds over time. The more lenders participate, the more complete the fraud picture becomes, and the harder it is for stacking and bust-out schemes to succeed.
The cost argument is immediate. Industry estimates place stacking losses at $50,000 to $100,000 per fraudulent merchant across multiple funders. A verification cost of $500 to $1,000 per questionable deal represents an ROI that any CFO approves on sight. The question is not whether consortium fraud detection is worth it. The question is when a US solution reaches sufficient scale to be effective.
Why Would You Share Data with Competitors?
The strongest objection to consortium models comes from the operations desk: why would I give competitors visibility into my deal flow? Other industries solved this exact problem decades ago.
In insurance, the CLUE database (Comprehensive Loss Underwriting Exchange), operated through LexisNexis, allows insurers to share claims data across carriers without exposing pricing or underwriting criteria. Every major carrier participates because the cost of insuring a claimant with undisclosed prior losses exceeds the competitive value of keeping that information private.
In banking, ChexSystems and Early Warning Services allow banks to share account abuse data. Zelle, used by over 2,000 financial institutions, was built on this consortium infrastructure. Banks that compete aggressively for deposits still share fraud signals because absorbing preventable losses is more expensive than the competitive intelligence they think they are protecting.
Credit bureaus are the original consortium model. Lenders share repayment data with Experian, TransUnion, and Equifax because the information asymmetry of not sharing costs more than the competitive risk of sharing.
The question for MCA funders is not whether you trust your competitors. It is whether the $50,000 to $100,000 per stacking loss is more expensive than the competitive intelligence you think you are protecting. For most high-volume shops, the math is not close.
Three Actions for Monday Morning
- Quantify your stacking losses. Pull every default from the past 12 months where the merchant was found to have simultaneous advances from other funders. Most MCA operations know stacking is a problem but have never calculated the dollar impact. You cannot budget for prevention without a loss number.
- Evaluate existing consortium and fraud-sharing options. DataMerch, ClearSale, and industry-specific blacklists each address part of the problem. Map what coverage you have today, identify the gaps (especially pre-funding, real-time signals), and set a target for where you need to be.
- Layer entity verification as the first check in your fraud prevention stack. Consortium data tells you what a borrower is doing at other institutions. Entity verification tells you whether the borrower is real. A shell company with a pristine consortium record is still a shell company. Confirm the entity is active, verify its formation date and registered agent, and flag any discrepancies before moving to credit analysis or consortium queries.
.jpg){kind=small}
.png)