For years, business pundits have predicted the demise of the accounting profession as technology replaces work traditionally performed by the industry. According to The Hackett Group, since 2004 the median number of full-time employees in the finance department at large companies has declined 40 percent, to about 71 people for every $1 billion of revenue.
The CPA profession acknowledges that the accounting skill set is changing, but the need for specialists in financial reporting, auditing, tax, and advisory services remains strong. For example, Wolters Kluwer N.V. uses Oracle’s Hyperion software to close its books in half the time it used to. “With fewer workers needed to collect financial information, Wolters Kluwer is hiring more analysts to help sift its data on profits, revenue, and cash flow, among other things, to help in planning and forecasting.”1
Many fear that machines will replace people. That might be the case for routine accounting tasks, such as bank reconciliations and invoice payments, but new opportunities have emerged. As artificial intelligence becomes more accepted in the workforce, it can be an essential partner. Artificial intelligence will not replace CPAs; rather, their duties, responsibilities, and contributions will change for the better. Artificial intelligence will eliminate certain day-to-day tedium, freeing up CPAs to work where human attentiveness is needed, adding value to the business.
We are at a point where accounting automation is the major technological innovation that will impact everyone working in the profession. This article explores components of this technology model and assesses factors limiting its current use, opportunities for the future, and skills needed for those working with innovative technology.
Defining Accounting Automation
In Accounting Today, Therese Tucker describes robotic process automation (RPA) as a system to capture, manipulate, and interpret transactional data flowing from myriad information technology (IT) systems and applications.2
RPA reduces the time spent performing repetitive tasks, and linking it to data analytics and artificial intelligence opens up a world where machines have the potential to augment (and perhaps replace) the work of humans. For CPAs, the key to a successful integration of accounting automation, or robo-accounting, will be a robust understanding of processes and the necessary internal controls.
What is commonly known as artificial intelligence really is divided into three concepts:
- Artificial intelligence (AI) – The broader concept of machines making decisions and performing “smart” tasks that normally require human intelligence. This includes learning, reasoning, and self-correction.
- Artificial superintelligence (ASI) – If AI is the ability of computers to mimic human thought, ASI goes a step beyond, where a computer’s cognitive ability is superior to a human’s.
- Machine learning – An application of AI based on the idea that when we give a machine access to data, it can continue improving its performance without human intervention.3
Robo-accounting is the combination of RPA, ASI, and machine learning.
Service companies are devoting resources to harness the power of big data. The input of data through optical character recognition (OCR) is transforming the preparation of tax returns via tax preparation software. Computer-assisted audit tools and techniques (CAATT) software, such as IDEA and ACL, give auditors the ability to access large databases and perform audit testing/functions at blazing speeds. This type of data analysis is being incorporated into software to provide continuous auditing of transactions with AI, reporting inconsistencies and exceptions to audit teams and management. Blockchain also has the promise to ensure the integrity of financial transactions, thus reducing or eliminating the need for traditional audit verification. Here are a few examples of some of the new software options:
- PwC and H2O.ai – This ongoing GL.ai project uses AI and machine learning to analyze billions of data points in milliseconds, to see what humans can’t, and to apply judgment to detect anomalies in the general ledger. PwC says the general ledger module examines every uploaded transaction, every user, every amount, and every account to find unusual transactions without bias or variability. And the more it is used, the smarter it gets.4
- KPMG and IBM – In its partnership with IBM, KPMG has brought to market two cognitive solutions that automate the extraction of lease contract data for financial reporting and that identify research projects eligible for credit subsidies for tax reporting. The systems use the IBM Watson Explorer.5
- Intuit – For the small-business market, Intuit has joined the AI bandwagon with some basic applications, including Expense Finder, Auto Categorization, and Mileage Tracking. Intuit claims to have approximately 100 patents pending related to machine learning and artificial intelligence.6
- Blockchain – This financial transaction system and global ledger technology connects millions of computers and servers. Transactions are grouped into a block, and that block is then sent out to the global network. In the global network, the block is time-stamped and added to the chain in chronological order, thus the term blockchain. Blockchain could reduce the need to maintain bookkeeping functions because the platform will maintain and track transactions on a real-time basis. That would free CPAs to work on other important tasks, such as valuation and planning, that can provide more business opportunities for companies. The structure of blockchain makes modifying recorded transactions difficult. Blockchain offers the assurance that underlying transactions are valid, exist, and are accurate and complete. When applying AI to a trusted database, the audit risk of a material misstatement is minimized in an environment of strong internal controls, reducing inherent risk of the data, and near 100 percent testing control detection risk.
