Agentic AI impact on CPAs

Agentic AI: Transforming the Accounting Profession

Agentic AI is poised to revolutionize the accounting industry by introducing advanced automation and decision-making capabilities. Unlike generative AI, which relies on step-by-step prompts, agentic AI operates autonomously, defining goals, creating plans, and adapting with minimal human intervention. This shift promises to streamline accounting workflows and enhance productivity, allowing professionals to focus on strategic analysis and decision-making.

Randy Johnston, CEO of Network Management Group Inc. (NMGI), highlights that agentic AI uses sophisticated reasoning and iterative planning to solve complex, multistep problems independently. This stands in contrast to generative AI tools, which require detailed user prompts to function effectively. Donny Shimamoto, founder of IntrapriseTechKnowlogies LLC, expects agentic AI to be widely adopted by the end of the year, integrated into various AI tools to support accountants in focusing on higher-value tasks.

Johnston, who will present at the AICPA & CIMA ENGAGE 25 conference, anticipates agentic AI will handle tasks such as customer service, sales support, supply chain management, and IT project management within the next one to three years. This technology is expected to significantly impact the accounting profession by automating routine tasks and enabling accountants to leverage their expertise more effectively.

Agentic AI vs. Generative AI

Both generative and agentic AI use algorithms to simulate human learning and decision-making, but agentic AI is designed to learn and adapt faster. Generative AI responds to specific prompts, such as summarizing meetings or converting reports into presentations. In contrast, agentic AI can handle multistep queries independently, researching tax law changes, creating client presentations, or managing marketing campaigns across multiple channels.

Agentic AI operates through a four-step process: perceiving input information, reasoning to set project goals, acting to complete tasks, and learning through machine learning and feedback. This iterative approach allows agentic AI to improve continuously, making it a powerful tool for streamlining workflows and increasing efficiency in accounting and finance.

Implementation of Agentic AI in Accounting

The integration of agentic AI into everyday accounting practices is expected to be seamless, similar to how AI was incorporated into search engines like Google. Users may not notice the transition but will benefit from more user-friendly interfaces and sophisticated results. Shimamoto predicts that agentic AI will be phased into existing accounting software, enhancing functionality without disrupting workflows.

Byron Patrick, CEO of VERIFYiQ, notes that combining agentic AI with DIY automation tools allows small firms and finance teams to improve processes by easily building automations and integrating agentic AI into their operations. This approach can lead to significant productivity gains and better resource allocation, enabling accountants to focus on strategic initiatives rather than routine tasks.

Agentic AI vs. Robotic Process Automation

Agentic AI addresses the limitations of robotic process automation (RPA), which often failed due to inflexibility and complexity. Unlike RPA, agentic AI can handle multiple tasks simultaneously, acting as an orchestra of decision-making agents. This capability allows for continuous deployment and exponential improvements in workflow efficiency.

Patrick suggests that agentic AI could alleviate the staffing crisis in the accounting profession by taking over time-consuming tasks, increasing productivity, and allowing accountants to focus on more valuable activities. However, he emphasizes that human oversight remains crucial, as AI cannot replicate the creative thinking and strategic decision-making that humans bring to the table.

Challenges and Considerations

While agentic AI holds promise, its adoption requires behavioral shifts and a willingness to accept its capabilities. Patrick notes that understanding and trusting the results produced by agentic AI will be a key hurdle. Although agentic AI is less prone to hallucinations than generative AI, misinterpretations are still possible, and accountants may question the processes behind its decisions.

Shimamoto emphasizes the importance of due diligence when selecting AI vendors, ensuring data security, and compliance with privacy regulations such as the CCPA and GDPR. Organizations must also evaluate the functionality and accuracy of AI products to ensure they meet their intended purposes. Patrick advises using existing processes to vet vendors and requesting SOC 1 or SOC 2 reports for assurance.

In conclusion, agentic AI represents a significant advancement in the accounting profession, offering the potential to automate complex tasks, enhance productivity, and enable accountants to focus on strategic initiatives. As the technology continues to evolve, it will be essential for professionals to stay informed, conduct thorough due diligence, and integrate agentic AI thoughtfully into their workflows.