Orbii

Strip out the models and Orbii is back to the slow manual credit checks it exists to replace. Machine learning models process raw financial data instantly to produce lending decisions in seconds. A separate layer analyzes and enhances that raw data into a comprehensive view of borrower financial health, and the same models drive underwriting, disbursement and ongoing monitoring. Assessing a business from point of sale and resource planning activity rather than from a credit file is inference throughout.

Automation extends past the decision to the money. The platform underwrites, disburses and monitors loans in real time with decisions returned in seconds, and the chief executive states the ambition directly, that credit decisioning will not be a process but a reflex, embedded in the systems businesses already use.

That is a coherent product vision and it leaves no described human checkpoint anywhere: nothing states what confidence triggers referral, whether a declined applicant reaches a person, or what limits apply to automated disbursement. For a lender the review layer presumably exists internally and none of it is published.

No accuracy, default rate, approval lift or validation result is published. The claims rest on descriptions such as extremely precise credit judgments and higher quality lending decisions. Transparent underwriting technology is named as a component of the approach and never explained, which matters because that is precisely the property a lender's own risk function and its regulator would examine. The company is two years old and its models have not yet observed a full credit cycle in the segment they serve.

Volume is quantified without being large, at thousands of applications processed and millions of dollars in approved loans, against a stated ambition of powering $1 billion of small business lending. The investor base is the stronger signal for a company founded in 2024: a $3.6 million seed led by one of the world's largest technology investors, with a regional fintech specialist, two Saudi focused strategic investors including a sovereign linked accelerator vehicle, and a further venture firm.

The lead investor's head of Middle East investments is quoted specifically on implementation speed and tangible client results. Customers are described by type as digital lenders, fintechs, payment companies and business ecosystems, and none is named.

Orbii publishes no data boundary statement. The platform serves digital lenders, fintechs and payment companies competing for the same small business borrowers within a concentrated regional market. Its models improve with exposure to more repayment outcomes across that shared book. Nothing says whether performance data from one lender's portfolio informs the models scoring applicants for another, or whether a client can decline to contribute. What happens to accumulated borrower profiles when a lender leaves is also unstated.

The data reach is deep. The platform collects borrower financial information automatically from banking sources through interfaces or statement uploads, and connects into point of sale and resource planning systems. It observes a business's transactions, inventory and operations rather than a submitted application. For small businesses that material frequently concerns the owners personally. Yet no data protection agreement, retention schedule, subprocessor list or deletion commitment is published, and both markets served have their own data protection expectations that go unaddressed.

No attestation, certification, trust center or enumerated framework is public. Banks and licensed lenders are named buyers, and their supplier assessment processes gate any system that touches borrower data or moves money. Publishing an assessed control set is therefore the practical prerequisite for the bank segment the company is targeting alongside its fintech base.

No supervisor, statute or instrument is named. Both markets the company operates in have active financial regulators with specific regimes for lending, open banking access to account data and outsourcing by regulated institutions, and the platform touches all three by collecting bank data, underwriting and disbursing on behalf of licensed lenders. Nothing published identifies any of it.

The access argument is structural and well matched to its market. The company says it was created to address the limitations of traditional credit systems and fragmented data environments, aiming to make business credit fairer as well as faster and to expand access for underserved businesses. The mechanism makes that credible.

In markets with thin credit bureau coverage, reading point of sale and resource planning activity assesses a business on its actual trading rather than on a credit file it may not have at all. In such markets the alternative to observed data is not a worse file but no file. The two sided outcome claim of higher approvals with lower defaults supports it. Evidence is what is missing: no outcome data, no fairness analysis, and transparent underwriting asserted with no mechanism described.

There is no published guarantee, indemnity or correction process. The lender can measure approvals and defaults against its own book over time, which is real if slow feedback. The small business has none and is decided upon in seconds. Nothing describes what a declined applicant is told, whether reasons are supplied, or how a business corrects misread point of sale or accounting data that determined the outcome. That matters more where the borrower has no credit file to fall back on.

Input categories are named clearly, covering banking sources reached through interfaces, digital statement uploads, point of sale networks and resource planning systems, so a buyer understands what kinds of evidence feed a decision. No individual provider, aggregator or model supplier is identified, which matters in these markets because account data access depends on specific open banking infrastructure whose coverage varies by country and by bank. No subprocessor list or hosting arrangement appears.

Integration is the strategy rather than a feature. The platform connects directly into the systems businesses already run, named by category as banks, fintech platforms, point of sale networks and enterprise resource planning software. Data comes in through interfaces, or through digital statement uploads where automated access is unavailable. That combination is what allows assessment of businesses whose formal financial records are thin. No named system, provider or aggregator is published, and there is no public developer documentation.

Where the platform is hosted is not published, and neither is any region selection, residency commitment or private deployment option. The omission carries weight because both markets served maintain expectations about local processing of financial data. The company's stated expansion plan is to deepen integrations with regional financial systems, which will bring those requirements forward.

Pricing is not published, nor is any packaging or basis of charge. The commercial argument is framed as avoided cost: clients can launch, test and manage lending products without building credit functions from scratch. The lead investor praises how quickly the solution can be implemented. Neither states what it costs, and nothing indicates whether charge falls per application, per approved loan or by volume.

Five buyer types are addressed, spanning banks, digital lenders, fintechs, payment companies and business ecosystems, and functional coverage runs the whole lending lifecycle from underwriting through disbursement to monitoring rather than scoring alone. Modularity means a client can take one part or the whole. Geography is currently Saudi Arabia and the United Arab Emirates with wider regional ambition, and the borrower segment is deliberately narrow at small and medium businesses.