1108440

research-article2022
AUM0010.1177/03128962221108440Australian Journal of ManagementSeethamraju and Hecimovic

Article

Australian Journal of Management

Adoption of artificial intelligence in

2023, Vol. 48(4) 780­–800
© The Author(s) 2022
Article reuse guidelines:
auditing: An exploratory study sagepub.com/journals-permissions
DOI: 10.1177/03128962221108440
https://doi.org/10.1177/03128962221108440
journals.sagepub.com/home/aum

Ravi Seethamraju and Angela Hecimovic

Discipline of Accounting, The University of Sydney Business School, The University of Sydney, Sydney, NSW, Australia

Abstract
Artificial intelligence (AI) is likely to transform the audit profession. This exploratory study,
by utilising the technology–organisation–environment (TOE) framework and data from semi-
structured interviews, identifies several technological (perceived benefits, compatibility, maturity),
organisational (readiness, data quality, trust) and environmental (audit standards, regulation, client
readiness) factors influencing the adoption of AI tools in audit practice. While AI has the potential
to improve audit quality and deliver value-adding services to audit clients, our study concludes
that AI adoption requires a rethink of audit practice considering the perceived lack of control in
AI ‘black-box’ potentially rendering audit practice even more hidden from view and exposed to
increased scrutiny over audit quality.
JEL Classification: M42, O31, O32, O33

Keywords
Adoption, artificial intelligence, audit, audit quality, technology–organisation–environment
framework

1. Introduction
Transformational technologies such as artificial intelligence (AI) have the potential to revolution-
ise auditing (Association of Chartered Certified Accountants (ACCA), 2019). Traditionally,
auditing is a late adopter of new technology; however, its labour-intensive nature, compliance
burden and competitive pressures make it a prime target for automation (Issa et al., 2016). With
the Big4 accounting firms (i.e. PwC, EY, KPMG, Deloitte) currently investing US$9 billion in AI
(Kapoor, 2020), it is anticipated that 30% of financial audits will be performed using AI by 2025

Corresponding author:
Ravi Seethamraju, Discipline of Accounting, The University of Sydney Business School, The University of Sydney, Sydney,
NSW 2006, Australia.
Email: [email protected]

Final transcript accepted 1 June 2022 by Sue Wright (AE Accounting).

Seethamraju and Hecimovic 781

(World Economic Forum (WEF), 2015). Australian Big4 accounting firms employ over 28,860
employees, many contributing to generating combined audit revenues of $1.561 billion in 2020
(Australian Financial Review, 2022), with firms actively leveraging AI technology in a quest to
deliver more cost-efficient and high-quality audits.
The adoption of information technology (IT) tools in auditing is not new as audit firms have
utilised electronic data processing, computer-assisted audit techniques (CAATs) (Kend and
Nguyen, 2020) and expert systems (Vasarhelyi et al., 2005). Enabling automation of structured
audit tasks such as sending electronic confirmations to debtors has contributed to improved effi-
ciency and audit quality (Rozario and Vasarhelyi, 2018). While these tools have saved auditors’
time, their adoption across audits is not extensive with evidence that 50% of auditors have never
used CAATs (Pedrosa et al., 2020), other than Excel spreadsheets (Bakarich and O’Brien, 2021;
Khan et al., 2021). AI tools incorporating technologies such as robotic process automation (RPA)
and machine learning (ML) while perceived to be transforming current audit practices, empirical
research examining their adoption in audits is limited. Audit research relates to discussions (Issa
et al., 2016; Kokina and Davenport, 2017) and further research calls (Duan et al., 2019; Sutton
et al., 2016), while studies in healthcare claim trust and performance expectancy are key factors in
AI adoption (Fan et al., 2020; Prakash and Das, 2021).
This exploratory qualitative study answers these calls by investigating the factors influencing
the adoption of AI specifically in audits. Using the technology–organisation–environment (TOE)
framework (Tornatzky and Fleischer, 1990) as a broad theoretical anchor for data analysis, we
investigated the influence of technological, organisational and environmental factors on the adop-
tion of AI in auditing. We conducted 24 semi-structured interviews with audit partners and manag-
ers, standard setters, IT specialists and managers from professional bodies.
We identify several technological (perceived benefits, compatibility, maturity), organisational
(readiness, data quality, trust) and environmental (audit standards, client, regulatory readiness) fac-
tors influencing the adoption of AI tools in auditing and extend the TOE framework to new innova-
tion and domain context. The findings have implications for the audit profession, standard setters
and in understanding the antecedents and challenges of AI adoption.
Next section provides an overview of the literature on prior technologies in auditing. Section 3
outlines the research method, and Sections 4 and 5 provide the findings and conclusion,
respectively.

