Wrapping legacy banks in AI agents - so institutions can modernize without ripping out the core.
FlowX.AI - the AI-native platform for regulated industries, photographed here as its corporate mark. The company splits a US headquarters with an engineering base in Romania.
Every large bank carries the same quiet burden: a core built decades ago, often on COBOL and mainframes, that still processes mortgages, payments and claims every day - and that nobody dares to replace. FlowX.AI's proposition is to leave that core in place and build a layer of AI agents around it, so institutions can modernize their processes in weeks rather than the years a rip-and-replace migration usually demands.
Founded in 2021 by Ioan Iacob, Radu Cautis and Serban Chiricescu, the company describes itself as an AI-native platform for building and deploying enterprise-grade AI agents in highly regulated industries - banking first, then insurance and adjacent sectors. The pitch is deliberately unglamorous: not a chatbot demo, but audited, on-premise, production-grade automation for systems that cannot afford to be wrong.
The company keeps a headquarters in the United States - registered in Delaware, with a New York address - while its core engineering team works out of Bucharest, Romania. That cross-border shape is common among European enterprise-software firms selling into US and Western European markets, and it has not slowed FlowX.AI's fundraising or its reach into large banking groups.
What makes the approach notable is where it starts. Rather than asking a bank to migrate, FlowX.AI plugs into what already exists - core platforms like Temenos, Finastra, FIS and Jack Henry, plus mainframes and transaction-management systems - and layers agents on top. The result, the company argues, is faster time-to-value and measurable return on investment.
It is a crowded promise. Plenty of vendors claim to modernize the enterprise. FlowX.AI's answer is to lead with the constraints of the regulated world - governance, audit trails, data privacy - and treat them as the product rather than an afterthought.
"FlowX.AI is the world's first AI enterprise application modernisation platform for large financial services organisations." - Company positioning, cited across launch coverage
The uncomfortable statistic that runs through FlowX.AI's marketing is that the large majority of enterprise AI projects stall before they ever go live. Pilots impress in a demo, then collapse on contact with integration, security review and compliance. The company frames its whole reason for being around getting customers into the small minority that actually ship - "Be in the 5% that succeed."
In banking, the gap between demo and production is especially wide. A model that occasionally invents an answer is a curiosity in a consumer app and a liability in a lending decision. So FlowX.AI leans on what it calls banking-grade safety: audit trails, on-premise execution to keep data inside the institution, and controls designed to eliminate hallucinated outputs in regulated workflows.
The differentiator it emphasizes is integration without disruption. Because the agents connect to existing core systems rather than replacing them, a bank can start with one process - onboarding, a lending flow, claims - prove the return, and expand. "Start. Prove ROI. Scale," as the company puts it.
That positions FlowX.AI against a familiar field of low-code and process-automation incumbents such as Pega, Appian, Mendix, OutSystems and Flowable, as well as newer agentic-AI entrants and RPA players like UiPath moving toward agents. Its wedge is the combination of pre-built financial-services agents and a compliance-first architecture aimed squarely at the regulated buyer.
An AI-native agentic platform combining enterprise architecture, AI-native development and banking-grade security in one foundation.
140+ pre-built agents (marketed as 220+ across categories) spanning onboarding, lending, underwriting, claims, invoice reconciliation and churn prevention.
Custom agent development, letting teams design and adapt agents for their own regulatory and product needs.
A multi-agent, "organically evolving" architecture - including an Auditor Agent that reads regulatory updates and proposes workflow changes for human approval.
Plug-and-play connectivity to Temenos, Finastra, FIS, Jack Henry, COBOL mainframes and TMS - without changing the core.
B2B platform licensing sold to banks and insurers, typically via multi-year engagements and system-integrator partnerships.
In the 2025 release, FlowX.AI described an Auditor Agent that interprets a regulatory update, identifies which workflows need to change, and submits proposals to developers. Once a designated human operator approves, builder agents implement the change in seconds. It is a concrete picture of what "agentic" is supposed to mean in a bank - not a conversational assistant, but a chain of specialized agents doing the tedious, high-stakes work of keeping software compliant, with a person kept firmly in the loop.
FlowX.AI's customers are large financial institutions and insurers in highly regulated markets, primarily across Europe. Coverage of its funding named several banking clients: OTP Group, a major European banking group; Banca Transilvania, one of the largest banks in South-Eastern Europe; and Alpha Bank, a large group in Central Europe. The company says its technology has driven hundreds of millions of dollars in ROI for leading institutions.
Those workflows sit at the center of the business, not the edges - retail and commercial lending, underwriting, onboarding, claims, mortgages and regulatory processes. That is the point: FlowX.AI aims at mission-critical systems, where the upside of automation is large and the tolerance for error is small.
Within the market, the company slots into the overlap of three trends - application modernization, low-code/no-code development, and the rise of agentic AI. Gartner listed FlowX.AI in its 2025 Digital Banking Market Guide among players bringing AI to financial services, a signal that analysts see it as part of the modernization conversation rather than a peripheral tool.
The wager is that regulated industries will be slower but larger adopters of AI agents than consumer software - and that the vendor who solves governance, integration and safety first will be positioned when those budgets move.
Ioan Iacob, Radu Cautis and Serban Chiricescu launch the company to modernize enterprise applications with AI.
Closes an ~$8.5M seed round led by PortfoLion to build out the modernization platform.
Dawn Capital leads a $35M Series A - reported as Romania's largest at the time - to fund product and global expansion.
Grows its catalog of ready-made AI agents across banking and insurance use cases.
Launches a multi-agent evolving architecture and is named in Gartner's 2025 Digital Banking Market Guide.
FlowX.AI has raised more than $43M in total. Its 2023 Series A of $35M was led by London-based enterprise-software investor Dawn Capital, with existing backers PortfoLion, SeedBlink and DayOne Capital participating. At the time, the round was widely reported as Romania's largest Series A - surpassing a record previously held by UiPath.
Led by PortfoLion.
Led by Dawn Capital; PortfoLion, SeedBlink, DayOne Capital participating.
Across seed and Series A rounds.
It provides an AI-native platform and pre-built AI agents that let large banks, insurers and regulated enterprises modernize legacy systems and automate mission-critical workflows without replacing their core infrastructure.
It was founded in 2021 by Ioan Iacob (CEO), Radu Cautis and Serban Chiricescu.
More than $43M in total, including a $35M Series A led by Dawn Capital in May 2023 and an earlier ~$8.5M seed round led by PortfoLion.
Large financial institutions in regulated markets, primarily in Europe. Reported clients include OTP Group, Banca Transilvania and Alpha Bank.
It focuses on banking-grade safety - audit trails, on-premise execution and integration with core systems like Temenos, Finastra and mainframes - and ships pre-built agents to reach production in weeks rather than years.