The chartered financial analyst (CFA) designation has long been considered one of the most rigorous credentials in finance. The three-part exam, particularly Level III with its essay component, is designed to test not only technical knowledge but also the ability to apply concepts in nuanced, real-world scenarios. For decades, passing the CFA has been viewed as a mark of extraordinary dedication, analytical rigor, and professional capability. Yet new research suggests that artificial intelligence may now be capable of achieving this distinction in minutes—something human candidates typically spend years preparing for.
A joint study conducted by GoodFin, an AI-driven wealth advisory firm, and the NYU Stern School of Business tested 23 leading large language models (LLMs) on mock versions of the CFA Level III exam. The tests included both multiple-choice and essay-style questions. Human graders and AI systems alike evaluated the responses. The results were striking: many of the top-performing models achieved passing scores, with several comfortably clearing the 65% threshold. OpenAI’s o4-mini model scored 79.1%, Google’s Gemini 2.5 Flash earned 77.3%, and Anthropic’s Claude Opus 4 came in at 74.9%. Even newer entrants like DeepSeek and xAI produced passing marks.
For financial professionals, these findings underscore how quickly AI is advancing in fields once thought to require uniquely human skills. Less than three years ago, when ChatGPT was first introduced to the public, early versions of AI systems could handle structured, knowledge-based questions but struggled with the more complex and interpretive essay portion of the CFA. In fact, a 2024 study concluded that while models could pass Levels I and II, Level III remained a barrier because of its emphasis on written analysis.
That limitation may no longer hold. According to the researchers, two key factors enabled this leap in performance. First, newer reasoning models such as o4-mini and Gemini 2.5 Pro are better at decomposing questions into logical steps, which allows them to simulate critical thinking rather than simply retrieving information. Second, the study made use of “chain-of-thought prompting,” a technique that encourages AI to lay out its reasoning step by step, considering multiple angles before arriving at an answer. These two innovations appear to have given models the edge they needed to handle the complexity of essay-style responses.
Anna Joo Fee, CEO of GoodFin, emphasized that the most impressive aspect of the results is that the models passed without specialized CFA-related training. “What’s so exciting is thinking about how much better you can make them when you add human expertise, targeted training, and contextual inputs,” she said. Still, she cautioned that these systems are far from perfect. Accuracy remains uneven, and AI is not adept at interpreting human intent—particularly when clients communicate through subtle cues, tone, or body language. “It will dramatically improve over time,” Fee added, but today’s tools are not yet substitutes for skilled professionals.
For those who have endured the rigors of the CFA process, the study’s findings elicit a mix of respect for AI’s progress and skepticism about what it means for practice. Kate Feeney, a wealth advisor at Summit Place Financial who passed Level III in 2016 after dedicating three hours a day for over a year, was not surprised that AI could now replicate the feat. “I personally had the hardest time with time management while taking the Level III,” she said. “And AI is incredibly fast.” The speed advantage cannot be overstated—what takes candidates hundreds of hours, models now complete in minutes.
Derek Williams, an advisor at Veratis Advisors who earned his CFA in 2023, highlighted the dedication required to prepare. “You can’t really wing it,” he said, recalling the hundreds of study hours that went into his success. But while he recognizes AI’s achievement, he does not believe it signals the end of human advisors. “My experience as a financial planner is that people don’t seek me out for technical knowledge,” Williams said. “Our industry is already moving towards not really relying on a person to do base-level analysis.”
That shift—away from purely technical expertise as the differentiator and toward relational, communicative, and planning skills—is one many advisors are already experiencing. Feeney noted that some aspects of advice delivery may be automated over time. Yet the deeper human elements, like empathy and trust, remain irreplaceable. “You have to be able to effectively communicate with your clients,” she said. “I don’t think empathy and that human component can be replicated right now.”
Mike de Massa, CIO of Forza Wealth Management, expressed a different concern: the potential impact on developing advisors. He drew an analogy between AI and calculators. Just as children who rely too heavily on calculators may never learn their multiplication tables, young professionals who lean on AI may miss critical stages of development. “If all the elementary schools gave kids calculators to learn their times tables, then when they get to algebra class, they’re not going to know what the heck they’re doing,” de Massa said. “AI is like a calculator for CFAs.” In his view, if entry-level analysts bypass the grind of building models, running calculations, and learning financial reasoning through hands-on work, they may lose the foundation that prepares them for higher-level problem-solving.
This worry points to an emerging tension in wealth management: AI has the potential to automate many of the tasks that once served as training grounds for analysts and junior advisors. While that efficiency could lower costs and free up senior professionals for more strategic work, it also risks eroding the apprenticeship model that has long defined career development in finance. For RIAs and wealth management firms, this raises important questions about how to structure training and professional growth in a world where AI can handle many of the technical tasks once assigned to associates.
Still, for most advisors, the idea of AI replacing human professionals seems far-fetched—at least for now. Clients do not hire advisors merely for their ability to pass exams or crunch numbers; they seek guidance, reassurance, and context. They want someone who can listen to their concerns, understand their aspirations, and help them navigate uncertainty. Those are human qualities that no algorithm has mastered.
At the same time, the pace of AI’s progress makes it impossible to ignore. In less than three years, models have advanced from struggling with Level III to achieving passing scores at rates comparable to highly trained human candidates. If these systems continue to improve at the same pace, they could soon outperform even the most diligent professionals on technical exams. The implication is not that advisors will be replaced, but rather that the definition of value in advisory services will continue to evolve. Technical expertise will remain necessary, but it will no longer be sufficient on its own. Advisors who differentiate themselves will do so through communication, empathy, and the ability to integrate technology into holistic client service.
For wealth advisors and RIAs, the study offers both reassurance and a challenge. On the one hand, AI tools may enhance efficiency, support research, and streamline tasks that once consumed hours of valuable time. On the other hand, the findings reinforce the need to double down on uniquely human skills and to rethink how junior talent is trained in an increasingly automated environment.
Ultimately, AI’s ability to pass the CFA exam does not diminish the value of the credential or the professionals who earn it. If anything, it underscores the evolving landscape of knowledge work in finance. Passing a test is not the same as guiding a family through a generational wealth transfer, helping a business owner navigate liquidity events, or managing the emotions of a client facing market volatility. The CFA charter will continue to represent a deep reservoir of technical knowledge—but for advisors, the true measure of success lies in how that knowledge is applied in human relationships.
The CFA exam has long been a crucible for the financial profession. Now, AI’s success in clearing its hurdles highlights a turning point. For the next generation of advisors, the challenge will be not only mastering the technical foundations but also cultivating the distinctly human qualities that clients value most. As AI grows stronger, the role of the advisor becomes clearer: not to compete with machines on speed or recall, but to bring judgment, empathy, and wisdom to the heart of financial decision-making.
