Six key takeaways from SIFMA Ops 2025

By James Maxfield, Chief Product Officer.

Earlier this month, Operations leaders from across capital markets attended the annual SIFMA Operations conference in Orlando. There was a heavy focus on panels and speaker sessions, featuring contributors from across the industry, covering how to face the challenges of today and also future proofing yourself against what is coming tomorrow. This provided a high-end learning experience for all the attendees, on top of the chance to forge new connections, share best practices and discover new technology solutions.

I had the pleasure of talking to many delegates and speakers about the biggest barriers and opportunities facing Operations teams. Here are my six key takeaways from the sessions and my conversations.

1. Resilience, robustness and reconciliation

Whilst much of the agenda at SIFMA was forward looking, a very common theme was the lack of resilience within many organisations’ operating models. Heavy reliance on spreadsheets and manual processes, commonly driven by poor data quality or lack of standardisation across market participants, continues to be a major headache for Operations leaders.

At our Lunch & Learn panel on automating the middle office, more than 60 people crammed into the room to hear the panellists discuss the challenges they face in day-to-day processing. With the recent market turbulence – the VIX breached 60 in early April – as a backdrop to the discussion, it was very clear how real this challenge was across the industry.

With a high degree of dependency on people to plug processing gaps, this puts significant emphasis on Operations leaders to make constant judgement calls around prioritisation and risk management, eg. where to focus time. This oversight model doesn’t scale in stressed market events, as the panel discussion highlighted, creating a real need for organisations to focus on understanding how to bring practical, best-practice reconciliation automation to create robust, scalable control environments.

2. T+1 is not just for Europe – and why the US needs to care

The challenges of T+1 for North America may well be in the rearview mirror for many US organisations. But the global nature of business means the push in Europe towards T+1 in 2027 is going to be a significant lift.

Whilst the large global institutions will leverage their European arms to solve the problem for them, those US-centric businesses with even moderate international flow into impacted markets will have to think hard around how they support these flows.

One wealth management business I spoke to explained that, due to the comparatively low international volumes relative to the rest of their business, they can currently get away with manually working through exceptions created for issues such as missing instruction data. These processes are typically run T+1 post-batch, which enables inventory to be positioned, currencies to be aligned and instructions to be matched (manually if necessary, via portals). 

Losing this day will now create a fundamental rethink in the operating model. A common discussion topic was how to work around batch driven system dependencies. Even for primarily US-centric businesses, the second order impacts of the shift in Europe will be significant. And batch processing is not going anywhere fast for most of the market…

3. AI: Right place, right time

Unsurprisingly as a vendor at SIFMA, I was asked a lot of questions around Duco’s AI strategy and how it impacts our roadmap. Whilst showcasing a lot of our current and future innovation (which does use AI) on our stand with customers, I also challenged visitors around the need to use the right tool for the job. Though large language models (LLMs) are all the rage for the value they create, they may be taking a mallet to do the job of a hammer in some cases.

Small language models have a place here, certainly where the use case is quite narrow and doesn’t need internet-wide scale – for example, where data is highly sensitive, such as personally identifiable information (PII). And, as importantly, the process can be run locally. Certain know your customer (KYC) use cases may benefit here, where the robustness and security aspects may outweigh the value of a much more comprehensive pool of data to scan across. 

There is a real demand for innovation from across the industry and the evolution of AI has shown real transformational capability. But, to repeat a much used cliche, technology has to be an enabler of transformation. It is not transformational in itself.

The ability of Operations leaders to balance their usage of the right AI tools, within the context of the use case they are trying to solve for, will ultimately define the value they create. And a vendor’s role has to be increasingly advisory around how they can bring value through AI, not just promote it as a new shiny toy.

4. Agents, agents everywhere…

With agentic technology everywhere – in our personal as well as our business lives – it was no surprise that this was a hot topic across the conference. Agents are AI programs that can autonomously fulfil goals through learning and taking action. Low-end use cases around administration, such as taking meeting notes, were clearly adding value to a lot of Operations leaders. (One commented on how this has reduced the amount of time spent at the start of the meeting arguing over who takes minutes…)

But other areas of value were less clear. Speeding up daily tasks such as exception investigation, that enable the human to take action faster, were common in a few use cases. However, having an agent do something like engage with trading desks over corrective actions was seen as overly ambitious.

In addition, the choice of in-house vs vendor agents was also widely debated. It certainly depended on the use case. The interoperability of agents to interact across platforms is going to be critical to unlocking value across a wider spectrum of use cases.

This will give choice – albeit securely permissioned – for users to determine why and how different agents are inserted into their processing value chain. And as importantly, what action they can take, and when.

5. The legacy technology challenge is still very real

The innovation experience for many organisations has improved significantly over the last decade, as newer technology has provided the tools to crack some of the stubborn use cases that historically had been very people-heavy. But, at its core, many architectures are still heavily reliant on aged technology that is deemed irreplaceable. (Or, more specifically, the risk of trying to replatform is deemed too high).

This was brought into stark contrast by one large, global consultancy I spoke to who talked about the rapid increase in demand they had seen for a growth area of their business. Namely, common business oriented language (COBOL) developers. Yes, that’s right – COBOL. The 60+ year-old programming language that most places haven’t taught for decades.

They had now trained over 300 of their engineers in how to code in this environment and these resources were heavily in-demand across various clients. With the engineers who built these applications now retiring, there is a shortage of these skills to both maintain these platforms and also try to enhance them.

Whilst CEOs across the globe are showcasing their cloud-powered, API-enabled business models, the reality on the ground is very different. For many Operations leaders, they are having to divert a significant amount of their budgets to run and change these 30+ year-old platforms. And at the same time as trying to embrace new innovation to automate and enhance their operating environments.

This is tough to architect and, in the absence of significant investment (or a partner who has appetite to fund this transformation), this situation is not going anywhere fast. Success here will require smart decision making around how to effectively renovate – not rebuild – post trade architectures to benefit from modern innovation, but also be practical around legacy domain areas such as integration and data accessibility.

6. Mind the (knowledge) gap

The employee experience and the challenges of retaining key talent have been a common discussion topic over the last 18 months.

It was no different at SIFMA, with several panels debating the challenge of giving freedom to Operations teams to innovate with new technology tools, whilst avoiding a ‘Frankenstein’s monster’ of poorly governed and bespoke processes. These limit resilience and eventually cause more problems than they solve for, where loss of knowledge retention becomes compounded by dependencies on ill-defined technology solutions.

Success here is balancing freedom to innovate – which is desperately needed – with providing career paths for middle management to evolve leadership skills that spread technology deployment, process standardisation/governance and capital markets domain expertise. Senior management are often having to play a top-down/gatekeeper role here, which is a by-product of a gap in this middle management layer.

Investing deliberately in this middle management layer not only creates an interesting career path, it also provides much needed organisational depth to solve for what is an increasingly complex span of control. The leaders of tomorrow will need to balance solving day to day challenges within an increasingly complex technology landscape, requiring ‘new world’ automation skills alongside ‘old world’ Operations skills around risk and control. Exciting opportunities indeed, but will need deliberate human talent management to nurture the right skills to succeed.

Summary

With the macro-economic landscape creating volatility and uncertainty, Operations leaders are likely to have their hands full in just keeping the lights on and maintaining control. Being able to address the challenges of today whilst thinking about the solutions of tomorrow is a tough ask, which can make leadership a lonely place to be.

Success here is around federating the load, through collaborating with partners to enable a best of breed approach. No-one has all the answers, from what I observed, which only reinforces the value of partnering with others to bring parts of the jigsaw puzzle together.

Why not talk to us about how we can help?