Despite this potential, many asset managers are discovering that there is a big difference between experimenting with AI and implementing it at scale. Nor is this a purely practical challenge. If asset managers are to succeed with AI, their clients and staff will need to trust the technology and understand that its goal is to augment people, not replace them.
To get full value from AI, asset managers will need to get three things right.
First, firms need to establish a strong risk and governance infrastructure around their AI-driven activities. This includes:
- A tailored oversight framework that monitors AI performance and risks, minimizing errors and ensuring full compliance with standards and regulation.
- A particular focus on the heightened risks — such as data security, hallucinations, “explainability,” confidentiality breaches or copyright infringement — that could arise from the use of GenAI.
- Strong operational guardrails that ensure AI governance controls are embedded in the processes, including setting-up operational playbooks to ensure AI does what it’s supposed to do.
- Appropriate staff training, including coaching employees who use GenAI on what questions to ask, how to ask them and how to interpret the responses.
Second, asset managers need to incorporate AI projects into differentiated strategies tailored to individual asset managers’ objectives. Some key steps of an AI-specific implementation approach could include:
- Rooting AI use cases in an implementation process that starts with a strategic vision, identifies the required operating models, then optimizes solution selection
- Using that top-down approach to prioritize AI investment where it will have the greatest impact — maximizing ROI by focusing on proof of value, not proof of concept
- Leveraging innovation learnings to boost agility and sharing the benefits of AI widely across the organization and its external ecosystem
Third, asset managers need to reverse their usual approach to innovation and embrace a transparent approach to AI. Openness about the potential benefits and risks of AI is key to building trust, including transparency over the data sets, methodology being used and traceability of outcomes to source. Firms must educate and empower their own staff, enabling them to communicate confidence and trust to clients – and to elevate investors’ own understanding of the role of AI.
Without the right support, asset managers risk committing a lot of time and resources to AI without generating sustainable improvements. They could even see their use of AI halted altogether by regulators or shareholders, setting back their transformation goals and creating significant reputational damage.
AI will truly be a game changer in asset management during the 2020s. The steps we have set out will be vital to keeping firms’ AI applications focused on the needs of the business; to maximizing the impact of investments in AI; and to delivering tangible benefits to investors, staff and other stakeholders. Even so, no one can be sure exactly what changes it will bring about. To explore the possibilities, along with seven potential 2030 scenarios for the industry, consult our full report on Reframing Asset Management.
