Cash & Liquidity Management

Q&A: future of banking and treasury is hyperconnected

Published: Nov 2026

BNY’s Bana Akkad Azhari, Head of EMEA Global Payments & Trade, joined Treasury Today Group to discuss AI’s impact on financial services and corporates, the rise of faster finance and liquidity and how a coalescence of technologies can improve customer experiences.

Connected traffic system.

Bana Akkad Azhari

Head of EMEA Global Payments & Trade
BNY

At this year’s Sibos event, Treasury Today Group’s CEO Sophie Jackson sat down with Bana Akkad Azhari, BNY’s London-based Head of EMEA Global Payments & Trade, to discuss her key show takeaways and the hyperconnected future of banks and their corporate treasury clients.

As on-chain finance, assets and tools like tokens and stablecoins emerge thanks to distributed ledger technology (DLT), alongside 24/7 traditional (TradFi) real-time payment (RTP) platforms that can bilaterally link or connect via Swift, the speed and ease of global cross-border money and capital movement is increasing. This feeds into faster liquidity and a more data-centric future, which is increasingly corralled and automated by artificial intelligence (AI) models.

During the event, the value of embedded finance, enhanced automation and hyperconnectivity – aided by the concurrent rise of open application programming interfaces (APIs) as an easier means of exchanging data – was evident to BNY’s Azhari.

What is your view of AI and its impact on your systems and those of your clients?

AI has certainly been a crucial area of technological innovation. I think it’s very important and provides a lot of opportunities for FS and for our clients. It’s been a hot topic in Miami.

Our CEO was one of the early AI adopters. He wants to see BNY use AI for the good of the institution internally and externally for clients and partners. BNY’s mantra for our own AI platform is to make it available internally ‘for everyone, everywhere and on everything’. It really will be that transformative.

Early on, BNY democratised the use of AI across our workforce so that we could try it out and innovate together. We’ve done that and learnt lessons. For example, it can help pull together data and presentations before a client meeting, almost like a super assistant.

We took a decision early on to embed the technology across the organisation, as we don’t want AI to be an elite type of activity. It’s for everyone, so we developed our own AI platform in a way that preserves and protects our client data and privacy. This is critical given our position in the financial ecosystem, while still delivering organisation-wide data availability, automation and fast interrogation and data delivery capabilities.

How we use the technology can be modular depending on use case. As our chief data and AI officer said: ‘it’s horses for courses’ AI is for automating processes, getting efficiencies, and serving our internal staff and external clients better. It allows us to derive actionable insights that enable us to make better decisions and offer better services in the payments, treasury and operations spaces.

What AI end uses do you see and which control mechanisms do you deploy?

AI is fast moving towards multiple practical end uses now. These include activities like exceptions handling, payment repairs, fraud prevention, risk monitoring or optimisation duties, and better client onboarding experience.

These are key AI end uses. Obviously, having good governance, buildout and operational guardrails is essential, as is having a human in the loop to ensure that AI models don’t go rogue. We made it a requirement for our people to take an AI ethics course before getting access to the platform. We do all this to ensure safety and maintain bank-grade trust.

Which key trends have stood out for you at Sibos 2026?

For me it’s the coalescence of different pieces of technology coming together to deliver better services and enhanced customer experience (CX) that is the key takeaway. For example, AI is allowing us to effectively use the large data lakes and volumes brought about by the earlier digitisation of trade, payments and assets. Allied to the structured data that banks already possess, this opens up huge new opportunities to get better data faster and in a more easily digestible form.

At the generative end of AI, that means better content for a staff member’s client meeting, for instance. At the agentic AI end of the spectrum, it means better training for the autonomous or semi-autonomous software systems that are powered by large language models (LLMs). These can set goals, make decisions, use external tools and execute multi-step workflows, with limited human supervision, to improve efficiency. Machine learning applications of the technology continue to offer enhanced automation potential as well, so AI is really coming to the fore.

The rise of TradFi T+1 settlement capabilities on capital markets – and of DLT-based finance, assets and tools like tokenisation – are all adding to the trend towards faster finance and powering the data deluge we’re seeing. Most immediately, the speed and 24/7 always-on nature of processing engines is feeding the need for speed. These trends can avoid cut-off times and enhance liquidity in payments and investments, aiding corporate banks and treasuries across the entirety of their operations.

