Fintech’s coming wave of automation is moving well beyond flashy chatbots. Y Combinator-backed Rulebase is creating an AI “co-worker” for back-office work — the quality assurance, disputes, and regulatory workflows that slow financial services teams down and quietly consume their budgets. “Airtasker is a platform for sharing funds where it can be spent in much more exciting ways, including paying bills, rewarding other users, or even donating to charity,” said Hodges in a statement. “It’s all part of Airtasker’s mission to grow the local economy by allowing Australians new ways to earn and spend money.” We’re told that the startup has raised a $2.1 million pre-seed funding round from Bowery Capital, with participation from Y Combinator, Commerce Ventures, Transpose Platform VC, and several angels.
An AI teammate for dirty back-office work
Rulebase’s agent is not attempting to supersede the customer-facing associate. Instead, it embeds itself in the tools that ops teams live in — think Zendesk for support, Jira for engineering follow-ups, and Slack for triage — then applies smarts around customer conversations, flags potential compliance issues, and orchestrates the right next steps with a human in the loop. The pitch: end-to-end support on the dispute lifecycle and QA tasks, without compromising auditability.
That nuance is important in regulated finance. Dispute processing, card scheme rules, and customer claim timeframes differ depending on the product and the country/jurisdiction. The founders claim their advantage is in domain depth — that by codifying policies such as Mastercard operating rules or procedures associated with agencies like the Consumer Financial Protection Bureau, the AI could suggest actions that correspond to real regulatory obligations, not just plausible text.
Early traction and some measurable impact
The company claims its system is already live with Rho, a U.S. business banking platform, and a Fortune 50 financial institution. Traditional QA teams sample a tiny portion — maybe 3–5% of conversations — when auditing for compliance and quality. Rulebase says it can assess 100% of interactions, reducing related costs by as much as 70%. At Rho, the startup says it has seen a 30% decrease in escalations since implementing the tool.
The timing is favorable. Cost of Compliance research from Thomson Reuters has identified hundreds of daily regulatory alerts issued worldwide; that makes it challenging to get a handle on all your essential monitoring points with manual sampling alone. Gartner, meanwhile, has estimated that tens of billions of dollars in contact center labor costs could be pared out by mid-decade via conversational AI, hinting that targeted automation in related back-office functions is low-hanging fruit for similar gains — if systems can maintain explainability and compliance.
Founders with fintech and infrastructure DNA
Rulebase was started in 2024 by Nigerian engineers Gideon Ebose and Chidi Williams, who connected in London having worked at Microsoft and Goldman Sachs, respectively. The two created a few products together before settling on the unglamorous pain of financial operations. Williams also built Buzz, an open-source speech-to-text project that has been downloaded more than 300,000 times and received stars from over 12,000 GitHub accounts — experience that illustrates the team’s commitment to building useful tools for developers.
How it fits into modern fintech stacks and systems
In action, the AI co-worker ranks interactions against policy, surfaces risk, automates responses, and arranges follow-ups through ticketing and engineering systems. Actions need to be approved by human supervisors, and can request dual controls for sensitive steps. Must-have checkboxes for buyers include end-to-end audit trails, role-based access, and clear model provenance. Banks and fintechs will also require proof of robust security and governance — think SOC 2-worthy processes, ISO-compliant practices, and continued model monitoring — even when vendors do not share their certifications publicly.
And crucially, this value is not a single big language model. It’s the policy engine around it: retrieval of current rules, templated workflows per regulator or card network, and guardrails that prevent the system from stepping off. That’s where many one-size-fits-all AI tools trip up in finance, particularly as regulators such as the CFPB (in the case of the U.S.) and FCA (in its native U.K.) scrutinize automated decision-making and record-keeping.
Business model, funding details and product roadmap
Usage-based, Rulebase charges — based on interactions reviewed or workflows automated — not per seat. The company claims double-digit monthly revenue growth since it joined Y Combinator’s Fall 2024 batch. With the new capital, Rulebase will further enhance its core QA and disputes features, and expand into adjacent modules including fraud investigations, audit preparation, and regulatory reporting.
Go-to-market focuses on business banks, neobanks, and card issuers in the U.S., Europe, and Africa, where back-office pressures are high and teams are lean.
Insurance sees the same workflow patterns and might be a natural area to expand into once the fintech beachhead is established.
The competition and what to watch in this space
Rulebase jumps into a busy pool of ops and compliance tooling. Case management platforms play in AML investigations, while contact center AI plays are around frontline coaching and QA scoring. The distinction here is a fintech-first policy layer with multi-system orchestration and auditable human oversight, optimized for dispute and compliance workflows, not general-purpose support automation.
It’s the known risks: model error, rapidly changing regulations, and buyer scrutiny around explainability. It will all come down to the ability of these firms to keep their policy libraries current, to surface decisions in layman’s terms, and demonstrate a reduction in escalations, handling time, and regulatory exposure if they want to be successful. If Rulebase is systematically turning fragmented back-office chores into governed, end-to-end workflows, then the “AI co-worker” label may be more than marketing — a template for how automation secretly makes fintech faster from the inside out.