A lender expanding into fix-and-flip and rental-property bridge loans in the US, and into real estate-backed and asset-backed SME lending in Nigeria, needed to move from manual, opaque processes to digital application, AI-assisted underwriting, and transparent pricing. Traditional hard money underwriting was taking days or weeks, with fees and timelines that were unclear until late in the process. In Nigeria, real estate-backed SME finance was largely manual with 4–8 week timelines. The client wanted to be the first technology-enabled hard money offering in the target Nigerian segment and to compete in the US on speed and transparency: 7–10 day close for fix-and-flip, 3–7 days in Nigeria, and 24–48 hour preliminary decisions so borrowers and brokers knew where they stood.
We were brought in to extend the client’s existing AI decision engine (Vida) for property-based underwriting and to design the digital experience and repayment infrastructure. The scope included LTV, ARV, LTC, exit strategy, collateral risk scoring, integration with valuation providers (e.g. HouseCanary, ATTOM in the US; Propvat or commissioned valuers in Nigeria), a transparent rate calculator on the website, a digital document portal with OCR, and API-driven collection accounts so repayments and servicing were automated and auditable. Regulatory structure was also in scope: the client would launch via a licensed partner in the US and an MFB partner in Nigeria, with technology and servicing retained in-house.
The challenge
Traditional hard money underwriting relied on manual document collection, fax or email shuffling, and opaque fee structures. Borrowers and brokers often did not see total cost or timeline until deep in the process. That created friction and drop-off. In the US, fix-and-flip and bridge borrowers need certainty: they are often under contract with a short closing window. Delays or surprises at the last mile are unacceptable. In Nigeria, SME owners with real estate collateral had few options besides banks (slow, document-heavy) or informal lenders (expensive, opaque). The client wanted to differentiate on speed, transparency, and a digital-first experience while meeting regulatory requirements in both markets.
A second challenge was underwriting logic. The existing Vida engine was built for unsecured or salary-backed lending. Property-based lending required new inputs: property value (AVM or appraisal), after-repair value for fix-and-flip, loan-to-value and loan-to-cost limits, exit strategy (sell vs refinance), and collateral risk grading. We had to extend the engine without breaking existing use cases, and we had to integrate with third-party valuation APIs and internal policy rules so that decisions were consistent and auditable. Personal guarantees and layered repayment security (e.g. ACH, assignment of rents, UCC, foreclosure) had to be designed so that investors and partners could see clear separation between collateral, guarantees, and recovery steps.
A third challenge was collections and investor reporting. The client wanted API-driven collection accounts (e.g. Stripe Treasury in the US, Flutterwave DVA in Nigeria) so that borrower payments would flow into accounts held in the name of the lending entity or SPV, not commingled with operating funds. That structure is required for investor confidence and for clean accounting. We designed the flow so that each loan could have a dedicated collection account, with auto-sweep to the main SPV or partner account and clear reporting for investors and auditors.
Our approach
We extended the AI decision engine for property-based underwriting. The engine now ingests property and project data, calls AVM or valuation APIs where configured, and applies policy rules (LTV, ARV, LTC, minimum credit score, minimum borrower experience). It produces a preliminary decision (approve, decline, or refer) and a risk grade. That output is available within 24–48 hours of a complete application, so brokers and borrowers get fast feedback. The final decision may still require a human review (e.g. for large loans or exceptions), but the majority of applications get a clear preliminary result that sets expectations.
We built a transparent rate calculator on the website. Applicants can enter loan amount, term, and property type and see an all-in cost estimate (rate, points, fees) before they apply. That reduces surprises and builds trust. The calculator is backed by the same pricing logic used in production, so the estimate is consistent with what they will see at closing. We also built a digital document portal where applicants upload documents (e.g. ID, bank statements, purchase agreement). OCR and auto-extraction pull key fields into the application so that staff spend less time rekeying and more time on exceptions. That cut manual collection and rework and shortened time to decision.
For repayment, we defined a layered stack: pre-authorized ACH (US) or NIBSS e-mandate (Nigeria), then assignment of rents where applicable, then UCC or lien enforcement, then personal guarantee, then collateral liquidation. Collection accounts are created via API (Stripe Treasury, Flutterwave DVA) in the name of the appropriate entity. No joint accounts; funds flow from borrower to collection account to SPV or partner account, with servicing and management fees paid to the client’s operating account separately. That separation is documented and auditable, which was required for the US partner and for investor reporting.
Results and metrics
In the US, the target is 7–10 day close for fix-and-flip (vs 14–21 typical) and 24–48 hour preliminary decisions. Borrowers and brokers get real-time status updates so they know where the application stands. In Nigeria, the target is 3–7 day close for real estate-backed SME loans; the client is the first digital, AI-underwritten hard money-style product in the segment. Transparent rate sheet and all-in cost before application are now standard. Investors get clear reporting and separation of SPV accounts from operating accounts, which supports future syndication or fund structures.
Key metrics: 7–10 day target close (US fix-and-flip); 24–48 hour underwriting (preliminary decision); 3–7 day target close (Nigeria real estate-backed SME); transparent rate sheet and all-in cost before application. The client can now scale originations in both markets without proportionally scaling manual underwriting or document handling. The portal and status updates removed the black hole that borrowers and brokers were used to with other lenders.
What this means for you
If you are a lender or a platform looking to offer hard money, bridge, or asset-backed products with digital application and AI-assisted underwriting, this case study shows that speed and transparency are achievable. The build required domain knowledge in real estate finance, valuation integration, and regulatory structure (licensed partners, SPV accounts). We brought that so the client could focus on distribution and capital, not building the stack from scratch.
We could see the total cost and timeline before we applied. The portal and status updates removed the black hole we were used to with other lenders.Borrower / broker
Lessons for similar projects
First, transparency is a differentiator. The rate calculator and upfront cost reduced friction and built trust before the applicant committed. Second, document automation (OCR, auto-extraction) cuts cycle time and errors; it is worth investing in early. Third, collection account structure and investor reporting need to be designed from day one if you plan to bring in partners or investors; retrofitting is painful. Fourth, regulatory structure (licensed partner in the US, MFB in Nigeria) allowed the client to launch without waiting for their own license while keeping technology and servicing in-house.
If you are exploring technology-enabled lending in real estate or asset-backed finance, we can walk through your product, your markets, and your regulatory setup in a discovery call. We will tell you whether our approach fits and what a path to production could look like.