Audit Automation Breakthrough: RPA-Powered Robot Streamlines Confirmation Processes
In a move that could redefine how accounting firms handle one of the most labor-intensive aspects of financial audits, researchers from Chongqing University of Technology have unveiled a new robotic process automation (RPA) framework designed to transform the traditional confirmation procedure. The innovation, detailed in a recent publication in Finance & Accounting Monthly, introduces an intelligent audit robot capable of automating key stages of accounts receivable and bank deposit confirmations—tasks historically bogged down by manual data entry, cross-referencing, and logistical tracking.
The study, led by Professor Cheng Ping and doctoral candidate Qian Tu from the School of Accounting and the Cloud Accounting and Big Data Intelligence Research Institute at Chongqing University of Technology, addresses long-standing inefficiencies in audit workflows. Confirmation procedures, while critical for validating asset existence and ownership, have long been criticized for their time-consuming nature, susceptibility to human error, and inconsistent execution across firms. With audit teams often juggling hundreds or even thousands of confirmations during peak season, the margin for oversight is narrow, and the risk of undetected discrepancies remains high.
Cheng Ping, a recognized authority in digital auditing and RPA applications in finance, argues that the profession is overdue for a technological leap. “Auditors spend an inordinate amount of time on repetitive, rule-based tasks that don’t require deep analytical thinking,” he explains. “This not only increases the likelihood of mistakes but also pulls professionals away from higher-value activities like risk assessment, anomaly detection, and strategic advisory.”
The research team focused on A Accounting Firm, a mid-sized Chinese practice with over three decades of experience and annual revenues exceeding 100 million yuan. Despite its strong reputation, the firm’s confirmation process mirrored industry-wide challenges: reliance on manual data collection from multiple sources, inconsistent verification protocols, and poor tracking of mailed correspondence. These inefficiencies not only slowed down audit cycles but also introduced significant control risks, particularly in verifying the authenticity of responses and detecting potential fraud.
To address these pain points, Cheng and Qian developed a comprehensive RPA-based audit robot model. Unlike general-purpose automation tools, this system is specifically engineered for the nuances of audit confirmations. It operates across three core layers: application, data, and deployment, integrating seamlessly with existing audit software and documentation systems.
At the application layer, the robot leverages a suite of automation components—web scraping, document processing, spreadsheet manipulation, and email integration—to simulate human interaction with digital systems. For instance, when verifying the details of a receivables confirmation, the robot automatically accesses public business databases such as Qichacha (a Chinese corporate information platform), retrieves the official name and address of the debtor, and compares it against the client-provided data. Any discrepancies are flagged and recorded in the audit working papers, allowing human auditors to investigate further.
This capability alone represents a dramatic improvement over current practices. Traditionally, junior auditors would manually visit external websites, copy information, and paste it into Excel sheets—a process that could take hours for a single confirmation and was prone to transcription errors. The RPA robot completes the same task in seconds, with perfect accuracy and full audit trail integrity.
Another major advancement is in the generation and tracking of confirmation letters. Once the sample is selected—either manually by the auditor or through predefined sampling rules—the robot pulls the necessary financial data from the client’s accounting system, populates standardized confirmation templates in Word, and prepares the documents for dispatch. It then generates shipping labels, records tracking numbers, and monitors the delivery status through courier APIs.
The return phase is equally automated. Upon receiving a response, the robot checks the sender’s address against the original recipient, verifies the postmark location, and assesses whether the reply came directly from the third party—key indicators of response reliability. If the return address matches the expected location and no red flags are detected, the result is logged as “no anomalies observed.” If inconsistencies arise—such as multiple responses arriving from the same address or mismatched sender information—the system flags the case for immediate human review.
This level of automated validation significantly strengthens audit quality. As the researchers note, one of the most common weaknesses in traditional confirmation processes is the lack of rigorous follow-up on returned letters. Auditors may assume a response is legitimate without verifying its origin, opening the door to potential manipulation. The RPA robot closes this gap by enforcing strict verification protocols on every single return.
Beyond efficiency and accuracy, the model introduces a structured approach to exception handling. During deployment, the robot is configured with predefined parameters and environmental settings to ensure stable operation. Throughout its execution, it monitors for common anomalies—such as delayed webpage loading or unstructured data formats—and applies corrective actions like retry loops or error logging. For rare or unforeseen issues, the system pauses and alerts the operator, preventing cascading failures while maintaining the integrity of the overall workflow.
Post-execution, the robot archives all processed data, generates detailed logs, and provides real-time feedback to the audit team. This creates a transparent, auditable record of every action taken, enhancing both internal quality control and external accountability.
The implications of this research extend far beyond a single accounting firm. As audit regulations tighten and stakeholders demand greater transparency, firms worldwide are under pressure to modernize their practices. While some have explored blockchain-based solutions or “internet+” platforms for electronic confirmations, these approaches face significant adoption barriers due to infrastructure requirements, regulatory uncertainty, and interoperability issues.
