VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2021)

ISSN(Online): 2581-7280

VIVA Institute of Technology

9th National Conference on Role of Engineers in Nation Building – 2021 (NCRENB-2021)

MachineLearningApplicationsUsed inAccounting and Audits

Rohan Sawant1, Prof. Neha Lodhe2

1
Department of MCA, University of Mumbai, Maharashtra
2
Department of MCA, University of Mumbai, Maharashtra

Abstract: AI is a territory of software engineering that gains from a lot of information, recognizes examples,
and makes expectations about future occasions. In the accounting and auditing professions, Machine Learning
has been progressively utilized over the most recent couple of years. Thusly, this investigation means to Survey
the current Machine Learning applications in accounting and auditing with a focus on Big Four Organizations.
In this study, the AI devices and stages created by Big Four organizations are analyzed by directing a content
investigation. It has been distinguished that Big Four organizations built up a few Machine learning devices
that are utilized for predictable audits coordination and the management, completely automated audits.
Accounting processes such as accounts receivable and accounts payable management, preparation of expense
reports, and risk assessment can easily be automated by AI. For instance, machine learning algorithms can
match an invoice received, decide the right business ledger for acknowledgment, and place it in a payment pool
where a human specialist can inspect and submit the payment request to the payment queue.
Keywords:Accounting,Artificial Intelligence,Auditing, Big four Organization, Machine learning.

I. INTRODUCTION

The technological developments such as AI that enable preparing data quickly and with no mistakes or
inclination are generally utilized in various zones. Hence, the potential for AI calculations to give accountants
and auditors upgraded information investigation is as needs be high.
AI tools designed by humans can perform numerous tasks that can help auditors and accountants. For
example, rather than inspecting information, an element's whole record might be audited with
robotization(Shimamoto 2018).
In this examination, the current machine learning applications developed by Big Four accounting
organizations were analyzed and summed up.

II. INTRODUCTION TO MACHINE LEARNING

AI is a region of software engineering which studies learning computer algorithms that utilization
measurements for deciding examples in colossal measure of information and making exact expectations for
obscure future occasions. Machine learning applications have been utilized in numerous various fields, for
example, schooling, wellbeing, science, and accounting (Dogan & Birant,2021, p.1)
During the 1930s, Thomas Ross worked on a machine that recreated a living animal's behavior, which
can be viewed as the main examination in the zone. Thomas Ross and his prof. Stevenson Smith built up a
Robot Rat that can discover the route through a fake labyrinth. Their examination showed strategies to program
a computer to carry on like creatures or people in the learning cycles (Bhavsar et al., 2017, Chapter 12, p.283).
However, the expression "AI" was presented by Arthur Samuel in 1959 and was characterized as "a field of

1
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2021)
ISSN(Online): 2581-7280

VIVA Institute of Technology

9th National Conference on Role of Engineers in Nation Building – 2021 (NCRENB-2021)

study that gives Computers the capacity to learn without being expressly customized". In 1997, Tom M.
Mitchell gave a more extensive meaning of machine learning as being worried about creating computer
programs that can consequently improve with experience. The two definitions included the possibility of
computers having the option to take in examples and patterns from authentic information and foresee the future
with precision by improving their learning execution. AI means to distinguish the model that gives the most
repeatable and precise expectation of future information. For that reason, calculations are sent to develop
numerical models dependent on example information (preparing information). At that point the exactness of the
prescient capacity of models is assessed (trying information).
AI includes both general and area information. General information isn't dependent on the information
or the undertaking and is connected to the premise of AI, for example, insights, software engineering, or neural
science. Then again, space information is about the field of information like designing, account, wellbeing, or
science. Space information has a fundamental job in AI as it helps planning better datasets (Deng et al., 2020,
p.2).
Since its first introduction, machine learning techniques turned out to be well known and are right now
utilized by numerous organizations. In view of past viewings on a site, you might be recommended a book or a
show that you may like, or your bank may educate you due to a dubious exchange that doesn't accommodate
your typical spending design (Ahmed, 2020, pp.2022-2023). These are a few instances of AI applications that
we face frequently in our day-by-day life. AI can investigate numerous information guides related toward a
given arrangement of circumstances to figure out which ones are important. At that point it can apply the
outcomes to another comparable arrangement of information to anticipate the results(Shimamoto 2018).
Before very long, like numerous different regions accounting and auditing will experience more huge
changes because of machine learning, AI, enormous information use, blockchain, and other mechanical turns of
events, and will be upheld more via mechanization(Türegün, 2019, p.91)

