The bank used to own the customer because the bank owned the ledger. That equation is breaking. As Ashish Kolte argues in a recent analysis for fintechnews.org, financial primitives — payments, KYC, lending decisions, FX — are migrating out of monolithic core banking systems and into the SaaS, ERP, and commerce platforms where economic activity actually happens. The institution that holds the deposit no longer controls the relationship. The platform that controls the checkout does.
That shift is rewriting how product teams think about payments infrastructure, where venture capital flows, and which incumbents get squeezed into commodity rails. Here’s what’s actually changing, and what it means if you’re the one wiring this stuff into a product.
The Transaction Layer Has Already Fragmented
Over the last decade, Kolte notes, processing has broken apart from closed bank-operated networks into lightweight processors, real-time settlement rails, and digital wallets — each callable as an independent service from third-party software. Checkouts, loan approvals, and reconciliations now happen inside ERPs and SaaS apps, not inside a banking portal.
That inverts where reliability engineering has to live. When a bank owned the whole stack, downtime was the bank’s problem. When your SaaS app embeds five different financial primitives from five different providers, downtime is suddenly your problem, and your customers don’t care which vendor failed. As Kolte notes, reliability, speed, and reconciliation have become first-class design considerations rather than back-office concerns.
If you run a B2B invoicing platform, this is concrete: the moment you add embedded payouts, you’ve taken on the operational burden of partial settlement failures, ACH return codes, and FX timing mismatches that used to belong to a treasury team at a bank. Teams shipping this kind of integration increasingly need custom payment gateway integration work that treats reconciliation as a product feature, not an afterthought.
My take: in 2026 and beyond, payments uptime SLAs will become a competitive feature on SaaS pricing pages, the same way security certifications became one in the late 2010s.
APIs Are the Product Now, Not the Plumbing
Kolte identifies three things that define a strong embedded finance API ecosystem: the breadth of primitives offered, the governance around quality and compliance, and the developer experience (docs, SDKs, sandbox). Companies that get all three right extend their distribution through partners. Companies that gate their tech miss the distribution entirely.
That cultural shift bites hardest at incumbent banks and processors. Treating an API as a product means version management, deprecation policies, a public changelog, a sandbox that mirrors production, and a developer relations function. Most legacy financial institutions don’t have any of that muscle. Most fintech challengers were built around it.
Picture a mid-market lender that wants to offer point-of-sale financing inside a Shopify-like commerce platform. If their underwriting API takes 800ms, has no sandbox, and returns cryptic error codes, the platform will pick a competitor whose SDK drops in cleanly and returns a decision in under 300ms. The lending decision isn’t the product anymore — the API is. Building that kind of integrations and custom API layer is now table stakes for anyone selling financial primitives to platforms.
My prediction: within the next two years, we’ll see at least one major bank publicly restructure its commercial banking arm around API product managers rather than relationship managers. The org chart will follow the revenue.
Capital Is Flowing to the Orchestrators, Not the Originators
Kolte’s read on capital allocation is sharp: investment is shifting toward firms that can extract embedded margins from data flow — credit scoring, dynamic pricing, micro-billing — rather than firms that book a single transaction fee. He identifies two distinct flows: capital into infrastructure orchestration layers that power many downstream experiences, and capital into vertical embedded-finance plays that pair domain expertise with financial services.
The risk he flags is over-allocation to companies optimizing distribution scale without underlying unit economics — a familiar pattern for anyone who watched the BNPL bubble inflate and deflate.
If you’re building a vertical SaaS for, say, dental practices, the point is clear: bolting on payments, payroll, and working capital lending against the data you already collect is now a more investor-friendly story than charging a higher seat license. The valuation multiple lives in the financial services revenue, not the subscription line. Teams in the fintech and banking software space should plan for audit-grade architecture from day one, because the moment financial revenue shows up, regulators do too.
Risk Migrates With the Functionality
When non-banks make decisions powered by bank-provided capabilities, compliance boundaries blur. Kolte points out that API downtime, latency, and sync failures can cascade across partner ecosystems, and that regulators are responding by focusing on interface resilience, contractual responsibility, and data lineage transparency.
Most product teams underestimate this. Embedding KYC from a vendor doesn’t transfer the regulatory obligation — it just adds a vendor to your compliance surface area. Circuit breakers on APIs, end-to-end observability, and graceful degradation in the customer UX (Kolte’s recommended mitigations) need to be designed in, not bolted on after the first incident.
My take: the next major embedded finance failure won’t be a fraud event — it’ll be a multi-platform outage triggered by one upstream API provider, and it will accelerate regulatory rules on third-party risk management for non-bank platforms.
Data and Underwriting Get a Quiet Upgrade
The last shift Kolte highlights: context upgrades underwriting. Transaction data, telemetry, inventory movement, and booking patterns let lenders and insurers build richer risk profiles than a credit bureau pull could ever produce. He notes the upside — better inclusion — alongside the obligation: model fairness and explainability become regulatory and reputational make-or-break issues.
A marketplace that sees a seller’s order volume, refund rate, and shipping consistency knows more about that seller’s creditworthiness than a FICO score does. That informational edge is the actual moat in embedded lending, and it’s why platforms with deep operational data are out-pricing traditional lenders on the same borrowers.
FAQ
Q: What is embedded finance, in plain terms? A: Embedded finance is the practice of delivering financial services — payments, lending, insurance, accounts — inside non-financial software where the customer is already working. Instead of sending a user to a bank’s portal, the SaaS app or marketplace handles it natively through APIs from a regulated provider.
Q: Why are banks at risk if they still own the ledger? A: Because, as Kolte argues, value accrues to whoever controls the customer experience and the data flow, not necessarily to whoever holds the ledger. If a SaaS platform originates the customer relationship and routes transactions through commodity rails, the bank becomes an interchangeable utility competing on price.
Q: What should a product team prioritize when embedding financial services? A: API reliability, observability, reconciliation tooling, and a clear contractual division of compliance responsibilities with the underlying provider. Kolte’s framing is that resilience and transparent commercial terms separate the platforms that capture embedded finance value from those that just take on its risks.
Key Takeaways
- Treat payment and financial APIs as core dependencies with their own SLAs, runbooks, and on-call rotations — not as vendor integrations you can forget about after launch.
- If you sell financial primitives, invest in your sandbox and SDK before you invest in your next feature; developer experience is now the distribution channel.
- Vertical SaaS teams should model embedded finance revenue into their 24-month roadmap, because investor expectations are already pricing it in.
- Build observability and circuit breakers around every third-party financial API before regulators force you to — the next embedded finance enforcement wave will target non-bank platforms, not just the banks behind them.
- Audit your data and underwriting models for explainability now; the platforms that win the next cycle will be the ones whose risk decisions can survive a regulator’s questions.