Explainable Risk Scoring Targets the Shell Firms Behind AI Access
Trulioo has launched Business Risk Horizon, a business risk scoring product that traces company ownership through layered corporate structures, flags sanctioned owners or ties to restricted jurisdictions, and attaches a plain-English explanation to every score. The company says a leading artificial intelligence lab has already signed on as an early adopter, using the tool to decide which businesses can reach its models. The product, a form of explainable risk scoring, went live on 15 October 2026.
The timing is deliberate. Trulioo frames Horizon against the White House Accord on Super Intelligence, the voluntary pact signed on 29 September 2026 by six leading developers: Google, Anthropic, Meta, OpenAI, xAI and Nvidia. The accord commits its signatories to four layers of internal controls and external audits, including independent board-level oversight. For AI companies now expected to show how they govern access to frontier systems, the identity of the business behind a corporate account has shifted from a compliance detail into a decision they may have to defend.
What does Trulioo's explainable risk scoring actually do?
The core idea is that the riskiest customers look ordinary on paper. A company's registered address is real and its filings are in order, yet Horizon finds dozens of other entities registered at the same address, several with owners linked to businesses flagged before. Trulioo says the tool weighs that concentration in context: routine at a Delaware registered-agent address, where shared addresses are common, and far more unusual in Sweden. Horizon then sets out, in plain language, how each finding contributed to the score, and leaves the final decision to the enterprise's own policies.
That written rationale is the point of difference. Horizon combines authoritative global Know Your Business (KYB) data with automated research into a company's operations and relationships, so an analyst sees the reasoning behind a score alongside the score itself. On Trulioo's own figures, the wider platform covers 195 countries, verifies more than 14,000 identity documents and over 700 million business entities, and screens against more than 6,000 watchlists. Those metrics trace to the company's materials rather than to independent audit.
Why are AI labs vetting business customers now?
Because access to advanced models has become a risk surface that boards and external auditors now expect developers to police. The September accord routes findings from internal and external reviewers to an independent board committee, which raises the bar on documentation: a lab that grants a business access to its models may have to explain, later, how it judged that customer. A scoring tool that produces a defensible written rationale maps neatly onto that need. Trulioo says one leading lab is already using Horizon for precisely this purpose, though it has not named the firm and the claim rests on Trulioo's own announcement.
What does Horizon change for onboarding costs and fraud teams?
It moves the hard calls earlier. Horizon flags likely fraud and businesses outside an enterprise's risk appetite before a case reaches full manual review, which Trulioo says can cut unnecessary analyst work, shorten decision times, and reduce the long-tail losses that come from onboarding customers who should have been declined. Chief Product Officer Zac Cohen argues that ownership frequently runs through several entities across jurisdictions, with the link to a shell company or sanctioned owner sitting several steps back from the visible business, and that the value lies in surfacing those links with evidence attached.
Chief Executive Vicky Bindra frames the commercial logic from the other side: a higher-risk customer can still be a profitable one if the exposure is understood and the right controls are applied, which is the judgement Horizon is meant to support rather than replace. The same approach extends past AI, to payment providers and marketplaces running enhanced due diligence on prospective merchants and sellers, where telling a legitimate supplier from one concealing illegal sales of controlled substances carries direct liability.
Where does Horizon fit in Trulioo's strategy?
Horizon is the latest move in a repositioning Trulioo began at Money20/20 in October 2025, when it recast KYB as a continuous intelligence layer and previewed business risk and reputation scoring. In early 2026 it added credit decisioning, reporting a 102% year-on-year rise in US KYB transactions on its own numbers. Horizon folds those threads together and puts explainability at the centre, which is where the competitive pressure now sits: the global verification market is crowded with capable scoring engines, and the harder problem is giving enterprises a record they can stand behind when a regulator, an auditor or a board asks how a decision was reached. Trulioo says it plans to widen Horizon across payments, marketplaces and financial services over time.
Why This Matters to FinanceX Readers
Vetting is becoming a documentation problem as much as a detection one. For finance and risk professionals, the signal is that onboarding decisions increasingly have to be explainable after the fact, not just correct in the moment, and that pressure is now reaching the customers of AI firms as well as banks and payment companies. For investors, Horizon is a read on whether Trulioo's pivot from identity checks towards lifecycle risk intelligence is landing with enterprise buyers, and a named early adopter in the AI sector, if the claim holds, is a useful proof point in a category where differentiation is hard to show.
