[ I N D E P T H | A R T I F I C I A L I N T E L L I G E N C E ]
who- while also praising AI’ s optimisation capabilities – tells The TRADE that an entire reliance on AI should be avoided as the industry begins to warm up to its benefits.
“ On the buy-side, AI will continue to grow as an effective co-pilot for idea generation, risk management, and execution, especially through models that integrate various data sources into one. But, it will not be the holy grail. Just because AI can perform a task much quicker than the smartest PhDs, doesn’ t mean the task itself requires a PhD in the first place.
“ To this tune, I expect a steady evolutionary progress that replaces grunt work and speeds up routine tasks- but for long-only asset managers like us, the core of differentiated investing, sound judgment, and accountability for fiduciary responsibilities will remain deeply human in the near to mid-term.” certain capabilities and processes.
Speaking to this, Mahon asserts that:“ From a trading desk perspective, we are not yet at the stage where the industry is ready to hand full trading autonomy to AI agents. While the capabilities are evolving quickly, trust, governance and control remain critical considerations.”
Although many smaller buy-side firms are finding the greatest benefits from AI automation stemming from efficiency enhancements to boost trading activity and functionalities, this is also key for many larger firms, such as Azimut, one of Europe’ s largest independent asset managers.
“ The most immediate and tangible benefit of AI on the trading desk is workflow optimisation and productivity enhancement,” Mahon adds.
“ Tasks that were historically manual and timeconsuming, such as preparing market updates or desk commentary, can now be completed significantly faster, enabling traders to allocate more time to higher-value activity and decision-making.”
Similar sentiment was also echoed by Zimecki,
A well-oiled machine When it comes to ensuring the AI used on desks and operations is effective and reliable, the often-recited phrase,‘ garbage in, garbage out’ resurfaces, specifically in relation to data. As with much of the technology and automation implemented across capital markets, the output of these processes is only as good as the data that powers it.
The importance of reliable data cannot be underestimated, and as Carey states:“ You need the right technology, data backbone, and service model to support AI at scale. Otherwise, you end up with isolated experimentation within teams. Our reflection on much of that experimentation is that it ' s valuable, but it doesn ' t necessarily deliver the outcomes an enterprise
Issue 88 // thetradenews. com // 33