2. Literature review – technologies in auditing

Audit firms have deployed several IT tools such as CAATs and expert systems in audits since the
1960s (Abou-El-Sood et al., 2019). Vendors such as CaseWare developed standardised CAAT
tools, including ‘Interactive Data Extraction and Analysis’ software, automating several manual
tasks to enhance audit efficiency and effectiveness. CAATs have assisted auditors with tasks such
as audit sampling, clients’ internal control testing (Pedrosa et al., 2020) and fraud risk interrogation
(Byrnes et al., 2018). Expert systems implemented in the 1990s incorporate some basic elements
of intelligence to support audit planning, risk assessment and compliance testing, mainly in the
early phases of the audit process (Dowling and Leech, 2014). Integrated audit support systems,
adopted by the Big4, help manage the entire audit process by facilitating the display of audit tasks
based on client information. The use of expert systems ended in the 1990s (Gray et al., 2014) with
attempts at embedding AI technologies into audit support systems (Sutton et al., 2016).
Despite technological capabilities assisting auditors in performing audit tasks, the audit phases
(shown in Table 1) have remained relatively unchanged and are guided by auditing standards (e.g.
Australian Auditing Standards (ASAs)).
782 Australian Journal of Management 48(4)

Table 1. Overview of the audit process and various phases.

Phase Typical activities undertaken in phase

Pre-engagement Accept/reject new client or assessment of continuation with existing client.
Ethical requirements as per APES 110 Code of Ethics for Professional Accountants.
Terms of engagement and issue engagement letter (ASA210).
Planning and risk Obtain knowledge of the business (ASA315).
assessment Perform analytical procedures (ASA520).
Risk assessment includes fraud (ASA240) and going concern (ASA570).
Materiality assessment (ASA320).
Control environment evaluation (ASA315).
Develop overall audit plan in response to risk assessment (ASA330).
Determine extent and nature of audit testing (ASA330).
Audit evidence Tests of controls (ASA500).
Evaluation of internal controls and modify audit approach (ASA330).
Substantives testing of transactions, balances and disclosures (ASA500).
Obtain representations from management, bank, solicitors (ASA580).
Opinion Review financial report (ASA520).
formulation and Review audit results (ASA450).
reporting Review subsequent events (ASA560).
Review fraud and error (ASA240).
Review-related party transactions (ASA550).
Formulate audit opinion includes Key Audit Matter (various ASA ‘700s’ standards).
Continuous Supervise and review work (ASA220).
activities Review client relations (ASA220).
Significant deficiencies in internal controls (ASA265).
Material errors or irregularities (ASA450).
Consider going concern (ASA570).

Source: Adapted from Martinov-Bennie et al. (2016).

ASA: Australian Auditing Standard.

While ASAs make some reference to CAATs, its adoption was relatively low for several rea-
sons. These include lack of confidence (Braun and Davis, 2003) and auditor skills (Ahmi and Kent,
2013), implementation costs and technical difficulties (Byrnes et al., 2018; Mahzan and Lymer,
2009), poor client system performance, lack of technical and organisational infrastructure
(Bierstaker et al., 2014) and auditors’ preference for Excel (Pedrosa et al., 2020).
Audit practice has generally lagged behind business in technology adoption due to the complex-
ity of audit tasks, multiple structures of source documents, client data availability, IT readiness and
the requirement of professional auditor judgement (Cohen and Rozario, 2019; Salijeni et al., 2019).
Auditors’ requirement to exercise professional judgement utilising their accounting, assurance,
industry knowledge and experience, while simultaneously drawing on their integrity, objectivity
and professional scepticism to evaluate audit evidence, and whether accounting treatment has been
applied correctly by client management is integral to audit practice (Harding et al., 2016).
Professional scepticism – the questioning mindset and attitude of the auditor – is essential to mak-
ing sound judgements, particularly being alert to conditions indicative of financial errors/fraud
(Trotman et al., 2011).
AI-focused systems differ from previous CAATs technology in trying to automate cognitive
tasks in the audit process that require professional judgement (Cohen and Rozario, 2019). AI in the
auditing context involves the adaptive use of various tools that incorporate technologies such as
RPA and ML that enhance audit tasks (Kachroo et al., 2020).
Seethamraju and Hecimovic 783

