Zac Maufe, global head of regulated industries, Google Cloud
Next year will mark a turning point for the financial services industry as generative AI moves from pilot programs to widespread adoption. This transformative technology promises to unlock new levels of productivity, extract valuable insights from data, and strengthen defences against fraud.
Financial institutions are drowning in data but struggling to extract its true value. Market analysts and compliance officers, for instance, can spend countless hours sifting through information scattered across various documents and systems. AI-powered intuitive search and advanced summarisation capabilities offer a solution, allowing employees to quickly locate and analyse information so they can focus on higher-level analysis and decision-making, rather than getting bogged down in data wrangling.
AI agents will further increase productivity by supporting routine tasks like summarising news articles and financial reports. These digital assistants will free human employees to focus on more complex and strategic work, adding value where human expertise is truly needed.
AI will also become a significant shield against financial fraud, and crucial in detecting and preventing cyber-attacks, protecting sensitive data, and maintaining trust. AI’s ability to analyse unstructured data, identify complex patterns, and prioritise alerts empowers security teams to detect and prevent fraud more effectively, safeguarding both institutions and their customers.
The success of these AI initiatives hinges on robust data platforms. Financial firms with these platforms can aggregate data from various sources, ensure data quality, and make it readily available for AI applications, enabling them to scale their AI initiatives effectively.
Kelvin To, founder and president, Data Boiler
Amid people having evermore data and AI power than ever before, it’s getting harder to discern truths. What was known in the past is becoming fragile because some fundamentals have shifted and people are ditching time-tested principles inadvertently. Knowledge is power, yet perfecting knowledge is hard, or it takes forever to achieve. The more one tries to finesse, the more noises will convolute the essential signals. AI destructs independent thinking.
Craving for evermore data is flawed unless it’s an AI foundational model or a totalitarian. Markets don’t get more transparent in a cyberpunk-era but people are, because one seldomly acts out of character. Data quality has a lot to deal with distilling onset signals from noises, as well as tolerating to imperfection, and making adjustments to structure and timeline of anything. Different perfectionists emphasise on different aspects of 4Vs. Equivalent exchange is nothing more than a power play between them in jockeying around to make money.
The trends with the greatest market impact boils down to (i) learn to discover unknowns; (ii) unlearn to avoid being the subject of reverse engineer; and (iii) re-learn “a good decision, made now and pursued aggressively, is substantially superior to a perfect decision made too late.”
Robert Miller, head of market structure and liquidity solutions, KCX
Artificial Intelligence (AI) and advanced analytics are reshaping trading and compliance by improving execution strategies, enhancing market efficiency, and driving cost reductions.
AI-powered algorithms are already analysing vast amounts of real-time data to optimise trade execution, identify patterns, and adjust strategies based on market conditions. In 2025, the adoption of these strategies will continue to increase, and I am sure we will see more of these execution products made available to clients.
Some algos, often using AI can now change strategy at any point in the execution life cycle, regardless of benchmark. The challenge the buy-side will face is how this aligns with their existing best execution frameworks. Traditional benchmarks used in best execution policies may be outdated when compared to trading against a dynamic benchmark that adjusts in real-time. This raises important questions about how transaction cost analysis (TCA) can be accurately applied when benchmarks are unknown or moving targets. The implications for performance measurement and compliance are significant, as TCA must evolve to reflect these dynamic conditions and companies must adapt their best execution policies to integrate a new breed of algo.
Balancing innovation with regulatory oversight to maintain standards and achieve the best outcomes for clients for be key for 2025.
Eugene Markman, chief operating officer (FX), ION Markets
The convergence of AI and the increasing availability of self-service tools will likely have the largest impact on the FX market in 2025.
AI tools are becoming increasingly common across the financial services industry, including within FX. Although not yet widely implemented, AI tools have the potential to enable traders to perform routine tasks faster, near instantly analyse vast quantities of data, expand the abilities of today’s TCA, and predict market trends with greater accuracy and speed than humans.
Simultaneously, self-service tools are increasingly being used to streamline the FX industry. These tools allow smaller firms and individual traders to access technological capabilities previously only available to large financial institutions.
In 2025, AI and self-service tools will likely converge, levelling the playing field and providing smaller players access to advanced trading capabilities. However, the path to widespread adoption of AI is not without challenges. Regulations and the technical complexities of AI implementation remain significant hurdles. As a result, while 2025 is likely to witness the expansion of AI and self-service tools in the FX market, this transformation is expected to be gradual rather than rapid and straightforward.