1stProtect, a runtime security company that governs what AI agents are allowed to do, and WealthRyse, a Hong Kong-based AI-native technology company that provides portfolio calculation and optimisation systems to licensed financial institutions, have announced a strategic collaboration to secure AI agents in financial services. The joint announcement was issued on 4 November 2026. The initiative is intended to integrate 1stProtect with WealthRyse's AI agents so that security policies are enforced in real time, before any action is executed.

The collaboration reflects a shift under way across financial services, as institutions move from AI that answers questions to AI that takes action. In that model, security must extend beyond reviewing model outputs to governing the actions agents are able to perform. Risks such as prompt injection, data exfiltration, and actions outside an agent's mandate create challenges for security and compliance teams. The partnership is designed to help WealthRyse clients enforce the boundaries those teams set, with each proposed agent action checked before execution.

Runtime security, as the term is used here, refers to controls that operate while an AI agent is running, checking each proposed step against defined policy rather than only inspecting what a model produces.

“AI can do the heavy lifting. It cannot take over fiduciary responsibility,” said David Lee, Co-Founder of WealthRyse. “Advisers and managers remain accountable for their decisions, however powerful the technology becomes. Human judgement must have the final say. That is the principle behind this collaboration: technology should support the people responsible for clients' interests, not replace their judgement.”

“Giving an AI agent more capability is only half the job. The other half is defining what it must never do,” said Kervin Pillay, CEO of 1stProtect. “An agent's mandate needs to be an enforceable limit, not just an instruction. Our role in this collaboration is to check proposed actions against policy before execution, so greater autonomy does not mean permission to operate outside that mandate.”

Under the planned integration, 1stProtect's SIGMA engine will sit alongside WealthRyse's agents, bringing policy enforcement into the portfolio workflows institutions already use. The deployment is designed to keep client data within the institution's own environment and to support the security and data-handling requirements of regulated deployments in Hong Kong and beyond. As set out in the announcement, the integration remains at the planning stage rather than in production: no rollout timeline, named client institutions, or commercial terms were disclosed.

The arrangement pairs complementary roles: WealthRyse supplies the AI-native portfolio calculation and optimisation systems used by licensed institutions, while 1stProtect provides the enforcement layer that determines which proposed agent actions may proceed. For readers tracking agentic AI in regulated finance, the near-term signals to follow are concrete delivery milestones for the integration and evidence of adoption by licensed institutions, in Hong Kong and in other jurisdictions consistent with the stated 'Hong Kong and beyond' scope.

Source: Global FinTech Series