III. ADVANTAGES AND DISADVANTAGES OF APPLYING MACHINE LEARNING

TECHNIQUES
The advantages and disadvantages of machine learning techniques in accounting and auditing are
followings:
Advantages:
• Consistency in decision-making
• Machine-driven activities
• Shorten the info process cycle (IAA, 2017, p.3)
• Increased efficiency and effectiveness
• Accuracy
• Models being incessantly updated with less effort because of constant changes in fraud patterns
Disadvantages:
• moral issues
• Risk of bias and logical errors within the model style
• Potential risks of security and violating privacy laws (Alarcon, Fine & Ng, 2019, pp.3-4)
• Need for an entire, well-balanced, and enormous dataset to assure prediction accuracy (Taniguchi,
Sato & Shirakawa, 2018, p.1)
• High price of development

IV. CURRENT APPLICATIONS OF MACHINE LEARNING IN ACCOUNTING AND

AUDITING

The vast majority of the accounting and auditing errands have a repairman and dreary nature, making
them reasonable for machine learning applications.(Zemankova, 2019, p.149) Routine accounting cycles, for
example, money due and creditor liabilities, the board, arrangement of cost reports, and danger appraisal can
undoubtedly be mechanized by AI. For example, AI calculations can coordinate a receipt obtained with the
connected buy request, decide the right business ledger for acknowledgment, and spot it in an installment pool
where a human laborer can inspect and present the installment solicitation to the installment line (Haq,
Abatemarco& Hoops, 2020, p.8) .

2
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2021)
ISSN(Online): 2581-7280

VIVA Institute of Technology

9th National Conference on Role of Engineers in Nation Building – 2021 (NCRENB-2021)

Due to the benefits of actualizing machine learning techniques, Big Four accounting firms have made
enormous interests in mechanical advancements and built up a few stages or apparatuses that machine learning
and artificial intelligence algorithms (Kokina& Davenport, 2017, p.116). In addition to these stages and devices
created, there are more modest undertakings for accounting reevaluating administrations in which machine
learning tools are utilized for accounting or expense announcement purposes.
As a big four organization follow the latest innovations in the field of accounting and audits
(Zemankova, 2019, p.151), the current applications of machine learning have been reviewed by focusing on the
techniques by big four organizations:

4.1. Earnst and Young (EY)

Machine learning techniques are utilized by EY to break down and separate information from
unstructured information to acquire further audit proof and to dissect enormous datasets for evaluating material
misquote chances because of extortion. AI applications increment the quantity of reports investigated
notwithstanding speed up.

4.1.1 Canvas:
is the main online stage that associates audit experts with their customers and empowers predictable
audit coordination and the management paying little mind to measure, area, or multifaceted nature. Material
gives constant checking of the audit and basic achievements in the audit measure, brisk customization of the
audit approach because of the adjustments in the administrative climate, and moment detailing of discoveries

4.1.2 Helix:
is a worldwide examination stage that is accessible to all audit groups and can deal with information of
any measure and implant the scientific audit approach into the organization's system. The stage dissects diary
passages, income and costs, exchange receivable and payable exercises, stock developments and records,
surveys the danger of a customer's home loan portfolio, and distinguishes any shrouded examples, patterns, and
irregularities. By Helix, instead of investing their energy in information assortment, the inspectors can zero in on
directing an excellent audit.

4.1.3 Blockchain Analyzer:

is increased transparency on the blockchain transactions of customers as help to audit processes.

4.2. PwC

4.2.1 GL.ai:
is a progressive bot that utilizes AI calculations to look at huge measures of information, for example,
transferred exchanges, the clients, the sums, and records in every exchange to recognize likely mistakes or
misrepresentation with no inclination. GL.ai is the principal module of PwC's Audit.ai, also, different modules
are being created to upgrade audit quality, proficiency, and customer administration.

