Andrew, it has been over a year since we last spoke. As predicted, it’s been a busy and disruptive year. AI has moved quickly from experimentation into practical workflows across capital markets. How has that changed the way you think about oneZero’s platform and the next generation of its products?
The past year has reinforced for me that trading technology for capital markets has entered a new paradigm with a new set of opportunities and challenges. Sustained volatility means performance and resilience remain fundamental, but AI is changing how people access information, investigate problems and interact with trading systems. This is not just an external input. The rapid evolution of AI models is only one part of that equation. What proving to be paramount in practice is the quality of the data, the context provided to the model, the interfaces around it and the controls governing what it can do.
We are seeing these trends emerge in every portion of our business, from our role as a platform provider with our Hub technology to the work that followed the integration of Autochartist into oneZero. Market analytics and engagement capabilities, combined with our infrastructure, data and global support, have developed into a broader Engagement product line as well as a clearly defined interface for our clients to adapt our products. We are bringing relevant market insight closer and closer to live execution and operational workflows. The opportunities that new tools are giving our team, and feedback from clients, are pushing us (and capital markets tech in general) towards connecting more of the journey from price formation and execution through risk management and client engagement.
Lastly, from a broader perspective, market counterparties are under pressure to differentiate while managing greater operational complexity. They do not solely need another isolated model, tool or dataset. They need an environment in which infrastructure, data and intelligence work together while the bank or broker remains in control. That is how I think about the next generation of our platform: AI becomes another way to interact with the trading environment, not a substitute for the proven systems at its core. It allows our customers who traditionally were not capable of building out their own tech to differentiate through new tools.
At the start of 2026, you identified persistent volatility, AI-enabled productivity and the growing utility of digital assets as three defining themes. To what extent has AI become THE connective thread among them, and how are these forces further shaping oneZero’s strategy?
The themes are closely connected because each one raises the standard clients should expect from their technology partners. AI does not make the established requirements of capital markets technology excellence less important. It increases the value of dependable infrastructure, real-time data and clear operating controls.
At the same time, volatility is no longer an occasional stress event. It has become a persistent feature of the operating environment, driven by policy shifts, geopolitical uncertainty and changes in market liquidity. During recent periods of exceptional activity, oneZero processed record volumes while maintaining greater than 99.99% uptime. That reflects years of investment in our Hub technology, a robust global EcoSystem and capacity management. The practical test is whether the market itself can continue to price, execute, analyse activity and manage risk when assumptions about liquidity, correlations and customer behaviour are being challenged at once.
AI is key here, as it changes how people work with the data produced by that environment. Through Data Source, oneZero captures and curates quote, order and trade data so clients can analyse liquidity relationships, simulate alternative configurations and feed insight back into pricing, routing and hedging decisions. AI can make that information accessible to more users and shorten the path from a question to a useful analysis. It can also help developers and operating teams investigate issues, test ideas and build tools more quickly. None of that works reliably if the underlying data is incomplete, delayed or stripped of context.
Digital assets bring the same discipline into a different market structure. Interest is moving towards institutional utility, particularly in stablecoins, settlement and cross-border funding, but those use cases still have to prove they can scale and create lasting value. oneZero’s liquidity-neutral, multi-asset architecture allows clients to incorporate digital assets alongside traditional FX while applying consistent controls to connectivity, pricing, execution and risk. AI may help firms interpret activity and test ideas, but it cannot manufacture a commercial use case or replace sound market infrastructure.
Across all three themes, the requirement is consistent: resilient systems, curated real-time data and technology that can adapt without giving up control. That foundation determines whether AI produces durable value or simply accelerates uncertainty.
oneZero has always been known for high-performance trading infrastructure. Where can AI add value today, and where do deterministic, low-latency systems evolve from here?
I would not frame AI as making latency or deterministic behaviour less important. Pricing, order routing and risk transfer sit at the core of mission-critical systems where decisions may be made in microseconds. Clients need predictable behaviour, throughput, resilience and the ability to understand exactly how those systems will respond. Current AI models do not remove any of those requirements.
The immediate opportunity is around that core. AI can help teams interrogate trading data, identify patterns, investigate operational issues, test configurations and customise the way users interact with systems. It can also improve the inputs to a decision. Clients want to know whether an order reached the right liquidity, whether pricing reflected current conditions and whether the transaction improved the economics of the overall relationship. That requires fill probability, market impact, profitability and risk to be viewed together.
This is where data and analytics become part of the execution loop. oneZero brings together quote, order and trade data so clients can examine behaviour, assess liquidity configurations and refine pricing, routing and hedging strategies. Examples like our Spot Rate Manager and Swap Curve Manager apply the same principle in live workflows by giving traders real-time visibility and control over market conditions, anomalies, spreads and skews.
As access to increasingly capable AI models becomes broadly available, what separates an AI-ready trading architecture from a conventional one? How important are APIs, user interfaces and liquidity connectivity?
The model itself is unlikely to be the lasting differentiator. Firms will be able to choose among models as the technology develops. The harder work is preparing the environment around them: clean and observable data, well-defined interfaces, timely access, appropriate permissions and enough context for a system to understand the task it has been given. In capital markets, those elements must also connect to infrastructure that can perform consistently under extreme conditions.
