Transforming Accounts Payable with Agentic AI
As finance leaders strive for enhanced ROI, deploying agentic AI for accounts payable (AP) automation is reshaping the financial landscape. This technology is transitioning manual tasks into autonomous workflows, yielding significant returns for organizations. While the average ROI for general AI projects reached 67% last year, autonomous agents achieved an impressive 80%, illustrating the need for CIOs to rethink their automation budget allocations.
The Shift from Theory to Tangible Results
Agentic AI is now moving beyond theoretical applications to deliver measurable outcomes. Unlike generative tools that merely summarize data, agentic systems execute workflows based on predefined rules and approvals. This shift is partly driven by increasing pressure from boards of directors, as nearly half of CFOs reported demands to integrate AI across operations. Yet, 61% of finance leaders admit that their organizations have only tested custom-developed AI agents rather than implementing them strategically to solve real business problems.
Challenges of AI Experimentation
Many of these AI experiments often fail to deliver value. Traditional AI models generate information or predictions that still require human interpretation. In contrast, agentic systems operate autonomously, integrating decisions directly into workflows. Jason Kurtz, CEO of Basware, emphasizes the urgency for boards and CEOs to see tangible results, stating that “AI for AI’s sake is wasteful.”
Accounts Payable: A Prime Candidate for Automation
In finance, the use of agentic AI is proving particularly effective in high-volume, rules-based environments like accounts payable. Approximately 72% of finance leaders view AP as the ideal starting point. This area is conducive to agentic deployment due to its reliance on structured data, encompassing tasks such as invoice capture, data entry, fraud detection, and overpayment reduction. Such applications are not theoretical; they function with considerable autonomy under the right parameters.
Data Quality and Scalability in Financial Processes
Success in automating accounts payable relies heavily on data quality. Basware leverages a dataset of over two billion processed invoices to provide contextually accurate predictions. This allows the system to distinguish between legitimate discrepancies and errors without requiring human intervention. Kevin Kamau, Basware’s director of product management for data and AI, describes AP as a “testing ground” due to its unique combination of scalability, control, and accountability.
Strategic Approaches to AI Integration
When it comes to acquiring these capabilities, finance leaders face the critical decision of whether to build or buy their AI solutions. Current definitions of “agent” range from simple workflow scripts to complex autonomous systems. In accounts payable, 32% of finance leaders prefer agentic AI integrated into existing software while 20% favor in-house development. In contrast, for financial planning and analysis, the preference slightly shifts in favor of self-built solutions. This divergence indicates that organizations should buy AI for standardized processes while building unique capabilities in-house.
Navigating Governance and Autonomy
Despite the potential benefits, fear of autonomous decision-making presents a barrier to adoption. Nearly 46% of finance leaders hesitate to deploy agents without established governance structures. However, the most successful organizations see governance as a means of evolving their AI deployment. Treating AI agents like junior colleagues, as advised by Anssi Ruokonen, head of data and AI at Basware, allows for trust to develop while keeping essential decision-making in human hands.
Conclusion: Moving Toward Purposeful AI Deployment
As businesses increasingly embrace agentic AI, the focus must shift from unstructured experimentation to strategic, purposeful deployment. Organizations that use agentic AI extensively report higher returns and improved operational efficiency. Leaders who effectively integrate AI into their workflows with diligent governance are poised to replicate the success of early adopters. “Agentic AI can produce transformational results, but only when deployed with purpose and discipline,” concludes Kurtz.
For finance teams eager to explore the possibilities of AI and Big Data, events such as the AI and Big Data Exhibition offer valuable opportunities to gain insights from industry leaders.