20VC Founder Harry Stebbings: AI Agents Will Redefine All Financial Software Moats
Harry Stebbings, founder of 20VC, cited SaaStr founder Jason Lemkin's view that the underlying rules of the fintech game are shifting from a "human interface user experience" to an "AI agent interface experience." In the past, success was based on visible dashboards, design, and interaction; now the real "buyer" is the AI agent, which fundamentally cares only about API quality, performance, and integration capabilities. This shift will impact the moats of traditional SaaS and financial infrastructure players. He named several leading fintech founders and investors to seek opinions, suggesting that this is not a marginal trend but a core variable reshaping future valuations and competitive landscapes.
Stebbings has repeatedly emphasized in previous articles and podcasts that the "invisible infrastructure" of the AI agent era will reconstruct the financial and enterprise software stack: more databases and business processes will be automatically created and invoked by AI rather than manually built by human developers. The true barriers for enterprises will shift from "interface and branding" to "underlying permissions, data structures, and reliable interfaces." Related discussions have rapidly spread on LinkedIn and industry podcasts, with more SaaS and financial infrastructure entrepreneurs beginning to view their product positioning through the lens of "Agent Experience (AX)" rather than traditional UX frameworks.
Source: Public Information
ABAB AI Insight
The shift from "human use to agent use" essentially rewrites the question of who the "end customer" of financial software is. Traditional fintech products, no matter how complex their backends, still have humans at the front end: operations, finance personnel, risk control teams, and end users. Therefore, design, branding, and interaction processes can create significant differences and bargaining space. Once the primary users become AI agents, they won't care about how aesthetically pleasing a page is; they will only care about whether the API is stable, the limits are generous, and the latency is low. This will compress a lot of the premiums that were previously maintained by "human experience" into a pure infrastructure competition.
This will lead to a two-layer moat reshuffle. First, application-layer fintech products that started with "interface + marketing" will be forced to compete on dimensions like "whose API is better, whose data is cleaner, and whose risk control rules are more machine-readable." The defensibility of human UX will be weakened, and traditional brand advantages will be nearly zero in the eyes of agents. Second, the real high barriers will shift to players who control clearing channels, regulatory licenses, underlying account systems, and high-quality data assets—because whether human or AI, to get close to money and risk, one must ultimately pass through these hard constraints.
On a deeper level, this is another abstraction following the "internetization of finance": the internet era abstracted branches into apps, while the AI era abstracts apps into API call graphs, further marginalizing front-end interactions and adding a layer of "machine agents" between humans and software. In this structure, whoever can become the "default ledger and payment layer for agents" will gain new pricing power and traffic entry; while those middle-layer products that previously gained premiums through UI innovation may be viewed as "replaceable intermediaries" after AI bulk comparisons, facing risks of being bypassed or packaged as white-label components.
For the capital markets, this means fintech valuation models need to be adjusted: it can no longer be simply calculated as "user count × ARPU × SaaS multiple," but must consider its position in the future agent economy—whether it is seen as a replaceable plugin by agents or as a foundational facility that controls access rights, clearing rights, and data rights. This type of structural migration is often reflected superficially through "API metrics, latency, and integration costs," but fundamentally represents a reshuffle of pricing power and bargaining order.