RPA, by contrast, offers a pragmatic, cost-effective alternative. It does not require overhauling existing systems or waiting for industry-wide standardization. Instead, it works within current technological environments, automating discrete tasks without disrupting established workflows. This makes it particularly appealing to mid-tier and regional firms that may lack the resources to invest in large-scale digital transformation.
“RPA isn’t about replacing auditors,” Qian Tu emphasizes. “It’s about augmenting their capabilities. By offloading mechanical tasks to robots, we free up human professionals to focus on judgment, analysis, and client advisory—areas where machines cannot yet compete.”
This human-machine collaboration model is already gaining traction in global accounting networks. Firms like Deloitte, EY, PwC, and KPMG have rolled out their own RPA initiatives, deploying digital workers to handle invoice processing, compliance checks, and tax filings. However, the application of RPA to core audit procedures—especially confirmations—has remained limited, often due to concerns about reliability, control, and professional skepticism.
The Chongqing team’s work directly addresses these concerns by embedding risk controls at every stage of the automation process. Their model doesn’t just execute tasks; it enforces audit standards, documents deviations, and ensures compliance with professional guidelines. This built-in governance framework makes the robot not just a productivity tool, but a compliance enabler.
Moreover, the research highlights the organizational and cultural shifts that accompany RPA adoption. As routine tasks become automated, the demand for entry-level auditors focused on data entry and verification is expected to decline. In their place, firms will need professionals who can design, monitor, and optimize robotic workflows—roles such as RPA solution architects, business analysts, and automation developers.
This transition presents both challenges and opportunities. For individuals, it means acquiring new technical skills and adapting to a more dynamic, technology-driven work environment. For firms, it requires rethinking talent development, performance metrics, and career progression paths. The most successful organizations will be those that view RPA not as a cost-cutting measure, but as a strategic investment in quality, scalability, and innovation.
The study also acknowledges the risks inherent in automation. A primary concern is stability—robots operate based on fixed rules and may struggle with unstructured or unexpected inputs. If a webpage layout changes or a data source becomes temporarily unavailable, the robot may fail unless properly configured to handle such exceptions. This underscores the need for robust testing, continuous monitoring, and clear escalation procedures.
Security is another critical consideration. Confirmation data often includes sensitive financial information, and any breach could have serious consequences. The researchers stress the importance of role-based access controls, data encryption, and secure deployment environments to protect against unauthorized access or data leaks.
Despite these challenges, the overall outlook is optimistic. The audit robot demonstrated in the study achieved significant gains in speed, accuracy, and consistency—key drivers of audit quality. By reducing the time spent on low-value tasks, it allows firms to reallocate resources toward more complex analytical procedures, ultimately leading to more insightful and reliable audit opinions.
Looking ahead, the integration of RPA with other emerging technologies could unlock even greater potential. For example, combining robotic automation with artificial intelligence and machine learning could enable predictive analytics—identifying high-risk accounts for confirmation based on historical patterns, transaction anomalies, or behavioral signals. Natural language processing could allow robots to interpret unstructured responses or handwritten notes on returned letters.
Blockchain technology, though currently difficult to implement at scale, could eventually provide a tamper-proof ledger for confirmation requests and responses, further enhancing trust and transparency. In such a hybrid model, RPA would serve as the operational engine, while AI and blockchain provide intelligence and security layers.
The research also points to broader implications for the future of auditing. As automation becomes more pervasive, the profession must redefine what it means to be an auditor. Technical proficiency in data systems, automation tools, and digital controls will become as essential as knowledge of accounting standards and auditing principles.
Regulators and standard-setting bodies will need to keep pace with these changes. Current auditing standards were written in a pre-digital era and may not fully account for the use of autonomous systems in evidence collection. There is a growing need for updated guidance on how to audit the outputs of robots, validate their configurations, and assess the effectiveness of automated controls.
Educational institutions, too, must adapt. Accounting curricula should incorporate training in RPA, data analytics, and information systems to prepare the next generation of auditors for a technology-rich environment. Professional certifications may need to include modules on digital auditing and automation governance.
For A Accounting Firm and others like it, the path forward is clear: embrace automation not as a threat, but as a catalyst for improvement. The confirmation robot developed by Cheng Ping and Qian Tu is more than a technical novelty—it is a blueprint for the future of audit practice. It demonstrates that with careful design, rigorous controls, and a focus on human-robot collaboration, technology can enhance, rather than undermine, the integrity of financial reporting.
As audit firms around the world grapple with rising expectations, tighter deadlines, and increasing complexity, solutions like this offer a way forward. They promise not only to make audits faster and cheaper, but also more accurate, consistent, and trustworthy.
In an era where financial scandals and audit failures continue to erode public confidence, innovations that strengthen audit quality are not just welcome—they are essential. The RPA-powered confirmation robot is a step in the right direction, proving that when technology and professional judgment work together, the result is a stronger, more resilient financial system.
Cheng Ping, Qian Tu, Chongqing University of Technology, Finance & Accounting Monthly, DOI: 10.19641/j.cnki.42-1290/f.2021.17.012