The credit gap is an operations problem.

The size of the unmet demand is not in dispute. The IFC puts the MSME finance gap in emerging markets at roughly $5.7 trillion, about 19% of those economies’ combined GDP, up from $4.4 trillion in 2015. Women-owned MSMEs account for some $1.9 trillion of it. In India alone, an RBI-commissioned committee estimated the MSME credit gap at ₹20–25 trillion. Zoom out to individuals and the Global Findex counted 1.4 billion adults still unbanked in 2021. Even in the United States, the CFPB found 26 million adults are credit-invisible and another 19 million unscorable, roughly one in five.

The reflex is to read all of this as a pricing problem: if only we could score thin- or no-file borrowers, the gap would close. Better risk models help. But they do not explain why so much demand goes unserved even where the risk is knowable. Something else is in the way.

The cost that never makes the headline

A loan is not a moment of approval. It is a relationship that has to be operated: onboarding, disclosures, statements, reminders, hardship handling, collections, complaints. Every step is bound by rules that differ by product, channel and jurisdiction. For small-ticket and thin-file lending, that operational load barely shrinks while the revenue per account does. The economics break not because the borrower is too risky to price, but because the borrower is too expensive to serve safely.

And the cost is not only labour. It is risk. As covered elsewhere in this series, a single mishandled communication carries real statutory exposure. Every cheap way to operate at scale tends to be the unsafe way, and every safe way tends to be the expensive, manual one. That trade-off is the quiet reason the frontier of who gets served sits where it does.

Lower the cost of safe, and the frontier moves

This reframes the problem in a useful way. If the binding constraint on inclusion is the cost and risk of operating a loan, not pricing it, then the highest-leverage move is to make safe operations cheap. Drive down the marginal cost of a compliant communication, a serviced account, a hardship conversation, and the math on the underserved shifts.

That is the leverage: when compliant operation stops being the expensive option, accounts that never penciled out start to, not by taking on risk the institution wouldn’t, but by making the safe way the affordable way. It is the same reason we are building toward the whole stack, origination and a safe AI underwriter included, so the system that makes operations cheap can also widen who is reachable in the first place. The credit gap was never only about who deserves a loan. It was about who could be reached without breaking the rules or the budget.