GenAI is reshaping investment workflows faster than most companies can adapt. The release of Claude for Financial Services is the latest step in the application of GenAI in the investment industry. The focus on domain knowledge and specialized workflows sets it apart from general frontier LLMs and raises important questions about how financial workflows will evolve, how tasks will be divided between humans and machines, and what skills will be needed to succeed in the future of finance.
Financial companies are facing the most dramatic overhaul of technology capabilities in a generation. AI-driven digital transformation is reshaping functions and investment processes, forcing professionals to rethink the boundaries between human and machine cognition, while companies work to upgrade their technological knowledge and human capital to stay competitive.
Amid this shift, companies and professionals must reevaluate the skills needed for success. Predicting how AI will change workflows and functions is challenging given the pace of technological progress and the uncertainty surrounding transition pathways. Yet this assessment is necessary for strategic planning, both for industry leaders and individuals considering their career path.
The CFA Institute continuously monitors and interprets AI developments and provides guidance and training to help financial professionals navigate the changing landscape and build the career skills they need to succeed. To further this mission, we are embarking on an ambitious project to analyze the structural implications of AI for the investment profession. We will explore scenarios for how AI will impact professional practice, judgment, trust, responsibility and career paths, building on our research to date.[1]
In this context, two questions often arise: will AI replace human professionals? And what is the relevance of the CFA program in a future environment where AI can perform most technical tasks?[2]
As we have noted elsewhere, we believe the future will be defined by the complementary cognitive capabilities of humans and machines, characterized by the ‘AI + HI’ paradigm and the continued importance of professional competence. To understand what this combination looks like, it is first necessary to assess the current scale of AI adoption in investment workflows, before identifying possible transition paths to future scenarios characterized by different mixes of human and machine interactions.
Current landscape
Early last year, the CFA Institute published a survey-based study: “Creating value from big data in the investment management process: a workflow analysis.” In it, we analyzed the rate of technology adoption for various workflow tasks performed across function categories including consulting, analytics, investment and decision-making, leadership, risk, and sales and customer management.
A key takeaway from this work is that investment professionals adopt a multihoming strategy, using multiple platforms and/or technologies to complete a task. In the Analytical Function category, three example workflows – valuation, industry and company analysis, and research report preparation – illustrate this pattern.
The table shows the share of respondents using different technologies for each of these tasks. Not surprisingly, traditional tools like Excel and market databases are still the most commonly used, but respondents also report integrating tools like Python and GenAI in addition to traditional software. For example, while 90% of respondents reported using Excel for valuation tasks, 20% also reported using Python in this workflow. For analytical roles, GenAI was most commonly used to assist in the preparation of research reports, which was mentioned by 27% of respondents.[3]
Source: Wilson, CA, 2025, Creating value from big data in the investment management process: a workflow analysis: https://rpc.cfainstitute.org/research/reports/2025/creating-value-from-big-data-in-the-investment-management-process.
GenAI in practice: a workflow example
Let’s take a look at an industry and company analysis. When our survey was conducted in 2024, 16% of respondents acknowledged using GenAI in this workflow. Us Automation moving forward content series, in the episode RAG for Finance: Automating Document Analysis with LLMsprovides a concrete example of how GenAI can improve this workflow.
The case study is supplemented with Python notebooks in our RPC Labs GitHub repository. It shows how RAG can extract details about executive compensation and governance from proxy statements of portfolio companies and present the results in a structured table, one of many tasks performed in this workflow.
Such a task is traditionally manual and time-consuming, with the effort required largely determined by the number of portfolio positions. GenAI allows the process to scale efficiently with only marginal additional computing power, eliminating the need for the analyst to perform manual data extraction and tabular comparison setup.
Because the tasks of data extraction and information presentation are outsourced to the GenAI model, the analyst can focus on data interpretation rather than preparation. Rather than crunching the numbers, the analyst focuses on evaluating the output by interrogating the model, checking the validity of the data, understanding the limitations of the analysis, correcting errors, and supplementing the output with additional information or insights from other sources, all with the goal of identifying potential governance risks for all portfolio holdings.
Rather than eliminating the need for a human analyst, this example shows how more value can be unlocked from human input by providing more time and capacity for critical thinking and decision-making. It also illustrates the limitations of AI (such tasks have imperfect accuracy scores) and the continued need for human oversight and judgment.
Evolution
Agentic AI has emerged as a powerful tool that can further improve workflows and deepen human-machine interactions. These tools build on some of the limitations of RAG and include chain of thought reasoning and calling external functions (see our article: “Agentic AI for Finance: Workflows, Tips, and Case Studies“) AI agents expand the scope of tasks that machines can perform and could shape the future direction of human-machine interaction.
Source: Pisaneschi, B., 2025, Agentic AI For Finance: workflows, tips and case studies: https://rpc.cfainstitute.org/research/the-automation-ahead-content-series/agentic-ai-for-finance.
In many ways, this evolution simply extends the multihoming strategy, combining multiple tools and platforms into a single user interface. Claude for Financial Services reflects this approach by connecting to market databases and traditional platforms such as Excel to produce reports and analysis for the user. In this way, AI functions as an application layer on top of other software tools, communicating with the human analyst who retains oversight and responsibility.
Professional judgment remains essential to test assumptions and validate data sources and references. Furthermore, the effective use of these tools also depends on strong fundamental knowledge in finance and investing, which allows analysts to trust and own the outputs of models and maintain a reasonable basis for investment decisions.
Professionals will also need soft skills that cannot be outsourced to machines, including building relationships and exercising duties of loyalty, prudence and care, grounded in ethical values.
Going forward, the CFA Institute will conduct deep research into workflows and skills as AI reshapes the investment profession. While the mix of tasks and the skills required to perform them will undoubtedly continue to evolve, and in ways we may not foresee, we expect that the AI+HI principle will remain the foundation of ethical professional practice and sound investment management.
We invite practitioners to share their thoughts on the shifts in skills and workflows you’re seeing in the comments section.
[1] Our research inventory on AI includes:
AI in asset management: tools, applications and boundaries
AI pioneers in investment management (2019)
T-shaped teams: organizing to apply AI and Big Data to investment firms (2021)
Ethics and artificial intelligence in investment management: a framework for professionals (2022)
Handbook on Artificial Intelligence and Big Data Applications in Investments (2023)
Unstructured data and AI: Refining LLMs to improve the investment process (2024)
AI in investment management: case study in ethics (2024); AI in Investment Management: Case Study in Ethics, Part II (2024)
Creating value from big data in the investment management process: a workflow analysis (2025)
Synthetic data in investment management (2025)
Explainable AI in the financial sector: meeting the needs of diverse stakeholders (2025)
Automation Ahead: Content Series (2025)
[2] See for example Tierens, I., 2025, AI can pass the CFA® exam, but cannot replace analysts
[3] An interactive version of this data is available in our RPC Labs GitHub repository: https://github.com/CFA-Institute-RPC/AI-finance-workflow-heatmap
#Finance #Changing #Workflows #Growing #Demand #Human #Judgment #CFA #Institute #Enterprising #Investor