Current trends and what comes next

Portfolio management has come a long way from the 1950’s reliance on static mathematical frameworks and fundamental analysis to the current quantitative, predictive and adaptive analytics techniques and AI-augmented portfolio management of 2026. However, industry reports such as BCG’s 2026 Global Asset Manager report indicate that asset managers trail banks and fintech firms in scaling AI across core processes, and that the industry remains in an early transition at present. This article examines how AI adoption is shaping the industry and future outlook.

AI adoption and value creation for investment managers

According to SimCorp’s 2026 research report, 70 per cent of investment managers deploy AI, but 63 per cent of firms still lack unified data across front, middle and back offices. This critical gap undermines AI readiness. In such scenarios, the value creation and return on investment for AI deployments envisioned by these firms, given their growing AI investments, are in question.  

Shabnam Sorkhi, CFA, PhD, senior principal at Simcorp, a leading portfolio management solutions firm, explains, “Most firms are creating value through targeted use cases rather than enterprise-wide transformation. The future return on investment is expected to come not only from cost savings, but from faster decision-making, improved risk management, greater automation, enhanced client outcomes and the ability to generate investment insights across the entire portfolio lifecycle.”

David Campbell, CFA, MBA, president and co-founder of ICP Securities, a registered Market Maker who develops and operates automated market making systems in Canada, explains, “Building a competitive edge such as a proprietarysecurity-specific dataset and a robust judgment overlay is more crucial to generating better return on investment on AI investment than building sophisticated workflows, since AI accuracy and efficacy rely on its inputs.”  

Prerequisites for utilizing full AI potential

A crucial prerequisite to fully utilizing AI’s potential is the user-friendliness of AI agents. This could drive AI adoption and efficiency higher. Campbell explains, “AI systems need to be intuitive, interactive and very simple to use, allowing vital judgment overlays so humans can navigate and control the agents.”

User-friendly versions of AI agent workflows for asset managers would most certainly require infrastructure and data support. The 2026 Simcorp report highlights that 58 per cent of firms are prioritizing vendor consolidation as a prerequisite for scaling AI. Sorkhi shares that the reason for this vendor consolidation push is that “AI agents can only operate effectively when they can access and act on consistent data across workflows; otherwise, they risk producing fragmented insights and suboptimal recommendations .”

AI augmented portfolio management software capabilities

AI-enabled portfolio management software in the modern age promises to radically transform the industry with the advent of some fascinating and ultra-powerful capabilities. Examples include generative AI solutions for instant insights and investment queries like Blackrock’s Aladdin copilot and strategy analytics solutions like Planisware Strategic Portfolio Management for creating a roadmap and objectives view to align investments with strategic themes. Another solution is Simcorp One’s integrated platform, which is designed to streamline operational workflows and provide real-time cross-asset data access and total portfolio views.

The industry is moving from generative AI to “agentic AI workflow systems,” which are more autonomous, requiring little human intervention. In such a scenario, buy-side firms may hesitate to hand over multi-step execution workflows to these agents due to risk management and operational challenges. A likely solution would be to adopt controlled, agentic AI workflows bound by risk limits rather than deploying fully autonomous AI agents to execute portfolio management functions. This view is also supported by Sorkhi, who says, “Buy-side firms are unlikely to hand over investment workflows to fully autonomous agents without oversight, but they will increasingly adopt governed agentic workflows where agents execute multi-step tasks within clearly defined controls.”

Reshaping risk management   

As AI agents handle more complex workflows, sufficient model risk management is also expected to become complex and difficult. Hence, the future outlook includes agentic AI enhancements beyond the traditional explainable (XAI) techniques (SHAP, LIME, etc.) so risk and compliance teams can trace exactly why a recommendation was made. Sorkhi also notes, “The focus is moving beyond traditional explainability techniques toward governed agentic workflows that provide end-to-end traceability of how a recommendation was produced.”

The traceability, explainability and transparency are not only crucial for regulatory compliance, but also to generate market and investor confidence. Campbell explains, “Markets run on confidence. AI may build that confidence by being on average more correct than a human, but there is a magic line of errors which, if crossed, can dwindle market confidence in AI agents.” 

Ethical concerns and resource constraints outlook  

As institutional investment managers utilize large-scale AI agents and infrastructure, questions arise about the ethical impacts of these on small investors, financial market integrity and society at large. The advent of AI, like any other technology, is seen as a harbinger of progress, sophistication, increased efficiency and cost reduction, especially in financial markets. However, it must be deployed responsibly within robust risk frameworks.

This is also reiterated by Sorkhi: “The key issue is not the use of AI itself, but whether it is deployed within a governed and transparent operating model. AI agents should operate within controlled workflows, trusted data environments, auditable decision processes and clear human accountability.”

Campbell presents an optimistic view: “During the last five decades, as institutional money managers have become sophisticated, they have also striven to establish best practices to safeguard investors and client interests. For example, just as CFA Institute established the CFA Code of Ethics in the past, regulatory bodies or professional organizations may come up with an AI usage code or standards to safeguard the broader society interests.”

As the computing and energy demands of AI infrastructure skyrocket, questions also arise about the environmental sustainability of scaling these autonomous agents. Campbell envisions, “This question is bigger than the industry, and in due time, a central organization may be established in future to create global standards or a framework and regulate AI resource use.”

Technological innovations as well as industry focus may also accelerate their efforts to resolve these constraints as value creation of using AI agents solidifies. Firms may focus on producing efficient, agentic workflows with lower computing costs and demands and technological innovation may reduce future demand. This optimism arises from past technological evolution trajectories, such as how much physical space 1970s computers required compared to present-day laptops, or how a 500 gigabyte hard drive was a massive data-centre concept in the early 2000s compared to being a baseline laptop capacity today. As we find more benefits in a technology, we have historically worked to make it cheaper and easier. Going by the same principle, if the industry sees true value in AI augmentation of its capabilities, more efforts to solve resource constraints will be made.

Written by

Attika Raj, CFA, is a seasoned quant finance professional currently working at TD. She has wide experience in quant research and modelling, investment management and financial advisory. Attika also actively volunteers with CFA Institute as CFA grader and Editorial Committee member for The Analyst.