The rise of enhanced real-time payment (RTP) and liquidity platforms, such as the RTR in Canada or FedNow in the US, is adding to the speed, 24/7 and enhanced data possibilities available to banks and businesses. Such platforms also tend to use more character-rich ISO 20022 messaging as well, which adds to the data deluge that banks and treasuries must be able to digest.

AI tools can make this 24/7 data more contextual, relevant and useful for dynamic cash flow forecasting (CFF), for instance, as well as enhanced liquidity management and numerous other end uses. All these things are coalescing to deliver better services.

There is a broader shift in the industry towards a more connected and always-on financial ecosystem with open APIs or blockchains driving greater connectivity and data exchange, alongside ISO 20022-enabled instant payment platforms.

Do you fear the practical implementation of FS tools will suffer if people focus too much on the technological possibilities of AI, DLT and the coming data deluge, without first finding real-world end uses?

Yes, that is fair. You do have to put technology to a good end use. Aligning people, processes and technology to deliver a better service is what matters. But I believe we’re more than capable of doing that – look at the AI super assistant tools I mentioned earlier that help our staff prepare for client meetings.

Collectively, the industry is also coming together to deliver anonymised federated data to spot suspicious activity patterns in the fight against fraud, while still respecting privacy. Deploying AI here is super helpful. After all, fraudsters use it and share target data on the dark web themselves. Bankers should respond accordingly. That is a practical end-use example.

I was pleased to see at Sibos that there was less talk about innovation for the sake of innovation. Instead, the focus in Miami was more on actual financial processes, the movement of value, efficiency, and getting better CX and compliance. This is what matters and what makes technical innovation practical.

Interoperability among market players across different asset classes, payment types, traditional or decentralised finance (DeFi) DLT-based platforms is vital in delivering widespread uptake and connected services, which equates to good, fast data-centric CX. The trust element comes from better compliance and bank-grade cyber security.

We need to work together to provide interoperability – Swift itself shows what can be achieved when we do. In this era of enhanced digitalisation and cyber threats, amidst a geopolitically unstable environment, the focus on resilience, trust, security and stability must be very strong as well. Again, federated data sharing to fight fraudsters is one instance of this trust element in action.

You’ve touched on real-time payment platforms and capabilities. But what does that mean for intraday liquidity, cut-off times and end-of-day positions for corporate treasurers?

Real-time cash visibility is what is essential, not just processing the payment in real-time. You need to be able to digest and use it. Banks can help with this.

For example, extra speed is only useful if a treasury can link it to its FX risk mitigation and sweeping or cash concentration procedures to get the best liquidity position. Ditto with aligning it to cash flow forecasting or dynamic trade finance, which can action payments as a ship traverses the globe.

It’s what you do with real-time 24/7 capabilities, which might emanate from an instant payments platform or an always-on DLT-based infrastructure, that actually matters – not the connectivity in and by itself. How do you manage funding, position and collateral management in certain instances, especially when agentic AI commerce and supply chains with conditional payments advance?

Being on top of your risk and liquidity management in the faster, more data-centric environment of today is essential. Transactions or milestones will increasingly be automated and linked to accounts, so think about the end-to-end processes and how to optimise your operations accordingly. This is my key advice.

It’s not about the payment itself anymore, but rather about the data it carries and how it converges and impacts a broader chain of value. Managing counterparty exposures better is a part of this too.

Transparency and visibility are super important as well. AI can help with this, in cash flow forecasting and many other end-use applications. It can also help me – and treasurers – get the optimal eight hours of sleep by advancing automation!

Join our global community

Creating a free account helps us to understand our community better, and tailor our content and events to suit your needs. You can unsubscribe at any time.

Already have an account? Sign in

Search for your company; if not found, select 'Other' and enter it manually below
  • • At least 8 characters
  • • At least 1 uppercase letter (A-Z)
  • • At least 1 lowercase letter (a-z)
  • • At least 1 digit (0-9)