While there is no commonly accepted definition of AI, we adopt the definition by Duan et al.
(2019), that AI is ‘the ability of a machine to learn from experience, adjust to new inputs and per-
form human-like tasks’ (p. 63). It is therefore argued that AI can accelerate auditors’ work not only
by enabling automation of tasks such as risk evaluation and 100% population testing but by also
improving judgement on audits (Munoko et al., 2020; Rozario and Thomas, 2019). Other potential
benefits include reduction in human errors and bias, facilitation of continuous auditing capacity,
ability to track patterns of data and the potential to lower audit costs (Brennan et al., 2017).
However, some caution the importance of understanding human capabilities, judgements and
skills, and at what point advanced technologies such as AI can reduce cognitive overload (Wiggins
et al., 2020).
The Big4 are currently building and deploying different AI tools to develop niche specialisa-
tions and to gain a competitive advantage in providing audit services (Ting and Vasarhelyi, 2018).
KPMG is currently trialling ‘Clara’, an intelligent audit platform combining data analytics and ML
techniques to scan information, model it against thousands of assumptions to highlight risks the
auditor can utilise in judgements such as evaluation of management’s debt provisions (KPMG,
2020). However, we are not there yet in terms of AI deep learning, where the application ‘Clara’
can ‘think’ for itself and make judgements such as which audit opinion will be issued in the report-
ing phase (Bakarich and O’Brien, 2021). PwC is developing ‘GL.AI’ that identifies abnormalities
in clients’ general ledgers and is currently trialling a complete AI audit of cash accounts (PwC,
2020). Similarly, Deloitte and EY are currently utilising a natural language processing technology
to analyse unstructured contractual information, used in both the planning and evidence audit
phases (Deloitte Insights, 2019; EY, 2018).
Potential factors influencing AI adoption include limited knowledge of AI-enabled auditing
and legal implications of relying on AI-based tools (Issa et al., 2016) and trust in technology
(Fan et al., 2020; Janssen et al., 2020; McKnight et al., 2011; Prakash and Das, 2021). Data
security and lack of guidance from auditing standards (Issa et al., 2016; Manita et al., 2020)
are also concerns. Working with AI requires a different skillset, and this is already influencing
hiring practices of Big4 firms, demonstrating a preference for individuals with fundamental
accounting knowledge coupled with programming/data management skills (Cooper et al.,
2019). Goh et al. (2019) argue that AI has the potential to enhance auditors’ technical skills
including coding, freeing up accountants to work on value-adding tasks such as promoting
best practice IT controls to audit clients. Others argue that AI cannot mimic the unique capa-
bilities of auditors, such as professional judgement and professional scepticism (Issa et al.,
2016), as its ability is limited to its programmed logic (Sidhu, 2019). Manita et al. (2020)
highlight the need to change audit standards by integrating guidance on emerging technologies
such as AI.
Recent research (Bakarich and O’Brien, 2021; Kend and Nguyen, 2020) acknowledges that
despite the potential benefit of AI to the audit process, the auditors’ requirement to understand their
clients’ business, its risks, compliance with standards, compilation of audit evidence and exercis-
ing professional scepticism is a long way off being replaced by AI audit technologies. Furthermore,
algorithmic bias potentially built into AI tools and contextual factors like – audit tasks, individual,
organisational and environmental – could all influence the adoption of AI-enabled tools (Kordzadeh
and Ghasemaghaei, 2021). Following this discussion and the limited empirical evidence, this study
aims to answer the following research question:

•• What are the technological, organisational and environmental factors affecting the adoption
of AI in auditing?
784 Australian Journal of Management 48(4)

3. Theoretical framework and methodology

3.1. Theoretical framework
To help understand the technology adoption and its effects on firms, several explanatory frame-
works have been developed. TOE, one of the most widely applied frameworks proposed by
Tornatzky and Fleischer (1990), is considered appropriate to understand the adoption of new tech-
nologies such as AI (Oliveira and Martins, 2011).
TOE framework considers the technological, organisational and environmental contexts and
helps in the identification and integration of different factors that influence new technology adop-
tion. The technology context in this framework refers to the technology available to the firm for
adoption while the organisational context relates to internal factors. The environmental context
reflects the external environment surrounding the firm in which it conducts its business (Tornatzky
and Fleischer, 1990).
Each technology innovation has specific attributes and characteristics, and different contexts of
adoption that cannot be considered by a generic model (Cruz-Jesus et al., 2019). A summary of the
studies that have adopted the TOE framework in different technology contexts (most recent for
each technology) and the significant factors presented in Table 2 demonstrate empirical support it
received (Clohessy and Acton, 2019; Furneaux and Wade, 2011; Li et al., 2018). TOE framework,
as illustrated in previous studies, allows researchers to comprehend the broader picture in which
innovation occurs without specifying factors in each (Mishra et al., 2007; Oliveira et al., 2019),
incorporating different factors influencing adoption into their respective contexts (Bose and Luo,
2011; Venkatesh and Bala, 2012).
A broad interpretation of the definitions of factors in studies listed identifies perceived benefits,
security, maturity and compatibility as the common technology-related factors, while management
support, organisational readiness, financial resources and data quality as the other significant
organisational factors influencing technology adoption. In environmental context, regulations,
external pressures and standards are significant. Factors, their definition and classification vary
according to technology context, deeming it necessary for the factors to be tailor-made for each
context (Cruz-Jesus et al., 2019). The exploratory nature of our study and the unique characteristics
of AI technology in auditing allow for the TOE framework to be utilised as an appropriate theoreti-
cal lens to identify the relevant TOE factors and their influence on AI adoption by audit firms.
Our application of this broad framework is guided by an understanding that adoption of AI is an
intentionally driven change (Weick and Quinn, 1999) and the views of decision-makers (such as
audit partners, technology specialists) reflect their attitudes, perceptions, intentions and motivation
to drive the adoption. As this framework is free from industry and firm size restrictions and does
not require a specific technology context and domain (Puklavec et al., 2018), it is useful for catego-
rising the forces contributing to AI adoption and identifying the sources of inertia arising from any
one of its three broad contexts.

