The promise of a fully automated finance function remains largely unrealized for many mid-market businesses. A new report indicates that while interest in accounts payable (AP) automation is high, widespread adoption is hampered by cost, integration challenges, and a lingering uncertainty about return on investment. The findings highlight a critical gap between the potential benefits of automation and the practical realities of implementation for companies with annual revenue between $20 million and $499 million.
According to a survey conducted in September 2025 by San Francisco-based AP automation firm Ottimate, only 4% of mid-market finance leaders reported having fully automated their accounts payable processes—meaning a seamless workflow “from invoice to payment with no manual touchpoints.” The survey, which included 225 respondents across the healthcare, retail, and hospitality sectors, paints a picture of a sector largely stuck in “partial automation,” where technology assists with some tasks but still requires significant human intervention. This incomplete automation, the report suggests, may be hindering the realization of substantial cost savings.
Nearly half—48%—of those surveyed said they’ve seen “little to no cost savings” from their current AP automation tools. Ottimate attributes this to the prevalence of partial automation, where tasks are split between technology and manual processes. A significant 89% of respondents reported utilizing partial automation, while 7% admitted to not using any automation at all. “Partial automation may seem like a step in the right direction,” Ottimate wrote in its report, “But it actually sets the stage for some of the key challenges finance teams face every day.”
The challenges are substantial. Half of the respondents are managing more than 5,000 invoices each month, and 38% are taking five or more days to process a single invoice. While a five-day processing time might not seem alarming in isolation, Ottimate points out that it quickly becomes unsustainable when multiplied across thousands of invoices. This slow processing speed, coupled with increasing invoice volumes, creates a strain on finance teams and increases the risk of errors and missed payments.
The demand for efficient AP processes is further underscored by the rising threat of fraud. Recent reports indicate that new artificial intelligence tools are being leveraged by fraudsters, increasing the sophistication of attacks. The Ottimate survey found that four in ten respondents had experienced either invoice fraud or overpayment in the last year. In response, mid-market firms are relying heavily on manual controls, with 52% citing manual invoice reviews as their primary fraud prevention measure and 50% requiring multiple approvals before releasing payments.
However, these manual processes are becoming increasingly vulnerable as fraud techniques evolve. “As AI evolves, fraud is getting more sophisticated and harder to detect,” the Ottimate report warns. “The manual processes and disconnected systems that define partial automation make organizations especially vulnerable.”
The Barriers to Full AP Automation
Despite the clear benefits of full automation—reduced fraud risk, faster processing times, and potential cost savings—several factors are hindering adoption. Cost remains a significant obstacle, with half of the survey respondents citing it as a primary concern. Difficulty integrating new automation tools with existing systems was as well a major hurdle for half of the respondents.
A substantial 40% of respondents expressed an “unclear return on investment” as a reason for delaying full automation. This hesitation is not unfounded, as the effectiveness of some newer technologies, particularly generative AI, is still being evaluated. A recent report by Workday found that almost half of the time saved using AI is spent correcting faulty outputs, raising questions about the net benefit of these tools.
The Promise and Peril of AI in Finance
While the ROI of some AI applications remains uncertain, other AI-powered solutions are showing promise in the finance sector. Finance-specific functions offered by Anthropic’s Claude model, for example, have garnered attention, though they also come with inherent risks, such as potential outages and data security concerns.
Finance leaders are increasingly scrutinizing the potential of AI across various functions, weighing the benefits against the risks. The challenge lies in identifying solutions that deliver tangible value without introducing new vulnerabilities or complexities. The current landscape suggests that a cautious, phased approach to automation—one that prioritizes integration and demonstrable ROI—is likely to be the most successful strategy for mid-market firms.
Looking Ahead
The path to full AP automation for mid-market businesses is proving to be longer and more complex than initially anticipated. The combination of cost concerns, integration challenges, and the need for a clear return on investment is slowing down adoption. As AI technology continues to evolve, finance chiefs will undoubtedly continue to evaluate its potential, but a pragmatic approach—focused on addressing existing pain points and demonstrating measurable results—will be crucial. The next year will likely see a greater emphasis on solutions that seamlessly integrate with existing systems and offer a clear path to cost savings and fraud reduction.
What are your thoughts on the future of AP automation? Share your experiences and insights in the comments below.
