HSBC has launched a new digital banking platform for corporate and institutional clients, but the most significant part of HSBCnio may not be its redesigned interface, mobile access or even its transaction banking capabilities. It is the decision to allow a company’s own artificial intelligence tools to access authorised banking information through a Model Context Protocol connection. That seemingly technical feature points towards a much larger change taking place in corporate banking. For decades, banks have designed digital channels around people: first branches, then desktop portals, mobile applications and APIs connecting enterprise systems. The emergence of enterprise AI creates another type of participant in that relationship. Increasingly, software agents may retrieve banking information, interpret cash positions, monitor payments and eventually initiate financial actions on behalf of companies. HSBCnio provides an early indication of how banks may position themselves inside that emerging machine-to-machine financial environment.
Moving banking beyond the bank portal
HSBC unveiled HSBCnio at Sibos 2026 in Miami, where this year’s theme, “Digital finance for AI-driven economies”, reflected how rapidly AI is moving into mainstream financial infrastructure. The platform brings together transaction banking services, embedded connectivity and AI interfaces, allowing corporate clients to interact through web and mobile channels, connect their own systems directly, or allow authorised AI tools to retrieve banking information.
At launch, clients can use HSBCnio to view cash balances and transactions, track payments and access services involving trade loans and foreign exchange. HSBC also provides APIs, developer resources, software development tools and a sandbox intended to make the platform easier to integrate with corporate ERP and treasury management environments.
Those capabilities themselves are not revolutionary. Banks have spent years developing APIs and connecting corporate banking services with enterprise systems. What is changing is the direction of the interface.
Traditional digital banking assumes that someone ultimately needs to interact with the bank. An employee opens a portal, checks liquidity, downloads information, reviews a payment or requests a transaction. APIs removed some of those manual steps by allowing software systems to communicate directly, but the underlying workflow often remained deterministic: a corporate system requested specific information or instructed the bank to perform a predefined operation.
AI introduces a different model. Instead of software merely following a fixed sequence, an AI system can potentially determine what information it needs, retrieve that information, interpret it and decide what should happen next within the permissions and policies established by the organisation.
That requires banks to think differently about connectivity.
The importance of giving AI permissioned access to banking data
HSBC says HSBCnio allows clients’ own AI tools to access authorised account and transaction information through its Model Context Protocol connection. The bank describes this as a way for companies to introduce AI into treasury workflows while using trusted banking data, existing permissions and appropriate controls.
The distinction between giving an AI model banking information and giving it permissioned access to banking infrastructure is important.
Corporate treasury contains some of an organisation’s most sensitive financial information. Cash positions, payment flows, foreign exchange exposures and transaction histories cannot simply be made available to a general-purpose AI system without controls governing identity, authorisation and access.
The emerging challenge for banks therefore extends beyond providing APIs. They need mechanisms through which AI systems can interact with financial data while remaining subject to the same security, permission and governance structures expected of other banking channels.
This is where HSBCnio becomes more interesting strategically. HSBC is effectively preparing an interface through which an authorised AI system can become another way of interacting with the bank.
Initially, that interaction remains largely informational. HSBC says it is also developing “Ask HSBC”, an AI-enabled capability intended to allow users to query and interpret their banking information. But once AI can securely understand a company’s banking position, the natural next stage is for those systems to assist with increasingly complex treasury decisions.
Corporate treasury is becoming an AI battleground
The timing matters because corporate treasury is emerging as one of the areas where enterprise AI could have a particularly significant impact.
The Association for Financial Professionals’ 2026 Treasury Benchmarking Survey found that AI and automation had entered the top five priorities for treasury organisations, cited by 30% of respondents. At the same time, 38% identified managing the opportunities and risks associated with AI as a significant challenge, while 35% cited using AI to automate manual processes. Cash and liquidity forecasting remained treasury’s biggest priority and challenge.
That combination explains why access to trusted banking data matters.
AI can be powerful at interpreting information and identifying patterns, but treasury decisions depend on accurate and current information about balances, transactions, payments, currencies and liquidity. Connecting AI securely to banking data could therefore allow corporate systems to move beyond analysing information exported from banks and towards interacting with financial information as part of a continuous workflow.
