CFA Society Switzerland: AI Pilots for Member Services & Operations
A controlled pilot exploring how generative AI can improve member experience and operational efficiency, with governance and human oversight.
This engagement illustrates Medusa's approach to applying AI in regulated, non-investment contexts.
Shared with client consent. Details are intentionally high-level.
1. Client Context
CFA Society Switzerland serves a broad community of investment professionals and stakeholders. As a member-based organization, it balances increasing expectations for accessibility and responsiveness with high standards of accuracy, trust, and governance.
Like many professional societies, it faces growing volumes of member inquiries and operational communications, while operating with lean teams and well-defined processes.
CFA Society Switzerland explored the use of generative AI to support member services and internal operations through a limited pilot initiative.
The objective was to assess whether AI-based tools could improve accessibility to information and support internal workflows, while maintaining appropriate governance and human oversight.
2. The Challenge
Two areas were identified as well-suited for a practical AI pilot:
- Member information access: Exploring whether an AI-powered chatbot could help members retrieve information about events, governance, and key contacts based on publicly available content.
- Email triage: Exploring the use of AI-assisted classification to support the routing of incoming emails, with validation by staff.
The objective was to explore how AI could support these workflows without disrupting existing operations or introducing unmanaged risk.
3. The Solution
Medusa designed and implemented two AI-based pilot solutions:
AI Member Chatbot (Pilot → Live Soft Launch)
A conversational assistant trained on CFA Society Switzerland's public content, enabling members to quickly retrieve information about events, governance, and key contacts.
AI Email Triage (Pilot Phase)
An automation workflow that classifies incoming emails and proposes routing, allowing staff to review and validate results before moving toward full automation.
Both solutions were implemented using a modular, cloud-based architecture with explicit support for human oversight.
4. Approach & Governance
A core design principle of the engagement was controlled rollout:
- Soft launch rather than immediate full deployment
- Human-in-the-loop validation during early stages
- Clear separation between pilot stabilization and live operational support
This approach ensured transparency, traceability, and the ability to adjust or stop the pilots if needed.
5. Early Outcomes
While still in an early phase, initial outcomes include:
- Faster access to information for members
- Tangible demonstration of how AI can enhance day-to-day interactions
- Improved visibility into email volumes and categories for internal teams
Just as importantly, the pilot helped stakeholders develop a practical understanding of the strengths and limitations of AI in a real-world setting.
6. Why This Engagement Matters
This engagement illustrates how organizations can move beyond AI experimentation and focus on pragmatic, value-driven use cases, even in environments where trust, accuracy, and governance are critical.
For Medusa, the project reflects our broader approach: helping organizations explore and deploy AI in a way that is concrete, responsible, and aligned with real operational needs.