Tax preparation software packages do not guarantee 100 percent accuracy for those tax practitioners who use OCR to input client data. In many cases, accuracy depends on the types of scanner, OCR software, and tax software compatibility. There are not many clients who are willing to accept 80 percent accuracy in preparation of their return. The tax code also adds to the complexity for data input with exotic credits, deductions, and revenue recognition rules.
When it comes to blockchain use, there are multiple platforms and protocols for blockchain transactions, and no platform has yet emerged as the dominant system. Thus, no technical or process standards are universally used.7
The trend today is for developers to look for industry solutions tied to a specific platform. Sandro Psaila, an audit and assurance manager with Deloitte, related a story of blockchain failure: “Although blockchain promises highly secure transactions, fraud instances cannot be fully eradicated. In July 2017, an unknown hacker managed to steal nearly 32 million U.S. dollars’ worth of Ethereum, one of the most popular virtual currencies. The root cause of this fraud was not related to deficiencies in the blockchain technology, but rather due to a vulnerability within the software that was used to manage Ethereum wallets. The fraud was quickly detected, and related parity vulnerability was mitigated accordingly to safeguard the remaining wallets.”8
Meeting the Challenge
During the May 2017 meeting of the Public Company Accounting Oversight Board’s (PCAOB) Standing Advisory Group (SAG), it was noted that, according to Deloitte’s 2016 Global Impact Report, many accounting firms are making significant investments in technology and new data analytic tools, which they have asserted could enable them to analyze large quantities of data more quickly and intelligently and to enhance the audit by automating time-consuming tasks that are currently manual and rote in nature.9
SAG noted that the PCAOB inspections staff has an ongoing project to understand the systems of quality control that firms have in place to provide assurance that the tools used to analyze the data meet the audit objectives; engagement teams are effectively using these tools and evaluating the results of screening large data populations; and engagement teams are applying due care, including professional skepticism, when using these tools during the performance of the audit work, including the evaluation of results of that work.10
Current auditing standards make it very clear that CPAs cannot avoid responsibility for understanding highly technical processing controls over financial statement reporting. PCAOB AS 2201, An Audit of Internal Control Over Financial Reporting that Is Integrated with an Audit of Financial Statements, addresses the use of others’ work in the performance of the audit (Section 19): “The extent to which the auditor may use the work of others in an audit of internal control also depends on the risk associated with the control being tested. As the risk associated with a control increases, the need for the auditor to perform his or her own work on the control increases.”11
That technical understanding of robo-accounting and its related controls are clearly placed with the CPA firm under PCAOB AS 1201, Supervision of the Audit Engagement. Section 05 states: “The engagement partner and, as applicable, other engagement team members performing supervisory activities should inform engagement team members of their responsibilities, including matters that could affect the procedures to be performed or the evaluation of the results of those procedures, including relevant aspects of the company, its environment, and its internal control over financial reporting, and possible accounting and auditing issues.”12
Opportunities to Gain Credibility in an Automated Accounting Environment
Auditors must be knowledgeable about the technologies employed by their clients, and understand the types of internal controls to prevent a material misstatement. One of the best ways to establish this level of expertise is to pursue a recognized technology credential. The AICPA established the Information Management and Technology Assurance (IMTA) section to support members who offer assurance services and information management support for their clients and decision-makers. Tied to this new section, the AICPA offers the certified information technology professional (CITP) credential, which acknowledges CPA professionals who receive additional training in the areas of emerging trends, security and privacy, business solutions, IT assurance and risk, and data analytics.13
To meet the CITP credentialing standard, candidates must do the following:
- Hold a valid and unrevoked CPA license, active permit, or certificate issued by a legally constituted state authority (the CPA license must also be in an active status).
- Pass the credential exam. At this point, the CITP exam is a four-hour, computer-based exam composed of multiple-choice questions. The standard AICPA-member fee to take the CITP exam is $400.
- Complete the online credential application.
- Attest to meeting the minimum business experience and education requirements, and pay the appropriate credential fee. The standard, initial CITP fee is $360. A CITP candidate must have a minimum of 1,000 hours of business experience in information management and technology assurance within the five-year period preceding the date of the CITP application. Credential candidates must complete 75 hours of precertification education within the relevant credential body of knowledge, based on the definition of continuing professional development. All hours must have been obtained within the five-year period preceding the date of the application.
