Jayna Bundy, Treasurer at Microsoft, shared the tech giant’s agentic AI use case at Sibos in Frankfurt last year when she explained how the treasury team are transforming their front-end payments experience and overall corporate treasury operations with a new agentic AI tool.
As part of the debate about what banks, technologists and treasuries can each learn from the other to deliver better payment and data insights and enhanced cash management, risk and liquidity operations, Bundy described how Microsoft has classified news in a data lake as low-, medium- or high-risk impactful.
“For example, regarding foreign exchange (FX) exposures in say Argentina, our new agentic AI tool can make recommendations [to mitigate risk],” explained Bundy, adding that this means she doesn’t immediately have to contact all her team to find out the risk of any market movements before taking any mitigation measures. “We have this agentic AI tool live now and will continue to evolve it.”
The treasury’s agentic AI tool at Microsoft can also make other recommendations, such as:
-
Cash management: ie it might suggest moving money from one account to another to avoid an overdraft and any penalties.
-
Interrogate the status of a payment: ie so that exceptions handling can be minimised and straight-through processing (STP) efficiency maximised on payment flows. It might also provide further down-the-line liquidity optimisation tips.
“Of course, treasurers will still make the ultimate decisions,” said Bundy, as she sought to reassure any audience concerns about governance, control and oversight.
“The tool could also be used to offload low-value tasks,” she added – for example, cash management automation end uses. “I’m super excited about the future of agentic AI.”
Shaping the future
“Digital transformation is key for Microsoft,” added Bundy, “so we’re also very interested in stablecoins, digital assets, and other things that will shape the future.”
“I encourage my treasury team to think about what the treasury of the future will look like,” she continued. “I want my treasury team to stretch our bank and payment providers and ask them what is next? How can you drive the finance industry forward and effect change through innovation?”
Banks and tech vendors share their tips
“AI can also be used on smart routing payment end uses and on anti-fraud and money laundering (AML) applications,” said fellow Sibos panellist, Carmen Rey, Head of Swift and International Payments at PagoNxt, a Santander fintech company that operates in 30 countries on 100+ currencies, handling two billion transactions and speaking alongside Bundy.
As the conversation widened away from AI towards other useful developments in the marketplace that may impact treasury in 2026, Andrew Blair, Head of Global Solutions at Kyriba, highlighted the importance of open application programming interfaces (APIs) as a means of achieving easier data exchange and connectivity. “APIs make things more seamless, payments frictionless, and they can help pre-validation anti-fraud checks and real-time connectivity as well,” he said.
24/7 availability on proliferating real-time payment platforms around the world is a key win for corporate treasurers seeking faster processing times, cash recognition, easy transactions across borders via Swift gpi and other such mechanisms, so anything that makes linking to these capabilities – as open APIs do – is welcome.
Additionally, stipulations, such as the EU Instant Payments Regulation (IPR), are buttressing the technological drive towards real-time capabilities.
Commenting on the Sibos event organiser’s announcement of a new Swift blockchain-based ledger, Kyriba’s Blair added that this should also allay some of Bundy’s fears about cross-border timezones and Microsoft possibly missing cut-off times for payment processing, with all the down-line liquidity, investment and other impacts that may have. Blockchain developments will also be a big tech trend for 2026.
Treasury’s fundamental requirements
Microsoft’s treasurer shared the fundamental requirements that she looks for in a payment processing partner. Outside of the AI driver that is evident across all operations for all financiers, Bundy must have:
“We need to know that our processors can execute and inform us about payments quickly and give us associated liquidity information,” said Bundy. “Confirmation and end-to-end tracking are super key too.”
The non-negotiable fundamentals of good payments processing and treasury operations, as listed above, don’t change despite the bells and whistles that new AI and distributed ledger technology (DLT) platforms might bring to the game.
Fellow Sibos panellist, Yoichiro Ushiro, in charge of cash management, trade finance and supply chain finance products at MUFG’s Transaction Banking Division, shared another fundamental – namely, clean data. “We need clean data – and beneficiary banks to be up-to-speed – to deliver yet more capabilities. It’s definitely progressing.”
It will be interesting to see how much the industry has progressed at trade shows in 2026 when many of these eternal topics of good payments processing practices, useful technology and data supports, plus efficient treasury operations will no doubt be to the fore once more.
