US-based ArcOne AI has expanded its BankOS platform to cover additional banking domains, including capital markets, wealth, and payments.
The expansion builds on the company's existing revenue intelligence offering, ArcOne EPM, and adds an AI command centre, additional agents, new connectors, and a unified semantic layer. The company has also updated its governance and compliance framework to align with current regulatory requirements for banking environments.
According to ArcOne AI, financial institutions have found it difficult to move AI initiatives beyond pilot stage due to structural data challenges. Large banks typically operate between five and 15 core banking systems concurrently, each with its own data model and integration requirements, while capital markets and global transaction banking environments introduce further complexity through proprietary data structures, real-time pricing feeds, and cross-border regulatory obligations. In addition, gslobal investment in AI within financial services was estimated at USD 35 billion in 2023, with projections indicating this figure will nearly triple, according to the company.
ArcOne BankOS is built on the Ocular AI platform, which the company describes as comprising three integrated components. The data layer, referred to as the Banking Domain Cartridge, provides over 60 connectors spanning banking cores and enterprise systems, and automatically maps over 80% of fields across major banking cores. The intelligence layer functions as an agent factory, using an orchestration engine called TERRA to coordinate a library of over 100 AI agents across banking domains. The agent layer comprises three products: Enrich360, covering pricing and profit intelligence; Experience360, covering customer experience; and Exceptions360, covering process and exception management. A multi-modal interface, LYZA, connects banking staff to the agent ecosystem through channels including web, voice, text, video, and document inputs.
Deployment timelines and governance
ArcOne AI states that certified channel partners can compress initial deployment to four to six months, following a three-step model of connecting, mapping, and activating the platform. The system is described as modular and cloud-agnostic, requiring no large data engineering team and no replacement of existing core banking infrastructure.
On governance, ArcOne BankOS has been built to ISO 42001 responsible AI standards, with audit-ready data lineage intended to support compliance with SOX, OCC, CFPB, and SR 26-2 model risk management requirements. The Ocular AI catalogue includes monitoring for bias, fairness, toxicity, hallucination, and lineage, which the company states is designed to make AI-driven decisions traceable across data, models, agents, and business processes.
The company also said the platform is intended to help banks move from contract to measurable outcomes rather than extended proof-of-concept phases.
Current deployments
ArcOne AI cites three examples of existing deployments: a regional bank running account analysis and deal management on ArcOne EPM for two years, a multinational financial services company operating in 196 countries using the platform for real-time dynamic pricing, as well as a global bank using Exceptions360 to automate quote-to-cash processes and reconcile fees with billing across multi-entity structures.
ArcOne BankOS, along with Enrich360, Experience360, and Exceptions360, is available to existing and new clients as standalone products or on top of any banking core or revenue management system. The company states it has active engagements with Fortune 500 clients and partnerships with fintechs, banking core providers, and systems integrators.
