[Editor’s Note: Ardent Partners recently published its Accounts Payable-themed report, “The State of ePayables 2025: AP’s Unfinished Journey.” Over the next several weeks, this site will feature our series highlighting key AP management strategies and top priorities for 2025.
AI is Here. That’s Clear. Start Using It!
While AI adoption use within AP is in the early stages, there is a strong view that it will have a major impact on operations over the next few years. A solid majority of AP leaders expect AI to deliver either a significant (42%) or transformational (23%) impact over the next two or three years (see Figure 9). This level of enthusiasm should continue to drive greater investment in ePayables solutions in the near term, particularly as the types of use cases for AI become more tangible and as more groups start using them.
What AP Hopes to Gain from AI
While AP leaders are optimistic about the potential of AI, their expectations remain grounded in practical goals (see Figure 10). The top objective, cited by a significant margin (65%), is to increase automation across AP processes. This reflects an ongoing desire to streamline the entire operation. Beyond automation, interest fragments across several areas. A cluster of secondary priorities includes improving forecasting and predictive capabilities (39%), reducing departmental budgets (35%), enhancing stakeholder satisfaction (30%), and strengthening cash management (26%). Each of these reflects a broader ambition to apply AI to key points of leverage, yet none of these goals commands widespread support. Surprisingly, better decision-making ranks lowest on the list of AI objectives. This suggests that many AP teams still see AI primarily as an operational efficiency tool rather than a strategic enabler and stands in contrast to the current view of AI’s overall impact.
Ardent Partners 2025 research on AI within procurement identified several obstacles to AI usage and optimization, which include a lack of budget to invest in more resources, data quality and access, and employees lacking skills to use AI properly. These are likely shared by their AP counterparts today and indicate that while AI is a promising technology, organizational readiness must be 23% addressed to fully realize its potential.
SIDEBAR
An AI Primer
Artificial intelligence (“AI”) adoption within AP operations is still in its early stages, with most organizations applying it in targeted, tactical ways rather than as a fully integrated capability. The path of AI within AP over the next few years will not be linear, but the direction is clear. AI will become a core part of procurement operations, so it is important to start investigating it, and using it with controls.
The ABC’s of AI
AI is a complex technology that can be looked at from different perspectives and classifications, such as general, narrow, or application AI (e.g., computer vision, speech recognition, robotics, expert systems, etc.). Another way of classifying AI, which has more relevance to this report, is by AI technique. Within procurement technology, machine learning, deep learning, and natural language processing (NLP) are perhaps the most important today.
- Machine Learning is a type of AI that learns from data to make predictions or decisions without being explicitly programmed. In AP, it can be used for invoice coding and data extraction, exception detection, and payment timing optimization.
- Deep Learning is a more advanced form of machine learning that finds complex patterns in large datasets using layered algorithms. It is often used behind the scenes in tools that handle tasks like OCR enhancement on poor quality documents, enhanced anomaly detection, or generating insights from unstructured data.
- Natural Language Processing (NLP) helps computers understand and work with human language in both written and spoken formats. In AP, NLP can be used to support automated vendor query response, PO text matching, and voice-enabled, mobile AP dashboards.
The newer AI innovations, such as generative AI (GenAI) and Agentic AI, build on these technologies.
Generative AI
Generative AI is a type of artificial intelligence that creates new content like text, images, or data based on patterns it has learned from large datasets. In procurement, it can help create written summaries for exceptions, invoice reviews, and general reports as well as drafting internal communications and process documentation. A common form of generative AI is the Large Language Model (LLM), which focuses on language tasks, such as summarizing documents, answering questions, or generating text. These tools respond to written prompts, so knowing how to give clear instructions is key to getting useful results. However, GenAI can sometimes produce information that sounds right but is inaccurate because the system lacks relevant training data or misinterprets the prompt. For AP teams, this means results should be reviewed carefully, especially when accuracy matters.
Agentic AI
Agentic AI refers to a more advanced type of artificial intelligence that can take action and make decisions on its own to achieve a goal. Unlike traditional tools that wait for instructions, these AI “agents” can plan, adjust, and use other systems to get tasks done with minimal human input. In an AP context, Agentic AI could eventually deliver autonomous correspondence to suppliers and stakeholders, self-service invoice help agents, and multi-system process coordination. It combines several advanced technologies, including generative AI, to operate more independently.
