TLDR: Gartner advises Chief Financial Officers to prepare for agentic AI, a technology capable of autonomous, judgment-based activities, which is set to revolutionize finance operations. With 57% of finance teams already adopting or planning to adopt it, CFOs must establish early governance, understand its limitations, and focus on augmenting human expertise rather than mere automation to leverage its potential benefits and mitigate risks.
Agentic AI is poised to fundamentally transform finance operations, moving beyond traditional automation to perform complex, judgment-based activities autonomously, according to a recent report from Gartner. The research firm emphasizes the urgent need for Chief Financial Officers (CFOs) to proactively prepare for this shift, noting that a significant 57% of finance teams are already implementing or planning to implement agentic AI solutions.
Unlike Robotic Process Automation (RPA), which relies on explicit inputs and predetermined outputs, or even generative AI (GenAI) that responds to user prompts, agentic AI systems are designed to make decisions, solve problems, and act independently. This capability allows them to handle intricate financial challenges with minimal human intervention. However, despite its transformative potential, agentic AI is still in its nascent stages, necessitating careful strategic planning from finance leaders.
Key benefits of agentic AI include its ability to autonomously tackle complex finance problems. Yet, this autonomy also presents challenges. Gartner highlights several limitations, such as reliability issues where systems can ‘hallucinate,’ get stuck in loops, or fail unexpectedly. The scope of these agents is also crucial; they perform best and carry lower risk when assigned specific tasks rather than general-purpose functions. Furthermore, while expanding an agent’s memory can enhance decision-making complexity, it can also lead to confusion and degraded performance.
To navigate this evolving landscape, Gartner advises CFOs to establish early governance approaches, ideally as an extension of existing AI protocols. This includes creating pre-established approved use lists, where sub-function leaders and process owners define which processes AI agents are permitted or prohibited from performing, based on compliance risk and potential financial harm. Human oversight is critical to ensure transparency, auditability, and accountability of the AI’s actions.
Experts like Kris Bennatti from Hudson Labs underscore the necessity of agentic workflows for precise financial queries. He explains that traditional language models often struggle with accuracy for specific financial data, whereas agentic frameworks coordinate multiple specialized models and data sources to deliver verifiable results. This approach moves beyond relying on potentially inaccurate ‘model memory’ to querying robust databases.
Matthew Griffin, speaking at Workday, London, emphasized that agentic AI is designed to ‘team’ with humans, acting as autonomous, action-oriented partners. He encourages organizations to focus on augmenting human experts rather than solely automating tasks, which he believes is the best way to boost productivity and output in a risk-averse manner. Griffin suggests starting with high-impact, low-risk use cases to experiment and learn.
Also Read:
- The Rise of Agentic AI: Engineering Autonomous Systems for Business Transformation
- EXL Recognized as ‘Rising Star’ in Agentic AI Services Amidst Enterprise Transformation
Ed Fitzpatrick, CFO Forum leader at Genpact, and Patrick Villanova, CFO at BlackLine, echoed these sentiments, noting that agentic AI is powering finance operations from automation to true autonomy. They believe that as agentic AI proves its efficacy, its adoption curve will become ‘hyperbolic,’ especially as the overwhelming flow of data cannot be managed through traditional technology alone. This shift marks an inflection point, redefining the finance landscape towards truly autonomous and intelligent operations.
