As AI becomes increasingly present in trading automation, the industry is beginning to adapt to its rise and recognise the need for necessary guardrails and regulations to ensure responsible use.
Experts speaking at TradeTech Europe 2025 identified how AI is transforming execution workflows, and highlighted its potential to enhance trade performance, analyse massive datasets, and guide dynamic decision-making in real time.
While discussing the evolution of algo wheels, Stuart Lawrence, head of European equity trading at UBS Asset Management emphasised how AI could provide benefits in a future which will include dynamic wheels that adapt strategies throughout the day.
He said: “The final step in the evolution would be using AI to actively pick where we should go, and I think a lot of the data is already there. It’s working out how to use that historical data and putting it into the decision-making of the algorithm.”
He also stated that for approximately 80% of the industry, this usage of AI is not yet a consideration, however indicated that in the next five to 10 years, firms will be looking to see how these changes can be actively implemented.
Currently, AI is not used in any way related to execution, however panellists also discussed the usage and processing of historical data as a space where AI can make a significant impact in the future, such as analysing performance strategies.
This was underlined by Joe Bellman, head of dealing at Arbuthnot Latham, who spoke of how AI can be integrated into algo wheels’ evolution for greater success.
He said: “The real use of AI and execution workflows is going to be to analyse data and help to make future decisions about how that work should be rooted. In short, AI will allow you to analyse your performance of various strategies and then help you define the rules go forward.”
However, he also highlighted the necessity for traders to be confident in using AI systems, rather than shying away from them.
“AI won’t replace traders, but traders who understand AI will replace those who don’t’.”
The need for AI guardrails
As discussions around greater integration of AI into trading practices gain more traction, the experts also agreed that robust governance and guardrails were essential to ensure transparency and fairness across models.
This was stressed by Ovidiu Campean, product director at LSEG, who said: “We need to have the ability to intervene, to have oversight and have kill switches in place. So, a trader would still need to be able to monitor the AI activity and halt the decisions that the models are making, if needed.
“AI should also be able to deactivate in case of a high market stress conditions or if the connectivity of the system breaks down. The model needs to log in continuously what the inputs are, what the decisions were and the reasoning behind those decisions. We need the ability to prevent bias and have fairness checks in place to avoid biasing towards or against different brokers.”