Transformation & Innovation

The clearest message emerging from SIFMA Ops 2026 was not simply that financial services is changing. It was that firms now need to prove they can operate differently.

AI, tokenization, accelerated settlement, extended trading hours, Treasury clearing, data modernization and interoperability are no longer separate transformation tracks. They are beginning to converge around the same operational question: are today’s operating models ready for the market structure now taking shape?

That is the context for the operating model now emerging across the industry. The observations below are informed by discussions across SIFMA Ops 2026, where the industry conversation was less about long-term innovation and more about practical execution. The most important takeaway may have been that firms are no longer talking primarily about systems modernization. They are talking about operating model redesign. One speaker summarized it succinctly: “Start with the operating model.” Firms cannot solve next-generation operational challenges simply by layering new technologies onto fragmented workflows, disconnected teams, and inconsistent data structures.

Firms are now walking a narrow line: moving too slowly creates franchise risk, but moving too quickly creates operational and control risk. That tension sits at the center of the next generation operating model.

The operating model became the real story

For years, operational transformation was often discussed through the lens of platforms, automation, and cost efficiency. Those issues still matter. But the agenda has moved on.

The industry is now being pushed to rethink the assumptions underneath post-trade operations: how work moves, how exceptions are managed, how data is governed, and how quickly firms can respond when markets move faster than their processes were designed to support.

That shift matters because the pressure is coming from multiple directions at once. T+1, extended trading hours, inventory visibility, corporate actions modernization, and increasingly real-time markets are all becoming part of the same operational conversation.

The Broadridge 2026 Digital Transformation & Next-Gen Technology Study reinforces this point. Eighty-four percent of firms now believe unified front-, middle- and back-office platforms are important to support emerging technologies and operating models. Firms are moving away from thinking about operations as isolated functions and toward orchestrated workflows built around shared visibility, unified data, and interoperable infrastructure.

Much of this transformation is being forced indirectly. AI requires cleaner data. T+1 requires faster coordination. Tokenization requires interoperability. Real-time markets require operational visibility. Different pressures are all pushing firms toward the same destination: a more unified and continuously adaptive operating model.

The tokenization conversation finally changed

Broadridge’s study shows that more than half of firms now believe blockchain and distributed ledger technologies will create new capital markets opportunities, while 54% are already making moderate to large investments in tokenization and digital asset infrastructure.

Tokenization is no longer being discussed primarily as a future-state innovation story. The conversation has become more practical, with the focus shifting toward collateral mobility, liquidity optimization, repo efficiency, interoperability, and settlement workflows – with tokenization is increasingly being treated as an enabler of market evolution. The important nuance is that the market is not looking for replacement narratives. The more credible path is extending existing infrastructure into new operating models.

That distinction reflects where the market stands today. Collateral mobility and repo remain among the strongest near-term tokenization use cases because the economics are easier to understand, the operational friction already exists, and the efficiency gains can be measured. Firms can connect the technology directly to liquidity and funding efficiency.

As one speaker put it, the current phase of market evolution is “less about creating digital assets and more about redesigning how assets move.” Tokenization is not just about changing the form of an asset. It is about changing the operating model around that asset.

Interoperability is becoming central to that discussion. Firms cannot operate parallel environments indefinitely. Tokenized assets still need to connect into existing post-trade workflows, settlement infrastructure, corporate actions processing, reporting, and governance frameworks. That may become one of the defining operational questions of the next several years.

The real AI story is happening inside operations

If tokenization is changing market structure, AI is changing operational execution.

The AI conversation has also matured. It is no longer centered on pilots and isolated use cases. The focus is now on orchestration, exception management, and operational deployment at scale.

AI is shifting from isolated use cases to the center of the enterprise operating model. That is important because the strongest early use cases are not always the most visible ones. Much of the real work is happening inside operations, in areas such as workflow triage, breaks management, reconciliations, KYC processing, inventory management, and settlement coordination.

One of the more revealing points raised during the event was how much time firms still spend trying to reduce operational noise. Teams are dealing with alerts, queues, emails, escalations, and fragmented workflows. AI is increasingly being used to help prioritize what actually matters rather than simply automate individual tasks.

The operating principle is becoming clearer: automate the normal and elevate the exceptions. The value is greatest in exception-heavy workflows where teams struggle to scale under compressed timelines. AI can help narrow the field of human attention toward the activity that genuinely requires intervention, judgment, or escalation.

