Thousands of assets and dozens of asset classes, high volatility, and increasingly stringent regulatory requirements are all realities of digital trading platforms that serve crypto exchanges, brokers, and institutional clients. Yet, despite today’s markets’ immense scale and swift speed, many still handle risk management in Excel. AI in digital asset trading is no longer optional — it’s becoming foundational. Can established practices adapt to the challenges of new circumstances?
The relationship between classic spreadsheets, API, and automated trading platforms is complex, raising questions of whether they are complementary or antagonistic. While classic spreadsheets may seem limited in handling large datasets and complex analyses compared to AI, combining them with APIs can bridge the gap and leverage the strengths of each technology. However, today’s rapidly changing world still prompts us to start employing AI tools ASAP because opportunity knocks only once.
The Risks of Sticking with Manual Spreadsheets
Firstly, comparing Excel with AI tools is inherently flawed. It’s similar to comparing traditional postal services with ISP mail delivery by an SMTP server. They solve fundamentally different problems: Excel was built to help users manually analyze and organize static data. AI and algorithmic systems, by contrast, are designed to process real-time information and take action automatically based on predefined rules.
In practice, the context of AI is often confused with algorithmic concepts; therefore, it is essential to differentiate between these categories. Terms like “AI-driven trading” are frequently used, but in reality, most solutions involve predefined algorithms or rule-based automation. Therefore, to avoid confusion, it’s more accurate to use terms like algorithmic trading, automation, or rule-based systems.
Even today, many investment firms process input data semi-manually — through spreadsheets and macros that require preliminary debugging, uploading, and cross-checking, all of which is very complex and error-prone. Large dealing or trading departments at well-established investment companies often rely on such outdated workflows, despite the availability of automated systems or intelligent execution frameworks.
Relying not only on Excel but also on general manual systems and outdated automation scripts to manage trading platforms during high-volatility situations can be quite risky. Mistakes in data entry, formula mishaps, and the absence of real-time updates can result in mispriced trades, delayed risk management, and even forced liquidations.
History offers painful reminders of what happens when spreadsheets fail. The loss that JP Morgan faced in 2012 can be traced back to a straightforward copy-and-paste blunder. The company relied on spreadsheets to develop its value-at-risk models (VaR), but an employee accidentally input incorrect data into the model. He simply copied the wrong information from one spreadsheet to another, which led to a model that underestimated risks. This mistake ended up costing the company a staggering $6 billion.
Another example comes from 2023, when Norway’s massive $1.5 trillion sovereign wealth fund announced that it had taken a hit of NKr980 million, which at that time was about $92 million, due to a mistake in how it calculated its required benchmark. That blunder resulted in a slight overweight in US fixed income compared to global fixed income. Once the issue was identified, the management jumped into action to fix it, but given the fund’s enormous size, the return only adjusted by 0.7 basis points. As a result, their previously reported positive relative return of NOK 118 billion was revised down to NOK 117 billion.
Excel just isn’t built for high-frequency trading; it lacks automation and struggles with large datasets. Plus, operational inefficiencies, downtime, and no disaster recovery options make it a poor fit for today’s trading landscape. However, a poorly written algorithm can also result in substantial losses. The risk isn’t limited to manual methods. Automation carries risks, too. In a striking example, Knight Capital in 2012 reported that a trading routine glitch cost the company some hefty $440 million. Neither manual tools nor automation can be trusted blindly — both require rigorous oversight and clear responsibility.
Laying the Groundwork for Automated and Algorithmic Trading
However, automation isn’t just about speed and scale. It’s also becoming a regulatory imperative.
Automation and advanced algorithms can also play a significant role in enhancing fraud detection and monitoring margins, which helps reduce repetitive manual tasks prone to various types of errors and boosts overall efficiency. With smart order routing and algorithmic trading, we can expedite execution and secure better prices.
