2026 investment management outlook | Deloitte Insights
Firms are scaling AI from isolated experiments to enterprisewide platforms
The reason for this talent acquisition shift is clear: Early indications suggest that some firms deploying AI are beginning to see promise from their AI deployments; as noted below, many are increasing their budget allocations in response. Unlike other technology upgrades that were often for either cost savings or marketplace advantage, AI has shown that it can help firms achieve both. Indeed, some investment management firms have moved from AI exploration to execution, delivering more concrete ROI in specific domains in the past 18 to 24 months.54 For example, some investment managers have successfully implemented gen AI throughout client communications and distribution processes, using it to write investment commentaries, generate sales leads, and capture client insights.55 Morgan Stanley’s financial advisers are rapidly adopting AI (98% penetration) and using the tools to help bring in new customers and assets at an accelerated rate.56 The AI-based tools can help advisers connect the right internal materials to each client, based on their unique needs, and document key points and action steps for each client interaction.
AI is also supporting some of the most intellectually demanding aspects of investment management. Schroders’ virtual investment committee agent was designed to analyze sector dynamics, evaluate business model implications, and assess potential risk factors to support investment decision-making.57 This effort could largely disrupt the heart of active investment management. Portfolio strategy and execution are at the core of the industry, and overseeing AI models in this space can present a major intellectual challenge. Meeting that challenge may require a rare blend of technical skill and investment insight.
Private equity (PE) firms also offer a compelling example of the accelerating tactical adoption of AI. Overall adoption of AI in the PE due diligence process is accelerating rapidly, with 64% of PE firms reporting that they are using AI to streamline the due diligence process.58 This development can help reduce costs and enhance market outcomes. Beyond efficiency, AI is increasingly being leveraged to identify prospective portfolio companies and support early relationship initiation. These capabilities often stem from a blend of learned success factors, advanced search functions, and sophisticated pattern recognition. This collaboration between human judgment and machine intelligence can help firms model traits of a strong portfolio company with greater precision. While AI can provide valuable insights to help guide investment decisions, the ultimate authority remains with firm leadership.
Projects like these can tackle cross-organizational processes, and leading approaches typically have C-suite sponsorship. While AI may be taking center stage, it is equally important to recognize that modernizing core systems is important for AI to function effectively and deliver real value.
A word on agentic AI
A recent Deloitte survey indicates that 26% of respondents reported that their organizations were already exploring autonomous agent development to a large or very large extent.59 The art of the possible for agentic AI is both broad in scope and deep in complexity. For context, AI agents are software systems that can complete complex tasks and meet objectives with little or no human intervention. They are called “agents” because they have the agency to act independently, planning and executing actions to execute assigned tasks consistently and reliably by acquiring and processing multimodal data, using various tools to complete tasks, and coordinating with other AI agents—all while remembering what they’ve done in the past and learning from their experience.
Let’s assume for a moment that a retail firm has advanced agentic AI capability with access to both operational and client interaction data. Agentic AI could screen interactions for common questions, creating behavior-driven segmentation, and then analyze portfolios looking for areas of improvement, such as raising dividend yield, improving diversification, and improving risk-adjusted return. Based on these findings, AI could then alert representatives with possible investment recommendations in language that appeals to the segment of the investor after checking that the recommendations are in compliance with proper documentation and prospectus delivery built into the process. This process is an example of strategic AI implementation that can potentially contribute to cross-organizational effectiveness and efficiencies.
AI graduates from sandbox testing to proving its worth
Use cases such as these are beneficial and will likely continue to be pursued because they can provide clear positive ROI.60 Some important metrics for tracking AI ROI may include time saved per workflow, incremental revenue or assets generated through AI-assisted distribution, reduction in operational errors, uptime and availability of AI platforms, and the ratio of models deployed in production versus those remaining in sandbox environments. However, the sum of tactical implementation benefits is less than the whole. Moving forward, AI funding is becoming a growing priority for many financial services firms; many have boosted their AI project budgets, noting its strategic significance.61 This funding will likely be deployed differently, with enterprise infrastructure and unified AI principles taking priority over discrete, task-based solutions that often excited firms in the early days of AI exploration.62
That said, many firm leaders should move beyond asking, “Can AI add efficiencies to a given task, process, or operating model?” Instead, the questions are: What are the priority projects? What are the risks of deploying AI? Which tasks should be performed internally versus outsourcing them to a service provider more cost-effectively? And as solutions expand, what is the governance model for AI-enhanced processes?
Without strategic frameworks, guidance, and governance, these projects can multiply into an unwieldy collection of mismatched policies, undocumented assumptions, operating procedures, and data management choices.
Despite increases in job postings citing the need for AI expertise, our analysis of investment management job postings also shows that current governance mentions remain generic and not AI-specific.63 While the SEC recently withdrew an AI rule it proposed in 2023, the regulator continues to prioritize enforcement against AI-related misconduct.64 This helps underscore a need for companies to establish robust AI governance frameworks—ideally led by a chief AI officer or equivalent—to proactively manage compliance with existing rules and have a human in the loop.
Imagine being the chief compliance officer, and you are asked for the firm’s AI data privacy and IP protection policy. The answer may not be straightforward; it can depend on when and where each AI use case was introduced, with numerous departments likely having implemented their own ad hoc solutions to achieve a short-term solution. At this point, while some tactical AI solutions may deliver satisfactory results, the overall approach is often fragmented, making a unified strategic framework important for both risk management and long-term planning.
To get the most from AI over time, leading organizations could turn to general, generative, and agentic AI managed by a central team with enterprisewide authority. To be more effective, such central teams should prioritize governance artifacts that help ensure both control and trust in AI adoption—for example, maintaining a model inventory, conducting model risk assessments, establishing clear data lineage, vetting vendors rigorously, defining human-in-the-loop thresholds, ensuring audit trails, and preparing incident response protocols
link
