The AI Agent That Never Sleeps: Rulebase Tackles Fintech’s Revenue Leaks

NEW YORK, NY – July 08, 2026 – In the digital labyrinth of modern financial services, the gap between a customer's initial interest and their first transaction has become a costly chasm. It’s a quiet failure point, a structural weakness where billions in potential revenue vanish not due to a single catastrophic error, but through a thousand tiny instances of incomplete follow-up. Today, New York-based Rulebase announced a new type of AI workforce, dubbed Revenue Agents, designed to relentlessly patrol this gap and recover the money left on the table.

These are not chatbots or simple automation scripts. Backed by Y Combinator, Rulebase has engineered what it calls “long-horizon” agents that treat each stalled customer application as a persistent case file to be worked over days or even weeks. The goal is to transform the industry’s chronic problem of revenue leakage from an accepted cost of doing business into a recoverable asset.

Plugging the Digital Gaps

The core bottleneck to growth in any regulated business is the friction-filled onboarding process. A potential customer starts an application, provides most of their information, but stalls on a single rejected document. An account is fully approved but never funded. A user makes one transaction and then goes dormant. In the high-volume, ticket-based world of operations, these cases are often flagged and then lost in a queue, as human agents are incentivized to close interactions and move on.

“Every fintech is leaking revenue it has already won,” said Gideon Ebose, CEO and co-founder of Rulebase, in a statement. “That revenue isn't lost to bad decisions, but to follow-up work that nobody owns long enough to finish.”

Rulebase’s solution is to assign a dedicated AI agent to each of these stalled cases. The company claims the results are dramatic, citing a 120 percent lift in recovered accounts and a 50 percent reduction in time-to-activation across early deployments with fintechs in the US, EU, and Africa. While these figures are company-reported and await independent validation, they point to a significant inefficiency in the market. The financial services industry already spends heavily on compliance and manual processing, with costs growing an estimated 15% annually. Rulebase is betting that a persistent, automated workforce can deliver a return that far outstrips its cost.

Beyond the Ticket: The Rise of the 'Long-Horizon' Agent

The technological foundation for this is what Rulebase calls its “Customer Agent Runtime.” This is a departure from traditional automation tools that are built for discrete, short-burst tasks. Instead, the runtime allows an agent to “own” a customer objective—such as completing onboarding or reactivating a dormant account—as a durable, long-lived task.

The system is event-driven. An agent “wakes up” when a specific signal occurs—a customer email, a time-based trigger, an update from a verification system—and then determines the next logical step. This could be drafting an email to request a new document, triaging an exception for human review, or sending a carefully timed reactivation offer. After taking or recommending an action, it records the result and goes dormant again, waiting for the next signal.

This architecture is designed for economic efficiency. By avoiding continuous model loops, the agent doesn't run up high computational costs while waiting for a customer to respond over several days. The economics of the agent remain tethered to the value of the revenue it is trying to recover. This ability to maintain context and state over extended periods without breaking the bank is the key innovation that enables a truly “long-horizon” approach to customer lifecycle management.

The Human in the Machine’s Loop

For all its autonomy, the system is explicitly designed to operate within the rigid regulatory and ethical boundaries of finance. This is not a “fire-and-forget” AI. An action policy governs what each agent is permitted to draft, execute, or escalate. Crucially, consequential customer-facing actions require explicit human approval, ensuring that final judgment and accountability remain with a person.

This “human-in-the-loop” design is a strategic imperative, not a technical limitation. With frameworks like the EU’s AI Act set to classify many financial AI applications as “high-risk,” the ability to demonstrate human oversight, transparency, and auditability is paramount. Rulebase appears to have built this principle into its DNA, stemming from its prior work in the compliance space. Its earlier AI agents, such as the compliance coworker “Gretta,” were designed to automate the review of 100% of customer interactions against dozens of regulations like BSA/AML and UDAAP, providing clients like business banking platform Rho with comprehensive, audit-ready evidence trails.

The Revenue Agents extend this philosophy from observing to acting. They pair the ability to see every signal with the ability to act on it, but always with a human safety net. This reflects a maturing understanding in the industry: the most effective systems are not those that replace humans entirely, but those that augment their judgment and free them from the monotonous, high-volume work that AI is uniquely suited to perform.

From Y Combinator to Wall Street's Back Office

Rulebase’s trajectory suggests a deep understanding of the systems it seeks to improve. Co-founders Gideon Ebose and Chidi Williams bring experience from Microsoft and Goldman Sachs, respectively, combining expertise in scalable product design with the rigorous demands of building defensive systems for financial institutions. Backed by a $2.6 million seed round from prominent VCs including Bowery Capital and Commerce Ventures, the company is well-capitalized to tackle this complex challenge.

Their strategy is not to rip and replace existing infrastructure, but to integrate deeply within it. The Revenue Agents are designed to connect with the systems where financial operations teams already live: ticketing platforms, CRMs, KYC/KYB vendors, and communication channels like email, SMS, and WhatsApp. By weaving its AI workforce directly into the operational fabric of modern finance, Rulebase is making a compelling case that the future of operational efficiency lies not in bigger teams, but in smarter, more persistent systems.