3.2. Data collection and analysis

Semi-structured interviews were undertaken with a range of stakeholders, including seven audit
partners across each of the Big4 professional service firms, with an average of 17 years’ experience
between them in providing audit services. In addition, seven managers predominantly from the
Big4 and one from a mid-tier firm were interviewed, with most having strong experience in using
audit technology on client engagements. Insights were also gained from managers from both a
standard setter (3) and professional bodies (3), pertinent to the study given these bodies are
Table 2. Key technology adoption studies using TOE framework and significant factors.

Author(s) Technology Technological factors Organisational factors Environmental factors

Abed (2020) Social commerce Perceived usefulness, Management support, organisational Consumer pressure, partner
security readiness pressure
Clohessy and Acton Blockchain Compatibility, maturity, Management support, size, Regulations, partner
(2019) perceived benefits organisational readiness support, industry pressure
Cruz-Jesus et al. Customer relationship Technology competence, Management support Competitive pressure
Seethamraju and Hecimovic

(2019) management data quality and integration

Oliveira et al. (2019) SaaS Technology competence Management support External pressures
Li et al. (2018) Audit analytics Technological competence Management support Standards, Professional help
Puklavec et al. (2018) Business intelligence Cost, BI system part of ERP Management support, project External support, size,
(BI) systems champion, organisational readiness industry
Xu et al. (2017) ERP systems Complexity, relative Management support, organisation Competitive pressure
advantage, compatibility fit, financial commitment
Martins et al. (2016) SaaS Relative advantage Technology competence, Coercive pressures,
management support normative pressures
Seethamraju (2015) SaaS ERP systems Perceived benefits, Organisational readiness, costs, data Partner support
complexity, compatibility quality, integration
Lian et al. (2014) Cloud health info Data security, costs, Management support, resources, Industry pressure,
system compatibility technical competence, CIO government policy
innovativeness
Lin (2014) Electronic supply Perceived benefits, costs Firm size, management support, Competitive pressure
chain mgt system absorptive capacity
Bose and Luo (2011) Green IT Technology competence Technology readiness, resources, size Regulations, competition
Furneaux and Wade IS discontinuance System (technology) Costs, integration
(2011) capability, compatibility
Hossain and RFID technology Complexity, compatibility, Resource-based factors, External pressure, external
Quaddus (2011) cost, standards, trailability management-based factors support, uncertainty

TOE: technology–organisation–environment; ERP: Enterprise Resource Planning; IT: information technology; CIO: Chief Information Officer; RFID: Radio Frequency
Identification; BI: Business Intelligence; SaaS: Software as a Service; IS: Information Systems.
Definition of factors and their classification into the TOE framework differs from study to study (e.g. costs) as shown in the above table and literature does not offer
consistent definitions. While this is the strength of the TOE framework, it is also a limitation as the interpretation is contextual with no consistency among various studies.
785
786 Australian Journal of Management 48(4)

Table 3. Details of interview respondents.

Label reference Respondent profile

Auditor 1 Manager with IT expertise in data analytics, working across various audit
teams (3+ years).
Auditor 2 Director of global audit at Big4 accounting firm with IT audit innovation
experience (18+ years).
Auditor 3 Senior audit manager at mid-tier accounting firm (6+ years).
Auditor 4 Non-executive director (4+ years) with prior CFO and Big4 partner
experience (20+ years).
Auditor 5 Senior manager auditing at Big4 accounting firm (7+ years).
Auditor 6 Partner at Big4 accounting firm with data assurance experience (14+ years).
Auditor 7 Senior manager auditing at Big4 accounting firm with data assurance
experience (6+ years).
Auditor 8 Manager auditing at Big4 accounting firm with data assurance experience (5+
years).
Auditor 9 Senior manager auditing at Big4 accounting firm with data assurance
experience (7+ years).
Auditor 10 Senior audit partner and director at Big4 accounting firm with strong IT
experience (17+ years).
Auditor 11 Audit partner at Big4 accounting firm (30+ years).
Auditor 12 Associate auditor in financial services team at Big4 accounting firm (2+
years).
Auditor 13 Director at Big 4 working in treasury (10+ years) and previously auditor for
5 years.
Auditor 14 Senior audit partner at Big4 accounting firm with financial services
experience (20+ years).
IT Specialist 15 Director of audit and accounting technology at a leading audit technology
firm (10+ years).
IT Specialist 16 Senior data scientist and audit team member at Big4 accounting firm (4+
years).
IT Specialist 17 Technology adoption consultant with renowned AI technology startup
experience (6+ years).
IT Specialist 18 Internationally renowned expert in audit technology and data analytics (40+
years).
Professional Body 19 Senior manager at an accounting professional organisation (8+ years).
Professional Body 20 Senior manager and knowledge leader at an accounting professional
organisation (13+ years).
Professional Body 21 Manager, reporting and assurance with policy and assurance compliance
experience (5+ years).
Standard Setter 22 Senior manager of standard setting board (6+ years) and previously Big4
senior manager (7+ years).
Standard Setter 23 Education director at an auditing standard-setting body (10+ years).
Standard Setter 24 Project manager in standard setting (4+ years) and previous assurance and
advisory in large accounting firm.

