MIAMI —Securities services are being compressed on several fronts. Europe is preparing to shorten settlement from T+2 to T+1 in October 2027, artificial intelligence is moving deeper into operations, and tokenisation is creating new ways to issue and service assets. Each also changes the controls, funding and processes around custody. For Asia Pacific institutions investing into Europe, the settlement change is particularly operational. Markets in the region open several hours ahead of Europe, leaving less time to allocate trades, arrange foreign exchange (FX) and ensure the right currency is available. BNP Paribas has already managed T+1 transitions in India and the United States and is using that experience with clients preparing for Europe. Franck Dubois, Head of Asia Pacific for Securities Services at BNP Paribas, said readiness before the custodian reaches the final settlement step is critical. The broader question is how a securities services provider can help clients as settlement shortens, artificial intelligence (AI) enters business workflows and digital assets require new forms of cash and control. T+1 turns settlement into a cash and pre trade problem Dubois said the main pressure points in moving from T+2 to T+1 are cash management and FX. A shorter securities cycle means the required currency has to be available sooner, while many FX transactions still operate on a T+2 basis. The custodian therefore has to help clients prepare their cash and transaction processes before the settlement deadline arrives. The challenge is sharper for Asia Pacific investors because of the time zone difference with Europe. Dubois said institutions in Asia can be six or seven hours ahead, and Australia and New Zealand further still, leaving less of the Asian working day available for intervention. Cash is only part of the problem. Dubois also identified trade allocation and pre trade readiness as mandatory because there is less time to correct information after execution, making the client's operating model as important as the custodian's infrastructure. AI is moving from augmentation into the operating model Dubois described AI adoption in three levels. The first is individual and technology augmentation: large language models can assist employees, while developers are using AI to accelerate coding. He said this is already changing technology delivery, including giving internal IT teams greater capacity to develop applications that might previously have required external vendors. The second level is agentic AI inside the technology environment. Dubois pointed to reconciliations, net asset value calculation and transaction monitoring as processes where agents could improve efficiency while remaining part of the technical infrastructure. The third level is more consequential. Dubois said AI is beginning to become part of the business itself, with agents that could perform complex workflows in autonomy, handling the execution while people would bring the judgement. He characterised this as emerging rather than established. Business-level AI makes governance and returns harder to separate Moving AI into the business raises a different control question. Dubois said institutions need to address continuity, governance and control, including how an AI agent performing a defined role remains available independently of disruption in the underlying technology. He pointed to regulatory guidance, including from Australia, as an indication of supervisory expectations. BNP Paribas has also elevated AI organisationally. Dubois said Charles Olive, Global Head of AI for Corporate and Institutional Banking, reports directly to the head of CIB, reflecting the view that AI is becoming a business capability rather than remaining solely within technology. The economics will have to become equally explicit. Dubois said AI comes with a cost and institutions need to measure return on investment through efficiencies, new services and revenues. As deployment moves closer to business activity, control and measurable value become harder to separate. Tokenised securities still need a cash leg Digital assets are further along as an industry discussion but still face a fundamental infrastructure constraint. Dubois said BNP Paribas is involved in initiatives including bond issuance. BNP Paribas Asset Management has also issued tokenised money market fund shares. The unresolved issue, in his view, is cash. A securities transaction can be represented and processed on tokenised infrastructure, but an industrial model also requires a tokenised form of payment. Dubois said initiatives are progressing, but the market does not yet have industrial scale tokenised payment capacity. The cash leg is therefore central to whether tokenisation changes securities servicing at scale. Without it, digital securities still depend on settlement outside the tokenised environment. Faster infrastructure increases the importance of operating discipline T+1, AI and tokenisation are different developments, but Dubois's comments point to the same operating consequence: faster processing shifts work earlier rather than removing it. T+1 requires earlier FX, cash and trade readiness; AI agents require accountability; tokenised securities require cash infrastructure capable of settling alongside them. BNP Paribas's experience from earlier T+1 migrations provides a practical framework for the first change. AI and digital assets are less mature, and Dubois distinguished what is already being used from what is only beginning to emerge. The securities services proposition therefore becomes less about adopting each change in isolation than keeping the operating model aligned as processing accelerates. For Asia Pacific institutions, the value will depend on whether faster markets can be matched by earlier funding, better automation and controls that remain effective as the time available to intervene becomes shorter.