This transition is already attracting attention across the transaction banking industry. BCG argues that corporate treasury is moving towards increasingly automated and agentic workflows in which AI systems could eventually route payments, manage currency exposures and optimise liquidity according to policies established by the company. It argues that transaction banks will increasingly need to design their systems for “machine counterparties”, exposing secure real-time information while maintaining authorisation, monitoring and audit controls.
HSBCnio fits directly into that trajectory.
The bank may disappear from the interface
There is another consequence that could ultimately prove even more important for banks.
If corporate AI systems become capable of interacting directly with banking infrastructure, employees may have progressively less reason to visit individual bank portals.
Imagine a multinational company working with ten banks. Today, its treasury infrastructure may already aggregate much of that information through a treasury management system. In an agentic environment, however, an AI layer could potentially monitor liquidity across institutions, identify upcoming obligations, examine foreign exchange exposure, compare available options and recommend or execute actions according to predefined corporate policies.
The employee would interact with the company’s AI environment rather than each bank.
The banking relationship therefore risks moving one layer further away from the financial institution’s own interface.
This creates an important strategic question for transaction banks: if AI becomes the interface through which companies manage money, where does the bank sit in that relationship?
Banks could remain largely providers of financial infrastructure behind third-party AI systems, or they could attempt to become part of the intelligence and orchestration layer themselves.
HSBCnio appears designed to accommodate both possibilities. HSBC is building its own AI experience through Ask HSBC while simultaneously providing connectivity that allows clients to use their own AI tools. That approach recognises that corporate customers are unlikely to operate within a single bank-controlled AI environment.
From APIs to agent-ready banking
Banking APIs were one of the foundations of open banking and embedded finance because they allowed financial services to move beyond the bank’s own applications.
AI may push that concept considerably further.
APIs allowed software to request banking services. Agent-ready infrastructure could allow intelligent systems to determine which banking services they require and when they should use them.
The difference sounds subtle, but economically it could be significant.
An AI treasury system might eventually identify surplus cash, evaluate liquidity requirements, assess currency exposure, compare available financial products and recommend an action without a treasurer manually moving between multiple systems. More advanced agents could potentially execute approved actions within predetermined limits.
Such a model would require considerably stronger governance than today’s conversational AI applications. Banks and their corporate clients would need clear rules around what an AI system can access, what it can recommend, what it can execute, when human approval is required and how every decision is recorded and audited.
This is why the development of secure, permissioned connectivity may ultimately matter more than the conversational interface itself.
Transaction banking’s next competition may be over orchestration
For years, competition in corporate digital banking centred on the quality of portals, APIs, payment capabilities and integration with treasury systems. AI adds another potential layer of competition: orchestration.
If AI agents increasingly determine how liquidity is allocated, which payments rail is used, when foreign exchange is executed or which financial product best satisfies a company’s requirements, the system controlling those decisions could become enormously influential.
BCG argues that as treasury becomes increasingly automated, transaction banks will need to protect their position in the corporate relationship by embedding intelligence within client workflows rather than remaining solely transaction processors.
That is the broader significance of HSBCnio.
The immediate product allows corporate clients to access banking services through several channels and gives authorised AI tools a controlled route into banking information. But underneath the product announcement is a recognition that the interface between banks and businesses is changing again.
Internet banking moved corporate finance from the branch to the browser. APIs moved banking capabilities into enterprise software. AI could move banking into an intelligent orchestration layer where machines increasingly retrieve information, interpret financial conditions and participate in decisions before a person ever opens a banking application.
HSBCnio suggests that large transaction banks are beginning to prepare for that environment.
What this means for the industry
- AI is becoming another banking channel. Banks increasingly need to design connectivity not only for people and conventional software, but also for authorised AI systems.
- Trusted data access will determine how useful treasury AI becomes. Secure access to live banking information could allow enterprise AI to move from generic analysis towards operational financial intelligence.
- Permissions and governance become critical infrastructure. As AI gains access to sensitive financial information, identity, authorisation, monitoring and auditability will become central to agentic banking architectures.
- Corporate banking interfaces could become less visible. Employees may increasingly interact with enterprise AI or treasury platforms while banks operate behind those environments.
- The strategic battle may move towards orchestration. The institution or platform influencing how AI evaluates and executes financial decisions could occupy a more valuable position than the provider that simply processes the resulting transaction.
Image Source: Pexels.com
