How asset managers and institutional investors may implement AI into their investment process.
When surveying 900 institutions across 25 countries for the Global Institutional Investor Survey, Fidelity found three-quarters of respondents think it’s unlikely the industry will be the same in seven years. Many cited potential disruption by artificial intelligence (AI), but optimism is generally high around the world about its potential incorporation into many high-value investing functions such as evaluating portfolio performance and risk and determining optimal asset allocation strategy. When it comes to this application, evidence from our survey responses suggests that many of these services are expected to be utilized as a tool to augment work, not fully supplant the roles of analysts or institutions’ investment partners.
Applications by Institutional Investors
Many institutions are already acting on their convictions about AI—34% are either currently testing or have already begun integrating AI into their decision-making processes, while another 35% are exploring its capabilities. Those institutions who have begun to implement AI indicated high confidence that the technology is delivering on its promise of improving decision-making; 35% reported that AI has removed some of their human biases and more than half claim to have already gained new analytical insights. Like other technological advancements, the implementation of AI will likely be a progression, where trust in the technology is built over time. The investment management industry at large remains in the early days of AI adoption, and to date, AI applications have focused largely on improving back-office operational functions.
Our survey results suggest:
- Institutional investors largely expect asset allocation modeling to shift to AI over the next several years.
- More than two-thirds of institutional investors felt they would be relying on AI for monitoring and evaluating manager and portfolio performance and risk by 2025.
- Nearly half of global respondents believe they will eventually depend on AI to recommend managers for a portfolio, while fewer will rely on AI to select managers.
- Forty-seven percent believe they will use AI to time their investments.
- Thirty-nine percent stated they will use the technology to bypass asset managers entirely and create customized portfolios directly from underlying securities.
Investment services that AI will be used for by 2025
Survey Question: Do you feel that, by 2025, it is likely you will rely on artificial intelligence for any of the following services? Responses show percentage of respondents who selected the listed service. Available responses also included “none of the above” (5% selected) and “Other, please specify” (0% selected).
Many institutions are still firmly in the educational phase of exploring AI and some have no plans to implement this technology on their own, possibly due to the significant financial investment and commitment required. But when it comes to applying this technology for the core functions of investing, a hybrid approach in which AI is used as a tool to augment human decision-making has the power to transform the investing landscape—so even for those with no plans to deploy AI for their own operations, possessing the skills and internal processes to enable discussions with investment partners and asset managers may become a more crucial part of an institutional investor’s toolkit.
Use of AI by Asset Managers
Led by quantitative hedge funds, large asset managers, and emerging fintechs, investment firms are moving quickly to embrace AI by aggressively recruiting data scientists and software developers to incorporate these techniques. Asset managers on the forefront of exploring and implementing AI stand to set themselves apart with the potential to improve investment outcomes, as advanced algorithms may assist in alpha generation, risk management, faster and more effective trading, and efficiency gains in the portfolio management process. While the adoption of AI for investment decision-making remains in its early stages, pricing and competitive pressures have triggered an acceleration of AI exploration and adoption to improve investment results in the asset management industry.
AI tools have the power to augment each stage of the investing process, at least in part by automating manual, redundant tasks, allowing humans to focus on their highest-value work.
AI has the potential to augment various stages of the investment process
For illustrative purposes only.
DATA AGGREGATION: In addition to traditional asset pricing, accounting, and macroeconomic data, an explosion of “alternative” data has created the need for automation to aggregate information from disparate places such as social media sentiment, media articles, geospatial imagery, and foot traffic.1 AI can be used to gather and convert data housed in non-traditional databases to a machine-readable format, enabling analysts to more easily extract insights and signals.
RESEARCH & ANALYSIS: AI tools such as natural language processing and text analytics are commonly used to contextualize words from company filings, financial statements, and earnings calls to gauge management sentiment and provide analysts with efficient access to potentially valuable insights.2 This research support may help analysts and portfolio managers build the mosaic of fundamental analysis around a security, and lead to greater efficiency and enhanced alpha generation.
PORTFOLIO CONSTRUCTION & RISK MANAGEMENT: Recent research has touted the potential benefits of using AI to estimate asset returns and volatilities.3 AI algorithms that identify nonlinear relationships in asset market data may produce portfolios better able to adapt to changing market environments and even anticipate future changes, which may result in more efficient portfolios.
TRADE EXECUTION: While ideal execution may be difficult to achieve, AI tools can help by analyzing millions of data points in very little time to allow trades to be placed at the best prices.4 Furthermore, AI can be used during other stages of a trade’s lifecycle, from forecasting capital market liquidity to detecting anomalies that may uncover trade order errors.5
PERFORMANCE MEASUREMENT: Investors are persistently looking for ways to enhance their performance. AI tools can help portfolio management teams analyze their investment processes and trading history, for example, to identify key drivers of performance and potential behavioral biases.6 With such real-time feedback, portfolio managers may be able to avoid suboptimal decisions and improve their results over time.