Another well-recognized organization for developing and training in IT controls is ISACA. This association engages in the development, adoption, and use of globally accepted, industry-leading knowledge and practices for information systems. ISACA led the development of COBIT 5, the business framework for governance and management of enterprise IT. This organization offers five IT certifications: certified information systems auditor (CISA), certified in risk and information systems control (CRISC), certified information security manager (CISM), certified in the governance of enterprise IT (CGEIT), and cybersecurity nexus (CSX).14
Arguably, the skill sets of the CISA and CRISC are most critical to the CPA auditor:
- The CISA is a recognized certification for information systems audit, control, and security professionals. To be eligible, five or more years of experience in information systems audit, control, assurance, or security is required.
- The CRISC certification is designed for those experienced in the management of IT risk, and the design, implementation, monitoring, and maintenance of IS controls. To be eligible, three years of work experience managing IT risk by designing and implementing IS controls, including experience across at least two CRISC domains (of which one must be in Domain 1 or 2) is required for certification.
Accounting automation is evolving to combine RPA, data analytics, and AI. It will impact all areas of accounting, including financial statement preparation, audit, tax preparation, and advisory services. At this point, there are no universally recognized standards and protocols for the software being developed. Major players are devoting significant resources to move the field forward with open source software to attract innovation and widespread acceptance. Some aspects of robo-accounting clearly are under the purview of management, but CPAs can apply this technology in practice management, audit, tax preparation, and advisory services. It will be critical for the CPA profession to gain recognition as experts in the internal control systems surrounding these technologies. That expertise can be gained with additional education, experience, and credentialing.
1 Bernard Marr, “Machine Learning, Artificial Intelligence – and the Future of Accounting,” Forbes (July 2017).
2 Therese Tucker, “Robots Are Not Accountants,” Accounting Today (March 15, 2018).
3 Marc Staut, “Rise of the Robo-Accountant: Artificial Intelligence and Emerging Technologies,” (Oct. 6, 2017). http://news.cchgroup.com/2017/10/06/rise-robo-accountant-artificial-intelligence-emerging-technologies
4 “Harnessing the Power of AI to Transform the Detection of Fraud and Error,” PwC Global. https://www.pwc.com/gx/en/about/stories-from-across-the-world/harnessing-the-power-of-ai-to-transform-the-detection-of-fraud-and-error.html
5 Ranica Arrowsmith, “KPMG Offers New IBM Watson-Enabled Accounting Tools,” Accounting Today (March 15, 2018). https://www.accountingtoday.com/news/kpmg-offers-new-ibm-watson-enabled-accounting-tools
6 “Machine Learning: Unlocking the Power of Millions for the Prosperity of One,” QuickBooks Online Team, The QuickBooks Blog. https://quickbooks.intuit.com/blog/innovation/machine-learning-unlocking-the-power-of-millions-for-the-prosperity-of-one
7 Eric Piscini, Darshini Dalal, David Mapgaonkar, and Prakash Santhana, “Blockchain to Blockchains: Broad Adoption and Integration Enter the Realm of the Possible,” Deloitte Insights, Tech Trends (Dec. 5, 2017).
8 Sandro Psaila, “Blockchain: A Game Changer for Audit Processes,” Deloitte Perspectives.
9 “The Use of Data and Technology in Audits,” PCAOB, Standing Advisory Group Meeting (May 24-25, 2017). https://pcaobus.org/News/Events/Documents/05242017-SAG-meeting/DTA-Briefing%20Paper%20May-2017.pdf
11 AS 2201: An Audit of Internal Control Over Financial Reporting That Is Integrated with an Audit of Financial Statements. https://pcaobus.org/Standards/Auditing/Pages/AS2201.aspx
12 AS 1201: Supervision of the Audit Engagement. https://pcaobus.org/Standards/Auditing/Documents/PCAOB_Auditing_Standards_as_of_December_15_2017
Paul R. Brazina, CPA, CGMA, CFF, is an assistant professor of accounting at La Salle University in Philadelphia. He can be reached at email@example.com.
Yusuf J. Ugras, PhD, is associate professor of accounting at La Salle University. He can be reached at firstname.lastname@example.org.