Citi Securities Services Evolution 2025

Securities Services Evolution 2025 | 33 processing volumes by helping to manage only inconsistencies is an elementary but essential requirement for today ’s regulated institutions. In a world where faster, cleaner onboarding literally means money (in the capacity to trade and to avoid regulatory penalties), this use case appears to be a perfect starting point and an opportunity to bridge the gap between retail and institutional clients (where onboarding speeds vary enormously today). Institutional investors see the potential Ahead of banks and brokers, it is the buy- side (most of all pension funds, insurers and sovereign funds) that appear to be most attuned to the potential of Gen AI, with 14% of institutional investors surveyed seeing GenAI as the most impactful post-trade change for 2025. Importantly, these firms are looking to take GenAI into their back offices, with up to 67% piloting the use of the technology in trade processing (i.e. clearing and settlement) and in reconciliations. In the face of increasing trade cycle pressures (including T+1, 24/5 clearing, etc.) the world’s institutional investors are clearly looking to GenAI to provide a bridge into a new era of operational efficiency and automation. And what are the obstacles? There is a distinct divide between the challenges faced by the sell side and buy sides when it comes to GenAI adoption today, with the former beset by challenges around regulation and the latter struggling more to hire and manage the talent required to scale its usage. Regulation around Artificial Intelligence (AI) is fragmented and in its infancy still. While the EU has its AI Act, the UK adopts a principles-based method, Singapore has issued strong guidance, and other jurisdictions are in consultation stages. Such a globally diverse regulatory landscape creates complex compliance challenges for all regulated firms, with a lack of underlying consistency that can support advanced levels of adoption. Without this consistency, global firms such as brokers, banks and custodians continue to face hurdles in using GenAI for anything more than heavily supervised, process-level use cases which according to our survey, account for 56% of all GenAI use today. Meanwhile, resource limitations pose an ongoing challenge for the buy side, primarily due to the scarcity of AI talent . Despite clearly seeing the potential for GenAI, the world’s asset managers and institutional investors are struggling to build the requisite skills sets to be able to use this technology in a live context . With over 70% of the world’s asset owners already undertaking significant transformation projects 32 in the middle and back offices, change-management talent is scarce and thinly spread for most major investors. Add to those challenges in data access and connectivity, which are felt to be a challenge by more than half the respondents in every segment , and you have a technology that is still constrained by integration, data quality and available skills sets. Indeed, Gen-AI models are models are largely determined by the data they consume, yet financial data is complex, unstructured, inconsistent , and fragmented across different in-house and external systems. Figure 18: Obstacles to deploying GenAI Question: What are the top3obstacles todeployingGenerativeAI in your business today? Expressed as: %of respondents citing each obstacle as a top 3, per segment. Buy-side Sell-side Regulation People Policy/Regulatory challenges Internal sign offs Lack of internal sponsorship IT infrastructure challenges Capacity Limitations Skill sets Data connectivity and access 40% 46% 20% 40% 46% 54% 54% 64% 64% 12% 38% 21% 48% 52%