By embracing automation and advanced algorithms, brokerages can enhance their accuracy, mitigate risks, and stay ahead in unpredictable and often volatile markets. Still, many tasks can be implemented through various situation handling rules and alerts, recognizing current limitations but maintaining ambition for the future.
While true AI in digital asset trading remains a future goal, automation and algorithmic systems are already transforming how platforms operate today. Algorithmic and automated trading is expected to become more prevalent, particularly in liquid asset classes like equities, government bonds, and listed derivatives. With 57% of respondents in 2022 reporting that over half their trades were executed via algorithms, adoption is clearly accelerating.
But let’s get back closer to the industry’s challenges. Trading platforms need a scalable, cloud-based architecture with real-time data processing, robust APIs, and seamless system integrations to implement AI effectively. A strong data culture requires clean, structured, and high-quality datasets, along with governance frameworks to ensure compliance and security. Cross-functional collaboration between data scientists, engineers, and traders is essential to maximize AI’s potential and drive competitive advantage.
Recent research found that 77% of executives believe unlocking AI’s true benefits can only happen when it’s built on a foundation of trust. Developers and users must build a bridge of trust in digital systems and AI models. As a result, all types of customers and the workforce will be assured of the AI systems’ accuracy, predictability, consistency, and traceability.
AI Meets Regulation
However, the value of automation and real-time monitoring tools extends beyond technology — they’re essential for meeting fast-tightening global regulatory standards. In a highly volatile environment, manual control is becoming ineffective and dangerous. As a result, firms are increasingly digitising compliance: a PwC survey found that 49% of respondents use technology for 11 or more compliance activities, indicating continuing momentum in digitising compliance models.
Authorities like the SEC, the FCA, and global frameworks such as ISO and SOC-2 are making automation a necessity in trading and risk management systems. These tools — often mislabeled as “AI” — represent the backbone of intelligent automation in trading platforms, supporting real-time risk monitoring, compliance, and data security.
Regulations such as MiFID II in the EU and the SEC’s Market Access Rule are pushing firms to implement real-time risk controls to curb market abuse, excessive leverage, and failures in algorithmic trading. Automated surveillance tools can spot anomalies, detect market manipulation, and prevent flash crashes far more effectively than human oversight ever could.
In the realm of automated compliance and fraud detection, frameworks such as AML (Anti-Money Laundering) and KYC (Know Your Customer) require ongoing scrutiny of transactions for any suspicious activity. AI-powered systems enhance this by analyzing large volumes of transactions in real-time, ensuring faster and more consistent compliance.
Data Security & Auditability are also crucial, with regulations like SOC-2 and ISO 27001 emphasizing the importance of data governance, security, and risk controls in financial services. Automation and real-time monitoring tools support cybersecurity by detecting intrusions, preventing data breaches, and automating regulatory audits.
Lastly, with the surge of high-frequency trading (HFT) and automated market-making, regulators call for tighter trading algorithms oversight. The FCA and SEC have rolled out stricter rules on algo-trading risk management, and automated systems can assist by backtesting, stress-testing, and ensuring that algorithms stay within regulatory boundaries.
The Cost of Falling Behind
Platforms that delay AI adoption risk losing out on a technological advantage and on resilience in the face of increasing market demands. Without automation, customer loss, risk management errors, and regulatory pressure all become more likely.
In the next 2-3 years, moving to a data-driven, automation-first model will no longer be a matter of convenience. It’s a step towards sustainability and scalability. Artificial intelligence is not a panacea, but it is what gives platforms the chance to cope with increasing complexity and evolve faster than the market. Technology must be ahead of the curve in a world where milliseconds decide everything.
Eventually, intelligent automation in trading platforms — especially in risk management — should be a top priority for digital asset platforms due to increasing market volatility, evolving regulations, and rising cybersecurity threats. Platforms without AI-driven risk controls will face higher exposure to fraud, liquidity crises, and regulatory penalties. Falling behind could result in reputational damage, loss of investor trust, and potential business failure in an increasingly competitive and regulated market.
