Algorithmic Trading & AI
FIX's work in this space has focused on developing standards for order submission, execution and data, helping firms to normalise and understand their data, and therefore measure their performance in a consistent and explainable manner.
From the earliest days of electronic trading, firms have used technology to automate many parts of the execution process.
FIX's work in this space has focused on developing standards for order submission, execution and data, helping firms to normalise and understand their data, and therefore measure their performance in a consistent and explainable manner.
FIX has also focused on governance, including testing guidelines and a full suite of messages and workflows for testing and certification of algorithms.
Though rooted in algorithmic trading, FIX is working on extending these principles and designs into a world of general automation, increasingly powered by generative AI.
The FIX Protocol supports algorithmic trading by providing standardised, machine‑readable ways to declare and carry an algorithm’s configuration and identity throughout the trade lifecycle.
The StrategyParametersGrp is used to convey named parameterisations of an algorithm (e.g., how a strategy is configured) in a consistent, extensible structure.
Complementing this, Extension Pack EP297 enhances algorithm transparency in executions by extending the ExecutionReport(35=8) message with fields to identify the algorithm and its lineage—specifically including AlgoCertificateID(3012) (certificate snapshot), AlgoTrialID(3097) (trial/configuration identifier), and LastAlgoID(3098) (execution algo identifier)—so firms can determine which algorithm/certificate (and, in trial contexts, which configuration) actually generated an execution.
Regulatory requirements such as MiFIR RTS 7 require that trading venues ensure their members certify algorithm testing to maintain market stability and transparency.
However, a lack of standardisation and limited electronic communication methods have posed challenges. In response, we have developed robust, standardised messaging workflows that enable meaningful, legally compliant, and comprehensive algorithm certification and reporting.
This approach provides precise algorithm definitions, thorough documentation of testing procedures and results, and clear pass/fail criteria, while also supporting interoperability between non-production trading systems and testing infrastructure to facilitate efficient test orchestration and information management. Details can be found in two extension packs - EP292 and EP295 plus our Recommended Practices on Unique Identification of Algorithms and Certificates
Regulators and financial institutions globally are becoming increasingly aware of the potential risks posed by the deployment of the latest AI technologies. To realise their full potential while limiting the risk of disruption, governance frameworks need to be adapted. We are building on our experience with algorithmic trading to consider this area and make recommendations to industry practitioners.
One example of regulatory interest in this area is the consultation on AI risk management guidelines from the Monetary Authority of Singapore, to which the FIX Trading Community issued a consultation response in March 2026.