SimCorp's introduction of AI-powered stress testing within Axioma Risk is exactly the kind of technology shift that creates compliance obligations before most firms realize it.
Investment managers adopting this tool will need to think carefully about how it fits into their existing compliance framework. The technology takes what used to be a hands-on, specialist job and turns it into a button-push. That's a big operational win, but it also hands you a new set of risks to manage.
Receive future blog posts by email. Leave categories blank to get every post.
What This Technology Does
SimCorp's AI capability generates stress test scenarios automatically, allowing portfolio managers to focus on interpreting results rather than building the underlying models. The system draws on market data and historical patterns to configure scenarios that would otherwise require manual setup by quantitative specialists.
For investment managers, this means faster turnaround on risk analysis. For compliance officers, it means a new category of model risk to understand and document.
The Compliance Implications
If your firm adopts AI-driven stress testing, whether this product or a competitor's, your supervisory procedures need to address several questions:
- Model validation: How do you validate that AI-generated scenarios are appropriate for your specific portfolio composition and investment strategy?
- Documentation: What records are you keeping about the scenarios generated, the inputs used, and the rationale for relying on automated outputs?
- Oversight: Who at your firm is responsible for reviewing AI-generated stress tests before they inform investment decisions?
- Regulatory expectations: If examiners ask about your stress testing methodology, can you explain how the AI works in plain language?
The SEC has been increasingly focused on how firms use and supervise technology. Their 2025 guidance on AI in investment management emphasized that firms remain responsible for outputs generated by automated systems. The tool may be new. The liability is not.
Model Risk Management
For RIAs and fund managers, this falls squarely within existing model risk management expectations. If you're using quant tools to make investment calls, you need to prove you understand how they work and where they can go wrong.
AI complicates this. Traditional models have explicit assumptions you can audit. Machine learning approaches can be opaque even to their developers. Your compliance program needs to account for that difference.
What You Should Do
If you're evaluating AI-powered risk tools, start with your written supervisory procedures. Make sure they address:
- Who approves the adoption of new quantitative tools
- What validation is required before deployment
- How you document ongoing supervision of automated outputs
- What training portfolio managers and compliance staff need
Nobody's saying you should slam the brakes on new tech. But if you don't bolt it down in your compliance program, you're asking for trouble. The firms that get this right will have a competitive advantage. The ones that don't will have examination findings.
Subscribe to Regulated Intelligence Brief
Get new compliance intelligence delivered to your inbox.
Key Takeaways
Do we need to update our WSPs if we adopt AI-powered stress testing tools?
Yes. Your written supervisory procedures should address the adoption, validation, and ongoing supervision of any quantitative tools used in investment decision-making. AI tools require specific attention to model validation and documentation of how automated outputs are reviewed before use.
What records should we keep when using AI-generated stress test scenarios?
Document the scenarios generated, the inputs and parameters used, who reviewed the outputs, and any decisions made based on the results. Regulators will expect you to demonstrate that automated outputs received appropriate human oversight before informing investment decisions.
How will examiners evaluate our use of AI in portfolio management?
Examiners will focus on whether you understand how the AI works, whether you've validated it for your specific use case, and whether you have appropriate supervision in place. Being unable to explain your methodology in plain language is a red flag.
The content in this blog is for informational purposes only and does not constitute legal advice, regulatory guidance, or an offer to sell or solicit securities. GiGCXOs is not a law firm. Compliance program requirements vary based on business model, customer base, and regulatory classification.