Expert: Dr. Boris Rankov, Head of Product Strategy International (EMEA and APAC), at Investcloud Facilitator: Stephen Wall, Founder at The Wealth Mosaic
Overview:
This session looked at various topics around artificial intelligence (AI) and its applications in the wealth management industry. The opportunities and challenges of leveraging AI, as well as the capabilities of large language models like GPT, are many.
Discussion:
It was generally agreed that AI can help improve productivity and efficiency in areas like documentation, reporting, research analysis, personalized investment recommendations, and client servicing through virtual assistants and chatbots. But effective implementation relies on having the right data architecture and vendor partnerships to drive system integration, and address legacy challenges around data quality, siloed processes and systems. Hence there is a need to assess gaps in data architecture, systems, and processes before deployment of any AI-based solution. Firms should also focus on their position in the value chain and leverage vendor partnerships strategically.
A phased, risk-managed approach makes sense. It means starting with lower-risk use cases that are easier to control, like automatic generation of meeting note summaries and action logs, rather than making immediate client-facing applications. By doing this, trust in the system can be built slowly and incrementally. In addition, issues around explainability, auditability, and data privacy need to be handled sensitively, especially when using third-party Cloud services. On-premise solutions may provide more reassurance to clients initially.
Finally, culture and change management are critical to ensure staff are on board and engaged. AI should aim to augment human capabilities rather than replace jobs and this needs to be clearly communicated from the start. Firms should also design training programmes to educate staff about AI.
Key findings:
- The opportunities and challenges of leveraging AI, as well as the capabilities of large language models like GPT, are many
- Data architecture and vendor partnerships are crucial
- A sensible approach is a phased one, with risk management built in
- Explainability, auditability, and data privacy need to be handled sensitively
- Communication that AI should aim to augment human capabilities rather than replace jobs is crucial
