AI in Asset Management: Appetite vs Capability — The Skills Gap in Investing

AI in asset management is the use of machine learning and, increasingly, agentic systems to support investment research, portfolio construction, risk management and operations across an investment firm. The definition is deliberately narrow: it concerns investing capital.

That distinction matters, because the same three words name two unrelated fields. In investment asset management, "assets" are securities, funds and the portfolios built from them, and the work is allocation, selection and risk. In enterprise or physical asset management, "assets" are pumps, turbines and vehicles, and AI does predictive maintenance and sensor analysis. Both are legitimate; only the first has anything to do with managing money. The variant phrasings, including "ai asset management" and the "role of ai in asset management" question, all point at this same investment field.

AI is not one capability inside an investment firm but several, mapped to different parts of the work. Five areas account for most production use today.

• Investment Research. Models read and summarise filings, transcripts and news at a scale no research desk could match. The output is a faster start, not a decision.
• Portfolio Optimisation. AI tests how a portfolio behaves across many futures rather than a handful, supporting construction, rebalancing and scenario modelling.
• Risk Modelling. The work here is monitoring: supporting value-at-risk estimates, flagging concentration and exposure drift, and watching positions continuously rather than at month-end. Model risk itself then becomes a concern.
• Operational Efficiency. Reporting and reconciliation are heavy with repeatable tasks, and automation removes that cost first. At the macroeconomic level, J.P. Morgan Asset Management estimates generative AI could lift developed-market labour productivity by roughly 1.4% to 2.7% per year over a decade (J.P. Morgan Asset Management, 2025). This is a whole-economy effect, not a figure measured inside asset-management workflows.
• Personalisation. On the client side, AI tailors reporting and how a portfolio is explained to its owner: less material to an institutional allocator, but real.

Grant Thornton's 2025 survey of around 500 senior executives globally (Grant Thornton / ThoughtLab, 2025) frames the breadth: AI in investment management is broad in intent, uneven in execution, and concentrated for now in operations and research.

The agentic shift: from chatbots to systems that act

The first wave was conversational: chatbots and copilots that answered questions, useful but bounded. The second is agentic: systems that do not just answer but act, executing multi-step tasks across tools and data with limited supervision. That shift, not the chatbot, is what makes AI an operating concern rather than a demo.

Appetite is now near-universal; operating capability is not. The gap between ambition and what actually runs is where the industry stands today, and most coverage steps around it.

The reason is structural: a model is not an operating system. A capable model, bought off the shelf, is the easy part. Making it useful inside an investment firm needs four things the model does not supply, data, clean and proprietary, which most firms do not hold in usable shape; workflow, so the model sits inside how the firm researches, decides and reports rather than beside it; oversight, with human review and model-risk governance built in, not bolted on; and integration, the engineering that turns a demo into a system relied on daily. Assemble those four and AI operates; leave one out and the firm owns a capability it cannot run.

"We use AI" and "we operate AI" are therefore different claims, separated by years of engineering rather than a procurement decision. Industry research has repeatedly found that only a small minority of organisations have scaled an AI agent into genuine production in any function. What that capability looks like in practice, production systems and not pilots, is the subject of agentic AI as an operating system. Production beats demonstration, and that gap is where credibility is won or lost.

"AI washing" is the practice of marketing AI capability a firm does not genuinely have, or overstating what it does. It thrives when claiming AI is rewarded and verifying it is hard, and the term now appears in financial regulators' own warnings. The useful response is not cynicism but a test.

An allocator can apply four checks before taking an AI claim at face value. First, what is in production rather than pilot: not "we are exploring AI" but what runs today, in which workflow, with what effect. Second, who operates it: a credible firm can name where a person reviews, overrules and signs off, because unsupervised models in an investment process are a risk to govern, not a feature to advertise. Third, what it replaced: a real capability shows up as a process that runs differently (faster diligence, continuous monitoring, automated reporting), not as a logo on a slide. Fourth, the governance: the model-risk controls, validation and documentation that separate a system you can rely on from one you merely hope works.

The test aligns with the governance stance of bodies such as the CFA Institute, whose work on AI in asset management stresses transparency, oversight and model-risk management as the conditions for using these tools responsibly (CFA Institute, 2025). It names no competitor; the questions do the work.

Most published guidance on AI in asset management speaks to public-markets managers, so the questions that matter to private capital tend to fall to the allocator. The places AI changes the private-markets process are specific: it widens the aperture on off-market deal sourcing; it compresses diligence and underwriting from weeks of document review toward hours, changing both cost and the number of opportunities a team can assess; and it turns illiquid-asset monitoring from a quarterly valuation into a continuous read of operating signals. This is native territory for a private markets investment group, and the same dynamic runs through adjacent strategies such as private-markets investing in credit. How AI reshapes the underwriting decision is taken up in where AI changes underwriting.

Whether AI is transformative is settled. Whether a firm can operationalise it is not, and over the next cycle that is where the dividing line in asset management will fall. Appetite is cheap; capability compounds. The firms that close the gap, with the data, workflow, oversight and integration the distinction demands, will carry that advantage across cycles. For how an operator approaches this in practice, see how IMS Group runs its AI lab in production (imslabs.ai).

This content is for information only and does not constitute investment advice or an offer of any IMS Group product or service. Figures cited are sourced and dated; forecasts are projections.

This article is provided by IMS Group for information purposes only. It does not constitute investment advice, an offer, or a solicitation. Figures are point-in-time and projections are estimates.