By Medan Gabbey, CRO, Quod Financial
(This is the third article in our multi-part series on industry themes, following Adapt to Survive – How Modular Deployment is Transforming Legacy Systems published on February 19, 2025.)
Goldman Sachs estimates that over the next five years Microsoft, Alphabet, Meta and Amazon will collectively invest over $1 trillion in AI. At the same time VC funding in AI startups reached $64.1 billion in 2024 alone.
With this kind of firepower being thrown about, few doubt that AI is going to have a significant impact in every industry and sector. And yet it’s surprising that trading firms have struggled to embrace what AI means for their Order Management Systems (OMS). Aside from the secretive HFT firms, success stories are thin on the ground. The number one reason cited for AI project failures is lack of a clearly defined goal. And coming in a close second is getting data into a format that the LLMs and Agents can go to work. And, as we all know, data in our industry is hugely fragmented and duplicated both in form and location. The simple fact is that AI is only as good as the data it feeds on.
This article looks at how OMSs need to change to become far more “data-driven” so that AI can really play its part. Overcoming this challenge is vital as a firm’s OMS is central to its business and helps frame its value proposition to the outside world.
What is a data-driven OMS?
The classic definition of data-driven refers to “making decisions, guiding actions, and
formulating strategies based on empirical evidence and analysis of relevant data”. In the OMS world this definition is more nuanced as data is spread over multiple systems and applications that historically have not worked in unison. Worse many of these platforms are legacies from a bygone era where developers never cared about what happened outside ”their” system.
So what is needed in a data driven OMS is a platform that can act as both a state engine and a single source of truth. This means close integrations with other systems so that the overall user experience can scroll smoothly. For sure, this data can then be loaded into an IMDG (In Memory Data Grid) so that system performance is never compromised but the data driven OMS acts like Central Station as nearly all data resides in it or flows through it.
Why is this important now?
The cycle for replacing OMS platforms is shrinking. In part this is because the huge leaps in technology mean that newer vendors have a distinct advantage. But also, the industry has changed since the legacy platforms were installed. Shrinking commissions, higher cost of business (due to regulation) and fragmented liquidity conspire to make the economics of the industry far more challenging. On top of this, clients are becoming ever more demanding from their broker networks. And, finally, the cost of failure in terms of regulatory sanction, reputational damage and hard cash is only going up.
Equally, the benefits of a data driven OMS can drive revenue through improving client service, understanding client behaviour and really understanding what is going on in your firm are considerable.
So, what does a data-driven OMS look like?
A data-driven OMS starts with the right data model. It must be inherently multi-asset — not simply because sell-side firms trade that way, but because their clients expect a unified cross-asset view. The OMS should be capable of ingesting and normalizing data from multiple sources and vendors, each typically using proprietary formats. By structuring data at the core, the OMS becomes the single source of truth, ensuring consistency across execution, compliance, and analytics workflows. Quod Financial’s data model is built for extensibility and integrity, allowing seamless adaptation to new asset classes, regulatory requirements, and client-specific workflows without major system overhauls.
Once the data foundation is in place, the next crucial element is a microservices-based architecture. In contrast to monolithic systems, where scaling is cumbersome and failures in one area can cascade across the platform, a microservices OMS runs as a series of independent, loosely coupled services—each representing a specific business function. This allows for independent scaling, high availability, and encourages technological diversity as deployment becomes far simpler, as you are not ripping out one monolith only to replace it with another.
Quod Financial’s architecture takes this further with a central messaging BUS, enabling real-time, in-memory data processing and integration with external solutions via high-performance APIs. Unlike legacy architectures, which struggle under peak loads or market spikes, Quod’s microservices model supports rapid scaling by simply adding more instances to the BUS, ensuring optimal performance in high-volume environments.
The third key feature of a data driven OMS is a robust and actively maintained API layer. The term API has become so ubiquitous that it masks a wide spectrum of implementations—many of which are inflexible and outdated. Legacy vendors often struggle here, as their architectures were originally built to lock clients into a single, closed ecosystem. In contrast, modern trading infrastructure is modular, relying on best-of-breed solutions that work together as seamlessly as the microservices inside the core OMS. A true data driven OMS must provide open, well-documented, and low-latency APIs that allow firms to integrate proprietary analytics,third-party algos, or alternative execution channels without friction.
All too often, techie rivalry or commercial jealousy gets in the way of multi-vendor platforms.Successful deployment of a data driven OMS needs cooperation between vendors and in-house resources against a single, shared goal.
The final ingredient has nothing to do with technology at all. It’s about the attitude of the vendor or vendors you are working with. Too often, techie rivalry or commercial rivalries create roadblocks to interoperability. Successful deployment of a data driven OMS needs cooperation between vendors and in-house resources against a single, shared goal.
Conclusion
Any firm that believes AI is in its future needs an OMS platform that can solve the data problem. A data driven OMS can be at the heart of this—provided it has a multi-asset, extensible data model, a microservices architecture, and rich APIs. And, of course, until AI is running the world, you’ll need a committed team of multi-vendor human resources aligned with the same goal as your internal teams.
