Explain Eeshita's thinking on construction lending to an unfamiliar outsider. Refrain from using private information (e.g. individual names etc.)
Eeshita’s construction-lending thesis is that this is a complex, still-manual financing niche where software can create outsized value by turning fragmented builder, project, and lender information into an underwriter-ready credit package. Rather than competing only on rate, she wants to make funding faster and more certain by mapping each lender’s real credit box—including qualitative decision gates—and telling builders what must change for a deal to qualify.
- She sees ground-up construction, teardowns, and infill as attractive because their documentation and underwriting complexity deter many lenders and technology providers; that complexity can justify higher margins rather than being treated as a commodity workflow.
- Her intended product is not merely document extraction: it would assemble a credit memo/package for the lender’s underwriter, automate much of early origination, and retain builder-performance, project-execution, construction-cost, property, and eventually liquidity data. This data could improve underwriting throughput and potentially support lenders’ securitization of loans.
- Her central defensibility view is to model each lender’s complete decision process—quantitative rules plus qualitative gates—not just publish prices. A builder should receive concrete remediation guidance, such as changing loan size, verifying liquidity, or obtaining a permit, before submitting a file.
- She believes speed and execution certainty matter materially to builders. Price and covenants are considerations, but a simple rate-comparison marketplace would become a race to the bottom; the more durable value is matching a deal to a lender that can actually close it quickly.
- Her approach is grounded in the operational reality that construction lending requires detailed, credible budgets and project evidence. Budgets are scrutinized for front-loading, missing or nonspecific scope items, and consistency with the appraisal; draws generally reimburse completed installed work and involve inspections, lien releases, and delinquency controls.
- She appears to favor starting with a narrow, repeatable loan category—such as smaller ground-up construction or fix-and-flip—then extending the automation to more nuanced loans. Her own early estimate was that 80–90% of a deal’s early stage could be automated, while human judgment remains important for context-sensitive review.
- She is pursuing an asset-light software and marketplace model rather than owning lenders, because she believes software can command better economics and is more resilient than balance-sheet lending during credit downturns.
- Her near-term data strategy is to codify how lenders assess the builder first, then incorporate property details and liquidity. She is seeking real loan files and builder document repositories under confidentiality to build this underwriting model from actual workflows.
- She is cautious about relying long-term on an external property/pro-forma data provider: she would use such a product only as a temporary bridge while building core capabilities in house, particularly if the provider’s maintenance commitment is uncertain.