IT: information technology; AI: artificial intelligence; CFO: chief financial officer.

responsible for issuing audit standards and policy guidance, respectively. Interviews were also
conducted with four IT specialists, three with extensive experience in audit technology and one
consultant with expertise in AI technology start-ups (see Table 3). Semi-structured interviews
Seethamraju and Hecimovic 787

Technical Organisaonal Environmental

Perceived benefits Audit firm readiness Audit standards
Compability data quality & integraon Regulaon readiness
Maturity Trust Clients' readiness

AI Adopon

Figure 1. Key factors shaping adoption of AI in audit practice.

(Lillis and Mundy, 2005) are appropriate for this study, given the use of AI tools in the audit profes-
sion is relatively new.
A loosely structured interview guide, while ensuring that all three contexts of the TOE frame-
work are covered, allowed flexibility to pursue participants’ responses. After obtaining ethics
clearance, 24 interviews were conducted (average duration 39 minutes), with 16 completed by both
researchers and 4 each individually (see Table 3 for summary respondent profiles). The majority of
respondents are leaders in the audit and assurance field and/or have experience with existing and
emergent audit technologies.
Two pilot interviews were conducted to establish and test interview protocol and questions (Yin,
2009), and a thematic analysis of the data was carried out independently by both researchers using
the three key broad contexts – technological, organisational and environmental, and reconciled.

4. Findings
In this section, we present our findings derived from the analysis of interview data using the TOE
framework contextualised to the adoption of AI in auditing. We found several significant factors
within each of the three contexts as exemplified below (see Figure 1 summary).

4.1. Technology context

As shown in Table 2, different factors were found to be significant in different technology adoption
contexts. While perceived benefits and compatibility were significant drivers in the adoption of the
blockchain (Clohessy and Acton, 2019) and SaaS (Software as a Service) ERP (Enterprise Resource
Planning) systems (Seethamraju, 2015), our data revealed the maturity of the AI technology also
significant.

4.1.1. Perceived benefits. Confirming previous research (Bakarich and O’Brien, 2021; Issa et al.,
2016), perceived benefits including cost efficiency is a significant driver for AI adoption. Our
study suggests the adoption of AI would make the audit process more efficient and effective by
automating certain tasks, enabling auditors to make better judgements, and thereby facilitating
better-quality audits.
AI tools could enable auditors to audit 100% of a clients’ dataset and contribute to a reduction in
audit time and costs (Bizarro and Dorian, 2017). In addition, AI could replace manual and repetitive
788 Australian Journal of Management 48(4)

audit tasks during the planning/risk assessment phase of the audit including ‘identification of risky
parts of business from board minutes and AI will give you the most efficient way of testing samples’
(Auditor 5), ‘reconciling the organisation’s structure with auditors’ databases’ (Standard Setter 24)
and in providing ‘deeper data sets to assist auditors in making judgement on risk areas of clients’
operations’ (Auditor 5). Our data suggest with the adoption of AI, it is possible to carry out a ‘truly
independent random selection of the relevant transactions and deliver a better outcome’ (Auditor
12), notably in the evidence-gathering phase. Further outcomes in this phase include efficiencies in
terms of ‘extracting outliers for evaluation’ (Auditor 3), ‘whole population testing, automatic con-
firmations circulation and data extraction for three-way testing’ (Standard Setter 24, Auditor 13),
‘reconciliations’ (Auditor 2), ‘analytical procedures’ (Auditor 5) and ‘handling of audit inquiries’
(Auditor 11).
While AI may utilise repetitive processing and algorithms to ‘learn from audit evidence and
select evidence without bias’ (IT Specialist 18), auditors in our study believe their judgement is
essential throughout the audit process to evaluate, interpret and communicate insights to their cli-
ents (Auditor 5, 6, 7, 11, 12).
The ability to expand the capability of audit firms beyond financial statements audit (IT
Specialist 18, Auditor 9, 11, Professional Body 19) and into continuous auditing to help investors
as well as clients (IT Specialist 18, Auditor 9) is another benefit. An IT Specialist suggests, ‘what
we need is assurance products that inform investors and management on what’s happening . . .
with AI tools this can be achieved’ (IT Specialist 18).
Furthermore, audit firms are planning to use AI to showcase their capability and offer a new
value proposition to their clients by the promise they ‘will keep bringing new technology, new
thinking, new tools and new processes’ (Auditor 9). Some accounting firms, however, are willing
to charge additional fees with a promise of delivering quality audits as a Big4 audit partner
observes: ‘we can also charge because we are using those AI tools . . . if the fees were to be
$500,000 based on auditor’s hours last year . . . implementation of AI tools we would have to
bump it by $200,000’ (Auditor 14). In contrast, another audit partner (Auditor 11) claims it is ‘pos-
sible that clients may seek fee reduction’, as AI will bring savings in labour costs to the audit firm.
Currently, Big4 firms appear to be using AI initially in their consulting services rather than in
financial audits which in contrast is subject to more regulation. As observed by a manager, ‘Big4
in the US are trying to move into this space with their advertised investment in AI, and they don’t
want to be on the back foot but a lot of them are used in consulting’ (Standard Setter 24), given ‘in
consulting the clients will pay’ (Standard Setter 22).
Respondents also wondered whether the AI adoption should be focused on efficiency (to save
audit costs) or on effectiveness. A Big4 audit partner illustrates this:

