The branch's monopoly ended on paper. It has yet to end in the price.

Digitizing the branch does not destroy it; it merely removes the chair — cost does not fall with permission, it falls with advantage: exclusive information, control of the flow, and better recovery.

The National Monetary Council has just authorized two species of credit institutions native to the digital environment: the Direct Credit Society and the Peer-to-Peer Lending Society. The branch's monopoly ended on paper. It has yet to end in the price.

The difference matters. For decades, the company that needed capital compared banks operating on the same logic: branch, account, service bundle, blanket guarantees, and concentrated decision-making. The internet made it possible to distribute credit without a branch, but distribution is not structure. A faster form can deliver the same bad product in fewer minutes. Digitizing the branch does not destroy it; it merely removes the chair.

The SCD will be able to lend with its own capital through an electronic platform. The SEP will be able to intermediate funds between creditors and debtors. They are different models, but both break a cultural association: credit does not need to be born inside a universal bank. That allows specialized companies to pick a segment, learn its data, build specific collection, and price risks the bank treats as noise.

Specialization can reduce fear. A creditor that serves only trucking companies understands truck values, freight seasonality, maintenance costs, shipper concentration, and fuel behavior. A creditor of clinics knows insurance receivables, claim denials, equipment, and payment cycles. The generic bank sees a taxpayer ID, revenue, and a personal guarantee. The vertical fintech can see the operation. Better information should produce a better price.

"Should" is the dangerous word. Technology reduces acquisition and analysis costs, but it also reduces the friction that protected the client from borrowing impulsively. An elegant interface can turn leverage into a one-click purchase. The most accessible credit is not necessarily the cheapest credit; it may just be the credit that best hides the total cost. When speed becomes the value proposition, the quality of the decision tends to stay off-screen.

The sector will celebrate alternative data, artificial intelligence, and new scores. There is merit in that. Small-company balance sheets are late, incomplete, and frequently useless. Bank flows, invoices, sales, payments, and behavior can reveal risk in real time. But an algorithm trained during an economic expansion may learn only that clients pay while credit remains available. Sophisticated correlation does not replace causal understanding.

The test is not predicting who will pay in the observed scenario. It is knowing what happens when fuel prices rise, a client concentrates revenue, the platform changes its rules, the sector loses demand, or the cost of capital increases. Models are excellent at interpolating the past and dangerous when marketed as knowledge of the future. The fintech that claims to have eliminated default has probably eliminated only humility.

The resolution creates a license to compete, not automatic economic advantage. SCDs need their own capital; they will grow until funding becomes the bottleneck. SEPs depend on investors willing to carry risk and on rules that prevent conflicts. Both will face client acquisition, fraud, collection, data protection, regulatory capital, and cycles. The bank has expensive infrastructure, but it has deposits, brand, and decades of recovery. The incumbent is not slow by accident; part of the slowness is scar tissue.

The biggest opportunity is not offering personal loans or generic working capital a few points cheaper. It is redesigning the obligation around the asset. Inventory financing connected to sales; receivables prepayment with verification at the source; equipment credit with monitoring and collateral; supply-chain capital where the anchor buyer reduces the supplier's risk. The platform should be less a loan website and more a control system for the financed flow.

When the creditor controls the information and the movement of the asset, he needs less indiscriminate collateral. He can charge less without pretending the risk disappeared. The traditional bank demands the owners' personal guarantee because it does not want to understand each operation. The fintech can replace part of the guarantee with data and control. That is the real attack on the branch: not removing the manager, but removing the ignorance that makes the manager powerful.

The temptation to sell credit as inclusion will also appear. Inclusion without adequacy is merely the distribution of fragility. An excluded company may be underserved or may be insolvent. The new entrant must distinguish the two. If it confuses repressed demand with good risk, it will grow rapidly until the first adverse cycle. The initial expansion will look like product proof; the mature vintage will reveal whether it was adverse selection.

I would invest in credit fintechs only where there was an advantage of proprietary data, integration into the operational flow, and funding discipline. I would avoid models whose differentiator was marketing, fast approval, and abundant third-party capital. Low acquisition costs can vanish, funding can close, and competitors can copy an interface. What remains is a position inside the transaction that produces exclusive data and superior collection capacity.

I would prefer platforms embedded in management software, marketplaces, or supply chains. They observe the business before the credit request. They know sales, inventory, recurrence, and behavior. The bank knows the client when he needs money, exactly the moment when the information is most biased. The operational platform knows the client when he is not yet asking. That precedence is worth more than a purchased score.

On the credit investor's side, I would demand alignment. The platform that originates and sells its entire exposure gains with volume and leaves the buyer with the loss. Some risk retention, performance-linked compensation, and vintage transparency are essential. The "originate to distribute" model is not new; 2008 already demonstrated how bad incentives hide behind financial innovation. An app does not repeal history.

Regulation should allow experimentation without turning the license into a seal of quality. The public will confuse authorization to operate with approval of the product. They are different things. The regulator verifies requirements and conduct; it does not guarantee that the rate is fair, the algorithm is good, or the portfolio will survive. The fintech will use the word "regulated" as a trust argument. The investor should read it as a minimum condition, not a recommendation.

Banks will respond. They will buy fintechs, create digital brands, cut prices in attractive segments, and use cheap deposits to compress margins. The new entrant that competes only on rate will face an adversary with potentially lower cost of capital. The defense must lie in information, niche, and product. Being friendlier than a bank is easy; being a better creditor is hard.

For the entrepreneur, the multiplication of offers will demand more capacity, not less. Comparing total effective cost, tenor, amortization, collateral, covenants, and effects on cash flow. An apparently lower monthly rate can hide insurance, fees, or an amortization that squeezes cash. The abundance of interfaces will create a false sense of choice. Ten apps selling the same loan do not constitute a competitive market; they constitute a shelf in ten colors.

The truly transformative fintech will start with the question the bank avoids: what is being financed? If the answer is "the company," it has already lost part of the opportunity. Companies are containers of different flows. Inventory, contract, receivable, equipment, and expansion carry distinct risks and natural buyers. Separating them lets the right capital finance the right asset. The branch mixes everything and charges for the worst.

I see a bigger business in the decision layer. A platform that recommends how much to sell on credit, what limit to grant, what collateral to require, when to collect, and when to cut exposure. Credit does not begin when someone asks for a loan; it begins when a company sells today and receives tomorrow. Millions of suppliers are already involuntary banks, without models, contracts, or collection. The fintech that understands this will find a market bigger than conventional lending.

Resolution 4,656 opens the legal door. It does not guarantee that anyone will walk through it intelligently. The first generation will probably copy banking products and win on experience. The second will integrate credit into operations. The third may turn risk into invisible infrastructure, priced at the moment of the transaction. It is in that third generation that the price can finally change, because the design will have changed first.

The branch's monopoly ended on paper. It has yet to end in the price because cost does not fall with permission, it falls with advantage. Whoever merely digitizes distribution will be compressed. Whoever reduces uncertainty, controls the flow, and recovers better can turn the license into a market. The rest will discover that a bank without a branch can still carry all the vices of a bank, minus the cheap funding.

Leo Bentier