Ampega and d-fine have conducted a joint proof of concept to investigate how agent-based AI can effectively support the validation of risk and valuation models in asset management.
The results are promising: through the close integration of multi-agent systems, ‘Talk to Your Data’ approaches and AI-supported analyses, routine tasks can be automated, quantitative evaluations accelerated and the workload on validators specifically reduced.
It is particularly encouraging that the added value lies not only in the automation of existing analyses, but has also become evident in the development of new analyses, scripts and data evaluations, as well as in additional, substantively valuable in-depth analyses.
Particularly good results were achieved in the analysis of data via ‘Talk to Your Data’, the handling of statistical queries, the performance of quantitative analyses, and the qualitative assessment and documentation of validation results. At the same time, humans remain consistently in the loop: validators continue to review, oversee and take responsibility for the results, whilst the AI frees them up to focus on core economic and substantive issues and quality assurance.
Our conclusion: The PoC impressively demonstrates that, when properly integrated with data management, technical systems and governance, agent-based AI opens up considerable potential for increased efficiency, scalability and quality improvements in model validation.
You can download the complete white paper at the top of this page.
