The Work AI Can Do Starts Internally

AI now handles much of the tedious prep work—summarizing notes, cleaning data, flagging anomalies—so financial professionals can focus on the strategic decisions that follow. For example, businesses of all types are using basic AI at no cost to handle tasks such as minute-taking, while some are paying modest subscription fees to use AI tools to do jobs such as editing verbal tics out of a video recording and producing a polished transcript. Here is an overview of the ways AI is currently being used in a variety of applications:

Investment management: Consultancy Oliver Wyman notes that generative AI can empower portfolio managers in the areas of by automating information collection, summarization, and data cleaning, investment research and risk analysis can shift focus to higher-value tasks like validation and idea generation—boosting productivity by up to 30%. In middle and back-office functions, efficiency gains of 25–50% are achievable in legal, compliance, and operations through enhanced task automation and democratized coding capabilities.

In sales and client service, the firm found that AI can automatically draft talking points before or during client conversations and help prioritize client outreach. It found that using AI tools produced a more than 20% upsell hit ratio for institutional clients targeted with tailored conversations, up to 30% in net new money based on successful placement of products on distributors’ fund buy list, and a 20% decline in AUM attrition through proactive engagements.

Risk Management: Here, the use of AI can improve the calculation of value at risk (VaR), which estimates the max potential loss of a portfolio in a given time frame, and expected shortfall, which measures the average loss beyond that threshold. Basically, a type of AI called a generative adversarial network (GAN) can play a game against itself and, in the process, determine the likelihood of outcomes in a vast array of forward-looking financial time series. Those simulations feed into VaR and shortfall models, improving how firms quantify risk within their specific portfolio and strategy tolerances.

Bridgewater’s AI Lab is an example that simulates how entire market systems behave. It uses machine learning to model risk in ways that respond to a portfolio’s specific structure. Its AI doesn’t just scan for red flags; it runs scenario simulations and stress tests based on each strategy’s exposures, from commodities to macro trends. That allows Bridgewater to adapt risk oversight in real time, rather than rely on static models.

Generative AI improves and customizes risk management by uncovering nuggets in unstructured data sets, then gleans insights specific to users’ specific strategies for mitigating industry or geographic risk. This allows funds to acquire or divest assets to properly balance risk and reward. However, few money managers are availing themselves of these risk mitigation benefits, according to Swiss financial industry regulator FINMA, so it represents an opportunity for forward-thinking firms to establish a competitive advantage.

Trade Execution: While the AI tools behind streamlining trade execution are further removed from investment strategy than the tools driving asset allocation or even risk management, they are still relevant, albeit less directly.

Tools like Tradeweb’s AiEX (Automated Intelligent Execution) or JPMorgan’s Execution Optimizer use rules-based automation and predictive models to route trades based on strategic preferences. This enables each strategy to execute trades according to its own logic—some may prioritize speed, others price, or market impact—ensuring execution aligns with portfolio intent.

Tradeweb’s AiEX, for example, lets traders set specific rules for how they want orders handled—pricing thresholds, timing preferences, or even which counterparties to prioritize. Once those rules are locked in, AiEX handles the routing automatically. So instead of spending time manually executing trades, a team can focus on strategy. It’s like each fund gets its own set of trading instructions that reflect how it actually wants to operate.

What makes that interesting from a personalization standpoint is how much flexibility it gives the investment team. One strategy might prioritize speed to capture short-term market moves, while another might want to optimize for price or minimize market impact. This kind of configurable automation helps funds implement distinct investment strategies at scale, tailoring execution to specific goals, risk profiles, and portfolio mandates.

The Key to AI Success is Good Data

While often taken as a given—which is not always the case—good, clean data is the essential foundational building block for all the efficiencies that AI can deliver. Its results, summaries and conclusions are only as good as the information the algorithms draw on. If your data is unreliable and incorrect, AI will just lead people into making bad decisions faster, making those decisions more difficult and probably more costly to unwind. That’s why data quality—and governance around how AI uses that data—needs to be part of the conversation from day one.

We cannot stress enough that the data layer is critical. From the quality and quantity of datasets to the database management, from the analytical tools to the dashboards that summarize their findings, it is crucial for any financial intermediary to put its data house in order. The more error-free and more easily referenced the data are, the more GenAI can deliver accurate, insight-driven outcomes. Even more so, the better the data are, the better basic, pre-GenAI tools function as well.

Institutional Investor Relations: A World Apart from Retail

While consumer banking and the client-facing arms of mutual fund complexes, for example, have embraced AI—Nvidia reports that more than half of all financial services firms surveyed now deploy AI in their chatbots, virtual assistants and similar devices—the institutional world will never shunt clients to an avatar.

A private capital investor calling to discuss a proposed multimillion-dollar trade, for example, expects a human being to answer the phone. And they should. Only a knowledgeable human can articulate an investment thesis, explain risk, and ultimately be held responsible if anything goes wrong. Nevertheless, that human being need not do all the iterative work. AI is helping firms align their strategies with each investor’s unique profile—in ways that simply weren’t scalable before.

If your firm is running into challenges like scattered data sources, limited IT bandwidth, or a reliance on a few key technologists, it may be time to explore the various ways GenAI can help. The right solution can make human expertise, and even personalization more scalable—not just possible—without adding complexity.