The breakthrough of generative artificial intelligence is the ability of technology to understand and respond to natural language, leading to the rise of sophisticated AI agents. These agents, whether autonomous or semi-autonomous, have the potential to fundamentally reshape how accountants, auditors, and staff members work, collaborate, and deliver value.
Before diving into the framework, it's important to acknowledge where the market stands. Although AI agents represent transformative potential, current data shows that while 65% of enterprises are experimenting with agents, only 11% have achieved full-scale deployment.
This gap isn't about lacking vision — it's about the practical realities of implementation, data governance, and demonstrating measurable ROI. For accounting professionals, this presents both an opportunity and a call for disciplined, phased adoption.
What are AI Agents and Agentic AI?
Most accountants are familiar with conventional accounting automation which focuses on rule-based approaches and repetitive task automation. There are many differing views on the definition of AI agents. Our view is that AI agents represent a technological leap beyond traditional Robotic Process Automation (RPA), extending their capabilities beyond RPA because of their ability to understand context and autonomously take action. Instead of following a series of pre-programmed tasks, AI agents leverage machine learning, natural language processing and other techniques such as computer vision to understand their environment, make decisions, and execute complex tasks that previously required elaborate human intervention.
AI agents vary significantly in their ability to respond to inquiries based on their complexity and autonomy in executing set goals. As AI agents gain higher levels of sophistication, they are often referred to as Agentic AI. Our view is that Agentic AI refers to advanced AI systems that use AI agents to perceive context and the environment, reason and act independently. For example, you could start thinking about AI agents as your personal accounting assistant.
Consider tax preparation in this workflow: traditional automation might simply copy information from the previous year’s tax forms or populate forms with data from another system. An AI agent, however, can analyze financial data and review tax regulations to identify tax savings opportunities, draft client communications highlighting these opportunities, analyze client responses for concerns or questions, and formulate recommendations for the accountant to review with the client. The evolution from basic automation to ‘intelligence’ represents a fundamental shift in how accounting work could get done.
What are the different capabilities of AI agents in accounting and auditing?
AI agents are the next significant step in practical applications of AI that will be widely embedded into technology. In part, this is due to their flexibility and the wide range of use-cases based on their capabilities, but also because agentic workflows emphasize the partnership between human and machine. The complexity of the AI agent can be increased or decreased based on the experience being optimized.
To better understand the potential use cases for AI agents in accounting and auditing, the KPMG “TACO” framework offers a useful structure to illustrate the four levels of agentic capabilities:
- Taskers handle specific, well-defined operations with single goals. These agents can be used for document classification and data extraction.
- Automators integrate end-to-end processes, managing multiple steps and basic decision-making across entire accounting workflows. These agents are ideal for executing client information requests or semi-automated review processes.
- Collaborators function as AI-based 'teammates' that work alongside professionals, augmenting human capabilities. Supporting tax and audit advisory work and conducting risk assessments are good use-cases for this type of agent.
- Orchestrators coordinate multiple agents, managing sophisticated interactions between AI systems and humans. Think of these as a copilot that allows a professional to interact with tax and audit software with natural language. These agents can assist with advisory, tax preparation, or staff resource allocation based on availability, load, skills, and workflow optimization.
This last category, Orchestrators, facilitates Agentic AI. By coordinating multiple agents and leveraging “multimodal” and “hyperautomation” capabilities, the power of agents expands significantly. Multimodal agents can process and synthesize information from diverse sources — for example, analyzing meeting transcripts, client calls, and emails simultaneously to provide comprehensive insights. In this case, an agent could review a recorded client meeting, extract key requirements, and automatically begin populating relevant sections of tax returns.
Hyperautomation takes this further by connecting AI agents across processes, effectively superseding traditional Robotic Process Automation (RPA). RPA historically offers rule-based automation of tasks, whereas hyperautomation with AI agents provides adaptive, intelligent orchestration across broader workflows. In tax compliance, this might mean an integrated system that automatically monitors regulatory changes, prepares client-specific analyses, and generates customized communications — while still having minimal human oversight.
How can accounting professionals use AI Agents?
As AI agents increasingly handle routine tasks, accounting professionals can focus on higher-value activities that require professional judgment and strategic thinking. This shift represents an enhancement of capabilities rather than a replacement of expertise, especially in three key areas of accounting practice:
- Optimizing the firm. AI agents support a data-driven approach to firm management that provides actionable insights to streamline processes, analyze revenue opportunities and provide additional support for junior staff.
- Elevating the professional. Firms could leverage AI agents to streamline time-intensive tasks — from automating data entry to summarizing complex research — increasing capacity for the professional to focus on higher value tasks and advisory.
- Enhancing client service. Firms can deliver higher quality results for their clients through faster and more accurate engagements, more personalized strategic advice and a proactive approach to compliance. Multimodal agents can analyze client meetings and communications to capture nuanced requirements that might otherwise be missed, while hyperautomation ensures consistent, proactive compliance and more personalized advisory services across the client lifecycle.
Considerations for accounting firms to successfully implement AI Agents
AI agents offer the potential to fundamentally transform professional service and client value delivery in the accounting profession. As these technologies mature, firms that thoughtfully integrate AI agents into their operations will be better positioned to thrive in an increasingly competitive landscape. However, successful implementation requires careful consideration of several critical factors:
- Integration with existing systems: Organizations must carefully plan how AI agents will interact with current systems of engagement and record. Recent industry data shows that 87% of AI projects fail to reach production, with poor data quality as the primary culprit. At Wolters Kluwer Tax & Accounting, we’ve built our agentic strategy on cloud-native architectures and Enterprise Data Lakes to avoid these pitfalls.
- Data security and privacy: As AI agents might require access to sensitive information, robust security protocols and privacy controls are sacrosanct. With risk management emerging as the top concern for enterprises in 2025, firms must ensure compliance with regulatory requirements and have strong operational efficiency.
- ROI and measurable outcomes: While data shows a significant increase in AI-based investments, demonstrating clear ROI remains critical. Crawl-Walk-Run remains an imperative. You could start with Taskers that deliver immediate efficiency gains, then scale to Orchestrators that are based on more proven value.
- Ethical considerations: Clear guidelines for decision-making and appropriate human oversight are critical. Firms must establish frameworks that maintain professional standards and ethical practices in an AI-augmented environment.
- Change management: Success depends on effective communication and training programs that help professionals adapt to transformed ways of working. This includes developing new skills and understanding how to effectively collaborate with AI systems.
The future of accounting lies not in choosing between human expertise and artificial intelligence, but in leveraging AI agents to enhance professional capabilities. Firms that embrace this transformation while maintaining their commitment to professional judgment and ethical standards will define the next era of accounting practice.