As one operations executive put it: “The edge cases are where you still need human judgment.”

The Broadridge study shows how quickly adoption has accelerated. AI usage across operations and processes jumped from 31% to 80% year-over-year, while 27% of firms say they are already realizing financial benefits from their GenAI investments.

At the same time, firms remain highly focused on explainability, governance, permissions, and supervisory controls. The focus has shifted from experimentation toward operational outcomes.

The industry is running out of operational slack

As one panelist put it: “T+1 didn’t remove friction. It compressed it.”

The industry is not just dealing with shorter settlement cycles. It is dealing with compression across decision-making, workflows, and risk management that historically relied on overnight windows and manual intervention.

Many of the industry’s existing operating assumptions were built around having time as a buffer. Overnight processing windows acted as shock absorbers. Batch processing created breathing room for breaks resolution and reconciliations.

The pressure is also expanding beyond T+1. Extended trading hours, rising message volumes, market fragmentation, Treasury clearing changes, and expectations for real-time visibility are all testing whether firms can operate safely when there is less room to recover from errors.

European T+1 implementation makes the challenge even clearer. Unlike the U.S. migration, Europe introduces additional layers of complexity through multiple central securities depositories, currencies, regulatory structures. and local market conventions. The operational burden is not disappearing. In many cases, it is moving earlier into the lifecycle.

Inventory visibility is one example of a basic but persistent challenge. Firms still need better visibility into where assets sit, which depositories hold inventory, and how quickly collateral or securities can move under compressed timelines.

Extended trading hours add another layer of complexity. Continuous or near-continuous trading puts significant pressure on infrastructure still designed around batch-era assumptions. One participant described the problem bluntly: “If you have continuous trading but not continuous settlement, that creates a problem.”

Operational resilience is increasingly becoming a question of whether firms can maintain control as conditions become more compressed.

Data is still the hardest part

Behind almost every major topic discussed at SIFMA Ops sat the same underlying issue: data.

AI depends on clean, accessible and governed data. Tokenization depends on interoperability between traditional and digital records. T+1 depends on accurate settlement instructions, inventory visibility, and workflow timing.

The challenge is not that firms do not understand the importance of data. It is that many still operate with fragmented data models, inconsistent definitions, and disconnected platforms.

One executive described discovering twelve separate Microsoft environments and multiple disconnected CRM and HR systems during what initially seemed to be a straightforward AI integration initiative. The point was not technological complexity as much as operational fragmentation.

This is why data modernization is no longer a back-office clean-up project. It is becoming a prerequisite for AI, tokenization, real-time operations and operating model redesign.

As one speaker warned, “Without a common data model, every AI use case becomes a separate build.”

Firms that continue building on fragmented data foundations will struggle to scale AI effectively, integrate tokenized workflows, and manage increasingly real-time operational environments.

What leading firms are doing differently

A clear pattern is emerging: firms making the most progress are not treating AI, tokenization, and operational modernization as separate efforts.

They are prioritizing a smaller number of foundational capabilities that support multiple strategic goals at once: investing in unified data environments, reducing workflow fragmentation, and building interoperability into operating infrastructure from the outset rather than treating it as a later integration challenge.

Importantly, many firms are also becoming more selective about where automation creates the most value by shifting away from isolated productivity gains and toward areas where operational bottlenecks create pressure across the organization.

Several speakers also emphasized the importance of sequencing. Rather than pursuing wholesale transformation, firms are modernizing high-friction operational areas to produce immediate efficiency gains while also supporting longer-term operating model flexibility.

That may ultimately become one of the defining characteristics of the next-generation operating model: the ability to evolve continuously without adding complexity.

What the next operating model requires

The next operating model is not about choosing between AI, tokenization, automation, real-time processing, or platform modernization. It is about understanding that these changes are connected.

The firms best positioned for the next phase of market evolution will be those that can bring several capabilities together at once: interoperable infrastructure, unified data, intelligent workflow orchestration, modern exception management, and teams equipped to operate in a more continuous environment.

That does not mean replacing everything at once. One of the clearest lessons emerging from tokenization is that the market will move fastest when new capabilities build upon what already works. The same is true across operations: firms need modernization that delivers value now while preparing for what comes next.

That is why operations is now strategic. It is no longer just support infrastructure. It is a source of resilience and competitive advantage.

The winners will not be the firms with the most pilots, the most experiments, or even the most technology. They will be the firms that build operating models capable of adapting as market structure continues to evolve.