you can either say the audit tool can do a sample of 200 in 10 hours, whereas it took a person 100 hours,
so more efficient tool. Or it still takes 100 hours, but able to look at the entire population . . . you don’t
save any time, but you get more effective as you have looked at more in the same time. (Auditor 11)

Adoption of AI, our study found releases auditors from mundane repetitive tasks mainly in the
planning and evidence phases, making some phases more efficient. It may also expand the capabil-
ity of audit firms into continuous auditing and other consulting services, enabling them to show-
case their technology readiness and innovative attitude to clients. While practitioners believed AI
technology adoption would add value to the client rather just a tool to reduce audit costs, others felt
audit firms would be using AI to offer high-value consulting services to its clients and/or charge
higher audit fees.
Seethamraju and Hecimovic 789

4.1.2. Compatibility. Several respondents recognised the need for the compatibility of AI tools with
the existing audit tools and accounting systems for effective adoption, supporting prior research
into the adoption of other technologies such as cloud health information systems (Lian et al., 2014).
Generic AI tools specially designed for specific audit tasks currently available in the market, our
study found, are not interoperable. As noted by a mid-tier senior manager, ‘it’s almost like an app
. . . you might have it on Apple, but when you move it across to Android, you’ve got to change the
format completely’ (Auditor 3). Unsurprisingly, when the AI tool implemented is compatible, it
complemented other audit tools such as CAATS, Excel macros and added value to specific audit
tasks claimed Big4 practitioners (Auditor 2, 6). Unfortunately, as claimed by a partner with strong
data analytics experience (Auditor 6), ‘AI tools available now are not readily compatible with the
data residing in the different accounting systems’ that are owned and managed by audit clients and
require data cleansing. This is resource-intensive as a partner (Auditor 14) observed, ‘you can burn
a lot of your audit time . . . just trying to get that information . . . you’ve killed ten percent of your
audit fee just trying to get something because the tools are not compatible’.
AI, if it is designed and deployed to be compatible with back-end ERP systems, could make data
transfers easier; however, an audit partner asserts there are limitations: ‘all these ERP systems have
some process automation, but somebody has to enter the information . . . even with AI, someone
needs to program it’ (Auditor 10). Given audit firms have diverse clients with different accounting
systems, a generic AI tool if developed could be deployed across several audit engagements, which
is economical and would especially appeal to mid-tier firms with limited resources. However, a
Big4 partner (Auditor 11) revealed his firm is developing an exclusive AI-enabled analytics tool
for auditing a large financial institution, indicating a need to also tailor a specific AI tool that could
be utilised on one single large engagement and likewise be economical for the Big4 firm.
AI tools must be compatible with the audit firm’s audit support systems and clients’ accounting
systems and align with their AI investment strategy. At present, the lack of compatibility of the AI
technology tools, given the nascent stage they are in, is impeding the adoption.

4.1.3. Technology maturity. Every new technology brings uncertainty and risk along with potential
performance improvements. A technology can only be considered mature if it is in use for long
enough (Tapscott and Tapscott, 2016), becomes integrated with products and processes (Ernst,
1997) and most of its initial challenges are overcome. The low maturity of AI technology currently
available for auditing is observed to be hindering adoption. As noted by a partner, ‘the tools on
offer are perceived to be immature and therefore the client acceptance (to use in their engagement)
is an issue’ (Auditor 6). This lack of maturity demands constant interaction with technology pro-
viders and audit clients. As pointed out by an audit partner, ‘the hardest issues is – having to learn
this information; going back to the provider and asking for help on how you get the information
into a usable format’ (Auditor 14). Because they are in an experimental stage, it involves regular
refining of the tools in consultation with stakeholders as another partner adds: ‘we’re struggling
. . . we’ve been looking at this [tool] and trying to work on it as a trial . . . sometimes it works,
sometimes it doesn’t’ (Auditor 10).
Although technology is available in proprietary and open-source formats to perform cognitive
tasks in auditing such as Deloitte’s tool in evaluating complex contractual documents (Deloitte
Insights, 2019), without AI knowledge and training, there are questions about their efficacy. As
noted by a Big4 manager, ‘they [senior partners] see the sales pitch of plug and play and then ask
“why isn’t this working? Why aren’t I getting a return on this?”’ (Auditor 9). Even though audit
firms would like to use this technology across all clients, management is unsure of the investment
value, especially with mid-tier firms. As noted by a manager, ‘we’d like to put it across all clients
. . . a bit of comfort if we can look at certain transactions, but we haven't seen the value yet’
790 Australian Journal of Management 48(4)

(Auditor 3). Current AI audit technology is still immature, and its potential value is insufficient to
accelerate adoption. Audit firms, particularly mid-tier, are hesitant to adopt, due to this lack of
maturity, specifically with the deployment of generic tools rather than in-house tools.