4.2.2 Cash.ai:
is another apparatus that utilizes computerized reasoning and AI procedures to direct a completely
mechanized money audit by dissecting money adjusts, bank compromises and affirmation letters, unfamiliar
trade and monetary state of the bank

4.2.3 Halo:
is an information evaluating innovation that investigates enormous volumes of information and gives
improved danger appraisal. It features designs and strange relations and determines high-hazard exchanges by
graphical visualizations (PwC, The PwC Audit). "Corona for Employee Expenses" identifies peculiarities,
adjusts approaches to decrease costs, and gives a precise grouping of costs for charge purposes

4.3. Deloitte

4.3.1 Argus:

3
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2021)
ISSN(Online): 2581-7280

VIVA Institute of Technology

9th National Conference on Role of Engineers in Nation Building – 2021 (NCRENB-2021)

is Deloitte's first intellectual audit application that utilizes progressed AI procedures to distinguish and
remove key monetary data from electronic records. According to Argus, rather than inspecting, information
removed from the whole populace can be utilized to decide likely dangers, irregularities, and patterns (Deloitte,
Delivering smarter audits - Insights through innovation, 2017, p.6).

4.3.2 Cortex:
is a cloud-based examination application that was at first intended for audit and duty customers yet
now is likewise being utilized for counseling, hazard, and monetary warning administrations. Cortex brings
together information stockpiling, huge information the board, progressed investigation, calculations, models,
also, designs.

4.3.3 Omnia DNAV:

is a computerized arrangement that upsets the audit of protections and speculations by joining
advanced advances, machine learning, information science with examiner judgment to convey great audits.
Omnia DNAV changes customer information and outer merchant information to build up another method of
performing valuations utilizing astute calculations.

4.3.4 Signal:
is a danger investigation programming that inspects openly accessible monetary data to recognize
potential dangers so an extensive hazard appraisal can be given to customers. By Signal, the review groups can
recognize likely dangers of disappointment, material misquote, and other danger factors.

4.3.5 Eagle Eye:

is an apparatus intended to scan the web for distinguishing whether an organization is probably going
to experience monetary pain soon. Conventional observing frameworks audit fiscal summaries, credit moves, or
ledgers. Then again, Eagle Eye searches for early notice signals by gathering immense volumes of information,
investigating and relating them, and perceiving explicit examples before the monetary statement-based trouble
pointers are notices (Deloitte Signal, Empowers more informed risk assessment, 2018, p.1).

4.3.6 Sonar:
is a device that assists with checking rapidly and precisely the data entered by people, for example, a
customer data set containing item portrayal, the VAT rate for every item, product code, the standardized
identification, and some other important data and recognizes marking mistakes in data sets.

4.3.7 Brain Space:

is a self-learning apparatus used to aid legitimate cases. Brain Space utilizes AI and group investigation
to look through unstructured information to recognize what can be utilized as proof for supporting the
customer's safeguard. It additionally gives outwardly engaging organizations to introducing the important
information.

4.4. KPMG

4.4.1 Clara:
is a savvy audit stage that embraces the most recent AI and man-made consciousness answers for
breakdown information, gives important designs, and recognizes dangers and abnormalities.

V.Tools Developed by Big Four Accounting Companies

EY PwC Deloitte KPMG

Tax & LEGAL Sonar

4
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2021)
ISSN(Online): 2581-7280

VIVA Institute of Technology

9th National Conference on Role of Engineers in Nation Building – 2021 (NCRENB-2021)

Audit Canvas GL.ai argus Clara

Helix Cash.ai Signal
Blockchain Halo Cortex
Analyzer
Omnia DNAV
Risk Advisory Eagle Eye
Financial Advisory Brain Space

Deloitte and PwC won the awards for the “Audit Innovation of the year” at the Digital Accountancy
Forum & Awards. It seems that big accounting organizations have already developed many platforms or tools
concerning machine learning algorithms and have integrated them into audit and consultancy processes
(Faggella, 2020).

VI. CONCLUSION

The new innovations, for example, AI, offered access to various methods of achieving assignments,
which likewise influenced the accountancy and auditing professions. Regardless of the downsides, actualizing
machine learning tools has numerous preferences, for example, expanded productivity and adequacy by quicker
information examination, top notch review, mistake decrease, early recognizable proof of dangers, and making
an upper hand. Subsequently, the Big Four organizations have created different instruments and are as yet
extending their arrangement of machine learning projects.

