Y Combinator-alum Rulebase is betting that the subsequent wave of automation in monetary providers gained’t be about flashy AI interfaces, however the unglamorous back-office duties like compliance.
The startup, based by Gideon Ebose and Chidi Williams, two Nigerian engineers who met in London, simply raised a $2.1 million pre-seed spherical led by Bowery Capital, with participation from Y Combinator, Commerce Ventures, Transpose Platform VC, alongside a number of angels.
Monetary providers companies spend monumental quantities of effort on help tickets, resolving disputes, making certain high quality assurance, and regulatory compliance. Rulebase’s software program, which it calls an agent coworker, replaces a lot of the handbook grunt work in these duties. Its AI agent can consider buyer interactions, flag regulatory dangers, and set off the appropriate follow-ups throughout instruments like Zendesk, Jira, and Slack with out dropping the human-in-the-loop oversight that monetary companies demand.
“Our ‘Coworker’ software integrates throughout platforms and collaborates with human brokers and back-office groups to totally handle the dispute lifecycle whereas saving time, decreasing errors, and sustaining compliance,” mentioned CTO Williams. Presently, the year-old startup is already deployed at prospects like U.S. enterprise banking platform Rho and an unnamed Fortune 50 monetary establishment.
Rulebase wasn’t the founders’ first swing. Ebose, a former product lead at Microsoft and Williams, a former backend engineer at Goldman Sachs, constructed a number of merchandise collectively like an AI buyer suggestions software earlier than finally selecting Rulebase. The thought took place after seeing how inefficient back-office operations have been in small and enormous monetary establishments, particularly when it got here to regulatory workflows.
The startup at present focuses on workflows triggered by customer support interactions, with its first wedge round high quality assurance. QA analysts in conventional monetary establishments usually manually evaluation 3–5% of help interactions to make sure reps comply with compliance protocols.
Rulebase now evaluates 100% of such interactions, reducing prices by as much as 70%, the founders say. Within the case of Rho, as an illustration, Rulebase has helped minimize escalations by as much as 30%.
Techcrunch occasion
San Francisco
|
October 27-29, 2025
“We automate workflows that begin with a buyer interplay, areas we’re already nice at dealing with end-to-end,” CEO Ebose mentioned in an interview with TechCrunch. “Whereas a lot of that’s QA, compliance, and disputes tied to buyer calls and messages, long-term our purpose is to tackle as many handbook back-office duties as potential by pulling these fragmented steps and tabs into one coordinated workflow.”
The brand new funding will assist it double down on engineering and finally add new options to AI Coworker like fraud investigation, audit preparation, and regulatory reporting.
Rulebase is targeted on monetary providers for now as a result of automation calls for precision. “It is advisable perceive MasterCard’s guidelines, CFPB timelines. That depth of area data is our moat,” Ebose mentioned.
They firm is concentrating on enterprise banks, neobanks, and card issuers throughout Africa, Europe, and the U.S. However the roadmap may finally embrace adjoining verticals like insurance coverage, the place comparable workflows exist.
Income is rising quick, with “double-digit” month-over-month development since becoming a member of Y Combinator’s Fall 2024 batch, the founders say. Rulebase’s enterprise mannequin is usage-based, charging per interplay reviewed or workflow automated.
As one of many few African founders to get into YC constructing AI instruments, Ebose and Williams’ recommendation to founders making an attempt to get accepted into the worldwide accelerator is to suppose globally from day one.
“We’re in a second the place small groups can ship extra worth, extra shortly, than ever earlier than, so limiting your self to ‘X for Y’ or a slender vertical looks like a missed alternative,” Williams remarked. “With AI, it feels apparent that it’s important to go after one thing large. Something lower than probably the most formidable model of your thought probably gained’t minimize it,” mentioned Williams, who earlier than Rulebase constructed Buzz, an early open-source speech-to-text software with over 300,000 downloads and over 12,000 GitHub stars.