4.2. Organisational context

In addition to organisational readiness and data quality that were found to be significant in the
adoption of business intelligence systems (Puklavec et al., 2018) and SaaS ERP systems
(Seethamraju, 2015), our data identified trust in AI technology also as significant.

4.2.1. Organisational (audit firm) readiness. Organisational readiness refers to the availability of finan-
cial and human resources, and the infrastructure in the organisation to adopt new IT innovations
(Wiggins et al., 2020). It includes a firm’s IT infrastructure, the commitment of financial resources to
new IT innovations and employees with the requisite knowledge and skills, as antecedents for new
technology adoption (Zhu and Kraemer, 2005). Our study found this ‘readiness’, a significant ante-
cedent for adoption, confirming previous research on the adoption of CAATs (Vasarhelyi et al., 2015).
While some noted resourcing and cost of technology are important considerations to embrace
adoption, the importance of investment return was significant. As noted by a manager, it is perhaps
‘one of those pieces of technology that’s not that expensive in the long run’ (Professional Body 20)
with a senior partner acknowledging that ‘any initial costs of deploying AI tools could be recouped
through efficiencies and/or additional fees’ (Auditor 5); however, ‘partners are wanting a return
straightaway . . . they’re not early adopters . . . they need to see it working and they need to know
how it’s working’ (Auditor 3).
There are also concerns about transferring data into AI tools and upskilling auditors to interpret
results, especially for the mid-tier firms as stated by an audit manager:

within our mid-tier network, there might be four IT auditors at most . . . just not a market for it at the lower
end . . . the cost, the investment, the training that puts pressure . . . if we don’t need to do it for the
regulator, why spend money on it? . . . if clients aren’t asking for it, there’s no point. (Auditor 3)

Big4 are developing specialist AI tools jointly with technology partners for their large clients
and positioning themselves in the market. As a Big4 partner observes, ‘we go well beyond the
ordinary part of our value proposition to the client at the beginning of the engagement . . . a kind
of a marketing tool to offer extra value’ (Auditor 6). The Big4, in displaying their optimism to
embrace new technologies, are setting up separate units with the necessary IT infrastructure and
manpower to develop AI solutions for specific audit tasks and large client applications. A Big4
manager states, ‘we have the digital lab . . . it changes how auditors think . . . we can always put
in a request and the lab builds a digital app’ (Auditor 5).
Other firms are relying on their global affiliations, establishing multidisciplinary teams com-
prising of auditors and experts in AI to develop AI solutions. A director of an audit risk committee
adds, ‘we have a push towards AI and analytics right now globally, so we can reduce a lot of physi-
cal work efforts from auditors . . . make everything more efficient and the clients are buying this
approach’ (Auditor 4). In terms of skilled human resources, Big4 firms are boosting their capability
through training as affirmed by a senior partner in financial services:

probably $200 million investment globally, putting lots of people through a year of dedicated technology
projects . . . we’re training more than a couple of hundred people . . . next five years we will end up with
a very scaled-up ability. (Auditor 14)
Seethamraju and Hecimovic 791

Mid-tier firms, on the other hand, would like to purchase off-the-shelf tools developed by tech-
nology firms for specific audit tasks and cannot afford to invest in bespoke AI tools as the follow-
ing quotes illustrate:

we hear the Big4 have developed in-house resources . . . we’re [mid-tier firm] more reliant on commercial
providers to come forward with the tools (Auditor 3), and

an IT expert agrees:

Big4 have a lot of capital . . . they produce the audit tool that supports their way. . . . Mid-tiers don’t have
that kind of money, they need a ready-made software solution. (IT Specialist 18)

While the Big4 are investing in infrastructure and resources, some firms are constrained by their
limited understanding of AI capability, as an IT specialist claims: ‘our client base is independent
audit firms . . . they want an AI tool, but they don’t know what all these things are and how it’s
going to benefit them’ (IT Specialist 15). A mid-tier senior manager who previously worked in a
Big4 claims that neither technology providers nor their auditors are ready and that support from
regulators is an important prerequisite for AI adoption: ‘requires education, support from regula-
tors and commercial software providers . . . we don’t have that understanding and skills now’
(Auditor 3).
Lack of organisational readiness will have implications on audit practice with the automation of
repetitive audit tasks relieving low-level audit staff, providing some extra audit capacity. However, to
assign them to tasks that require greater judgement requires other skills and expertise, which junior
staff generally do not possess. A Big4 auditor, referring to the potential deskilling effect, cautions,