There are projections that 30% of corporate reviews will be performed by AI and computerized
reasoning stages by 2025. This fast change in the calling additionally makes a requirement for moral
administration as far as usage. Also, administrative direction and oversight are likewise required concerning AI
and man-made brainpower devices utilized in the cycles by bookkeeping also, examining organizations.

REFERENCES
[1] Ahmed, A. (2020). “From Data to Wisdom” Using Machine Learning Capabilities in Accounting and Finance Professionals. Talent
Development & Excellence, 12(3), 2019-2036.
[2] Alarcon, J. L., Fine, T. & Ng, C. (2019). Accounting AI and Machine Learning: Applications and Challenges. Pennsylvania CPA
Journal, 2019 Special Issue, 1-10.
[3] Bhavsar, P., Safro, I., Bouaynaya, N., Polikar, R. &Dera, D. (2017). Machine Learning in Transportation Data Analytics. In M.
Chowdhury, A. Apon, & K. Dey (Eds.), Data Analytics for Intelligent Transportation Systems (pp. 283–307). Elsevier Inc.
[4] IIA. (2017). KüreselBakışAçılarıveAnlayışlar: YapayZeka - İçDenetimMesleğineİlişkinDikkateAlınmasıGerekenler, Kısım I.
Retrieved from https://www.tide.org.tr/file/documents/pdf/GPAI-Artificial-Intelligence-Part-I-Revised.pdf (Accessed: 6 Sept. 2020).
[5] Dogan, A. &Birant, D. (2021). Machine learning and data mining in manufacturing. Expert Systems with Applications, (166), 1 -12.
[6] Deloitte (2017). Delivering smarter audits - Insights through innovation. Retrieved from
https://www2.deloitte.com/content/dam/Deloitte/us/Documents/audit/us-audit-smarter-audits-dynamic-insights-through-
innovation.pdf (Accessed: 6 Oct. 2020).
[7] Haq, I., Abatemarco, M. & Hoops, J. (2020). The Development of Machine Learning and its Implications for Public Accounting.
CPA Journal, 90(6), 6-9
[8] Shimamoto, D. C. (2018). Why Accountants Must Embrace Machine Learning. Retrieved from
https://www.ifac.org/knowledgegateway/preparing-future-ready-professionals/discussion/why-accountants-must-embrace-machine-
learning (Accessed: 24 Sept. 2020).
[9] Türegün, N. (2019). Impact of technology in financial reporting: The case of Amazon Go. Journal of Corporate Accounting &
Finance, 30(3), 80-95
[10] . Zemankova, A. (2019). Artificial Intelligence in Audit and Accounting: Development, Current Trends, Opportunities and Threats –
Literature Review. 2019 International Conference on Control, Artificial Intelligence, Robotics & Optimization (ICCAIRO), 148-154
[11] Faggella, D. (2020). AI in the Accounting Big Four – Comparing Deloitte, PwC, KPMG, and EY. Retrieved from
https://emerj.com/ai-sectoroverviews/ai-in-the-accounting-big-four-comparing-deloitte-pwc-kpmg-and-ey/ (Accessed: 8 Oct. 2020).
[12] Kokina, J. & Davenport, T. H. (2017). The Emergence of machine learning: How Automation is Changing Auditing, Journal of
Emerging Technologies in Accounting, 14(1), 110-124.
[13] Munoko, I., Brown‑Liburd, H. L. &Vasarhelyi, M. (2020). The Ethical Implications of Using machine learning in Auditing. Journal
of Business Ethics, 1-36.
[14] Shimamoto, D. C. (2018). Why Accountants Must Embrace Machine Learning. Retrieved from
https://www.ifac.org/knowledgegateway/preparing-future-ready-professionals/discussion/why-accountants-must-embrace-machine-
learning (Accessed: 24 Sept. 2020).
[15] Deng, C., Ji, X., Rainey, C., Zhang, J. & Lu, W. (2020). Integrating Machine Learning with Human Knowledge. iScience, 1-35
(Journal Pre-proof). https://doi.org/10.1016/j.isci.2020.121656.

5
VIVA-Tech International Journal for Research and Innovation Volume 1, Issue 4 (2021)
ISSN(Online): 2581-7280

VIVA Institute of Technology

9th National Conference on Role of Engineers in Nation Building – 2021 (NCRENB-2021)