There is no single institutional operating model. Some firms want an end-to-end managed environment, while others have invested heavily in proprietary technology and need specific capabilities that fit an existing stack. Some workflows belong in a low-latency API; others need a GUI where a trader, risk manager or operations team can see what is happening and intervene. Agentic tools will add another interface, which makes controlled access and consistent permissions more important.
oneZero allows banks and brokers to combine pricing, execution, risk management, analytics and engagement around those business needs. They can choose liquidity relationships, deploy proprietary algorithms through programmable components and integrate with established customer workflows. Standardised connectivity and distribution reduce the amount of complex plumbing each firm has to recreate. The result is an architecture that can support new forms of interaction without forcing a disruptive replacement of the systems that already work.
Digital assets remain one of your defining themes for 2026. Does AI change the institutional case for the asset class, or do practical utility and infrastructure still come first?
Utility and infrastructure still come first. I have always viewed digital assets through the lens of market infrastructure and the problem a technology is actually solving. Clients still need connectivity, aggregation, pricing, execution, analytics and risk management, but those capabilities must reflect different liquidity patterns, trading hours and operational risks. The market operates continuously, liquidity can be fragmented, and relationships among venues, providers and regulated institutions are still developing.
A common technology framework is valuable because most institutions do not want a separate operating model for every asset class. Separate stacks increase cost, fragment oversight and make consistent risk controls harder to apply. At the same time, a common framework cannot pretend that digital assets behave exactly like spot FX. oneZero’s multi-asset, liquidity-neutral architecture accommodates those differences while giving the institution a coherent view of connectivity, price formation, routing and risk.
AI can help an institution interrogate fragmented datasets, compare liquidity behaviour and test how different configurations might have performed. That is useful, but it does not answer the more fundamental question of utility. The most interesting long-term opportunities may extend beyond trading individual tokens. Stablecoins, more efficient settlement and cross-border funding could address frictions that already exist in financial markets. We are also seeing more interaction between crypto-native firms and regulated institutions, which suggests digital assets will increasingly borrow from established market structure rather than develop in isolation.
This market still requires discipline. Institutional adoption will depend on transparent pricing, reliable infrastructure, effective controls and use cases that produce tangible value. Our role is to give clients the flexibility to participate as those use cases develop, without asking them to lower the standards they apply to their traditional trading businesses.
Where is oneZero investing to make banks and brokers ready for AI-enabled workflows, and what should clients expect from the next generation of your products and services?
I think about the investment in layers, with resilience at the foundation. Persistent volatility means clients need consistent performance across the full range of market conditions, not simply an attractive latency result during a quiet period. We continue to invest in capacity, stability, security and the operational visibility required by mission-critical trading businesses. AI can improve how teams interact with that environment, but the underlying platform must remain dependable before any intelligence is added.
The next layer is data and analytics. Data Source helps clients understand liquidity, customer behaviour, execution and profitability. It makes Hub quote, trade and derived data available in real time, near real time and historically, giving firms a consistent dataset for internal analysis, external tools and regulatory reporting. We see considerable potential for AI to make those insights easier to access and to accelerate analysis, simulation and optimisation. Today, oneZero technology handles more than $250 billion in average daily volume, 14 million transactions and 400 billion quotes each day. Operating at that scale has taught us that intelligence is only useful when the data beneath it is dependable.
The interface layer is equally important. As AI becomes another way into the trading environment, banks and brokers need secure, controlled methods for giving tools access to data and permitted actions. They should not have to rebuild the connectivity, low-latency processing and distribution infrastructure beneath them. Our long-standing investment in API-driven workflows and standardised connectivity provides that foundation while allowing each client to determine where automation belongs and where human approval remains necessary.
We are also investing where insight becomes useful to the end customer. The Engagement product line brings Autochartist analytics closer to live execution and helps brokers distribute relevant content through established client channels. Tools such as Swap Curve Manager remove fragmented manual processes while preserving trader control over market-data inputs, curves, tiers, spreads and skews. It can operate as a standalone tool, connect to an internal pricing engine through APIs or run with the oneZero Hub. Those examples reflect the same design principle: make intelligence easier to use without making the underlying workflow less controlled.
Agentic AI may allow users to construct their own trading interfaces and delegate actions using natural language. What does that mean for the relationship between people and machines, the guardrails banks and brokers provide, and the future of market oversight?
The greatest effect will be on where people spend their time and apply their expertise. Machines are exceptionally good at processing data, monitoring large numbers of variables and performing repeatable tasks at speed. People remain essential where context, accountability, creativity and judgement are required. The productive outcome is not to force every task towards autonomy. It is to give experienced people better information and remove friction from the work that surrounds consequential decisions.
Agentic access does introduce a new issue. Historically, a user interacting through a bank or broker’s front end encountered a defined experience with order-entry conventions, confirmations, risk warnings and pricing information. A sophisticated API user built a workflow explicitly and had to manage its logic and credentials. Natural-language agents can allow a much broader group of users to create a personalised interface and delegate parts of that workflow. That expands access, but it can also make the path from the user’s intent to the resulting order less obvious.
Regulators will need to form a view on the auditability of that path. If an agent interprets an instruction, chooses an instrument, determines timing and initiates an order, the bank or broker may need to reconstruct more than the execution record. The original instruction, available data, model context, permissions, intermediate steps and confirmations may all matter. Firms will also need explicit boundaries between actions an agent may recommend, prepare for approval or take autonomously. As access becomes easier, accountability cannot become weaker.
Technology providers can help make that framework practical. oneZero’s established connectivity and distribution models give clients consistent ways to reach the market, control permitted actions, observe activity and record how decisions move through a workflow.
As agentic access develops, that structure can support the transparency and auditability that firms and regulators will expect. The model will continue to change. Trusted data, resilient infrastructure and accountable workflows will remain. I believe the strongest outcomes will combine machine-scale processing with human expertise and clear control over where each is applied.
