Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Y combinator-alam Rulbes The bet is that the next wave of automation in financial services will not be about the slick AI interfaces, but the work of continuous back-office as consent.
The startup, founded in London, two Nigerian engineers, Gidion Abos and Chidi Williams, just raised a $ 2.5 million pre-$ 2.5 million pre-$ 2 million in addition to the participation of Wi-Combinators, Trade Ventures, Transpose Platform VC.
Financial service companies spend a lot of efforts on support tickets, resolving disputes, assurance of quality guarantee and regulatory consent. Rulbes’ software, which is called an agent colleague, replaces most of the Manual Grant’s work in these functions. Its AI agent can evaluate customer interaction, flag regulatory risks and triggers the right follow-ups on equipment such as gendesk, cumin and slack, which may not lose human-in-loop supervision.
CTO Williams says “Our ‘fellow’ tool consolidates the platform across the platform and helps the human agent and back-office parties to fully handle the life cycle of disputes while saving the errors and maintaining consent,” CTO Williams said. Currently, the old startup of the year has already been deployed to customers like US Business Banking Platform RHO and an unnamed Fortune 50 financial institution.
Rulbes founders were not the first to do. Former Microsoft and Williams Product Lead Ebos, former back -finger engineer of Goldman Shutch, made several products like AI Customer Response Empty before finally settled in Rulbes. This idea came after seeing how back-office operations were inefficient in small and large financial institutions, especially when it came about the regulatory work flow.
The startup is currently concentrated on workflies triggered by customer service interaction, with its first shoots around the assurance of quality. The QA analysts of the Traditional Test Financial Institution usually manually review 3-5% support interaction so that the reps compliance protocol follows the compliance protocol.
The founder says that Rulbes now evaluates 100% of these national interactions, reduces the costs by up to 70%. For example, in the case of RHO, Rulbes has helped to spend up to 30%extra.
TechCrunch event
San Francisco
|
October 27-29, 2025
In an interview to TechCrunch, CEO Abos said, “We automatically automatically start with customer interaction, we are well-managed from end to end,” “most of them, though most of it, conflicts with consent and customer calls and messages, long-term aims to make our aim and tabs as much as possible. Make. “
The new fund will help to double engineering and finally add new features to AI colleagues such as fraud investigation, monitoring preparation and regulatory report.
Rulbes is now focusing on financial services because automation demands accuracy. “The rules of your MasterCard, the CFPB timines need to understand that the depth of the domain knowledge is our rock,” said Ebos.
They are the targets of business banks, neobanks and card zerokeepers throughout Africa, Europe and the United States, but on the roadmap can eventually include adjacent verticals like insurance, where similar work flows exist.
The founder says the revenue has been increasing rapidly with the growth of the “double-digit” month since joining the 2024 batches of the Wi-Combinator. Rulbes’ business model is based on usage, every interaction review or workflow automatically charge.
As one of the few founders of African to go to the YC Building AI equipment, Ebos and Williams consultants tried to accept the global accelerator is advised to think worldwide from the first day.
Williams commented, “We at a moment where small groups can supply faster, faster than ever, so it seems like ‘x’ or a narrow vertical missed opportunity,” Williams commented. “With AI, it seems clear that you have to go after something huge. You probably never cut it less than the most ambitious version of your idea,” Williams said, who made more than 300,000 downloads before playing Rulbess and made an early OPEN-Source Spext Tool with 12,000 Github stars.