AI in audit would limit junior staff from going up . . . it literally strips them of their roles and how would
they [firms] be able to promote staff who haven’t gone through this tick and bash process . . . critical for
junior staff to touch and feel invoices. (Auditor 12)

Automating low-level tasks while leaving complex judgement to senior staff would result in
loss of learning opportunities for junior staff and may lead to deskilling. This aligns with Wiggins
et al.’s (2020) view that while new technologies can alleviate cognitive overload particularly in
new employees, there is also a risk they are no longer needed.
Our study suggests that while Big4 are investing in training, skills development and develop-
ment of AI tools and getting ready to apply more AI in their audits, mid-tier firms are struggling
due to their limited understanding of AI capabilities, lack of skills, non-availability of commercial
off-the-shelf AI tools and incapacity to invest.

4.2.2. Data quality and integration. Data quality and ease of integration with new technology tools
is a significant factor in technology adoption, our study observed, confirming the previous studies
in the adoption of business intelligence systems (Puklavec et al., 2018) and SaaS ERP systems
(Seethamraju, 2015). Integration and data quality, however, are dependent on the level of digitisa-
tion, as pointed out by audit practitioners below:

data and its quality is the challenge . . . getting the data, holding it, owning it, managing the data, scrubbing
the data, dealing with the irrelevant noise (Professional Body 20), and

clients are slow to invest in digitisation and data quality . . . still have piles of paper on their desks.
(Auditor 6)
792 Australian Journal of Management 48(4)

Furthermore, bias can be injected into the AI tool if the data quality is poor and inaccurate,
which in turn may influence human judgement as claimed by a mid-tier manager: ‘AI is such a new
field and people are discovering that removing bias in data sets is a challenge’ (Auditor 3). When
clients’ data quality is not reliable, this risk is transferred to audit firms. Furthermore, security and
privacy of the data that is stored in the databases of commercial providers and used in the develop-
ment of AI tools is a concern as a standard setter points out:

we noticed some struggles to have access to the right data . . . what are you doing with your data, what sort
of security are you having around your data? (Standard Setter 24)

For the data to be easily transferrable between the audit firms’ IT infrastructure, audit clients’
accounting systems and AI tools, quality of data and its ability to integrate with other systems is
more important than data security, claims an audit technology director:

getting data from their clients, that’s the biggest hurdle . . . had to spend millions of dollars on a global
level . . . getting the data out of their accounting systems can either be horrible, unreadable . . . takes
forever to format. (IT Specialist 15)

In our study, data quality from the audit client and the client’s ability to integrate with other
audit tools is found to be a significant antecedent for AI adoption. Client’s hesitancy to invest in
digitisation and data quality may give rise to unwanted consequences such as biased output, pri-
vacy and security issues – with all these risks transferred to the audit firm.

4.2.3. Trust. Auditors’ reluctance to trust AI tools is another factor limiting adoption. Trust in tech-
nology refers to individual beliefs about specific technology’s capability and the infrastructure that
supports it (McKnight et al., 2011). Our findings support recent studies in healthcare (Fan et al.,
2020; Panicker et al., 2016), and this lack of trust emanates from the respondents’ belief that AI
tools would not be able to support tasks that require auditor judgement and scepticism. In addition,
there are concerns in respect to legal implications of financial auditing work (Issa et al., 2016), lack
of transparency in the development and workings of AI tools (Miller, 2019), and potential inherent
bias (explained in section 4.2.2), and the general reluctance of accountants to move from tradi-
tional tools such as Excel (Schmidt et al., 2020).
A manager claims, ‘how can partners take ownership of everything the AI is producing and
sign-off if you don’t know how it is being produced’ (Professional Body 21). Similarly, another
questioned, ‘how do you audit the black-box to make sure the data is okay to use’ (Standard Setter
24). A Big4 auditor elaborates, ‘AI wouldn’t have sufficient experience to work efficiently and be
professionally sceptical’ (unlike people), therefore suggests for ‘things that involve high judge-
ment, high estimation, there should be a lot of restrictions in AI use; we don’t want to depend on
AI completely’ (Auditor 12).
As observed by a senior audit manager, ‘more people would be more hesitant towards it . . .
because it is an unknown’ (Auditor 5). Auditors’ lack of trust in AI tools may also stem from their
remoteness from the development. As a Big4 manager states, ‘I don’t think people will accept an
order put through a machine . . . real risk because we’re using tools we didn’t build’ (Auditor 8).
As an IT specialist also opines,

an AI tool that does a task fast and doing all the stuff that otherwise a human being would be . . . it is a sort
of loss of control, with no knowledge of what drives the decisions, who are drivers . . . reasoning is
missing. (IT Specialist 18)
Seethamraju and Hecimovic 793

Our study found that inherent lack of technical skills and knowledge to confidently work with
AI tools (as explained in Section 4.2.1) further accentuates auditors’ lack of trust. Auditors are
required to interpret the knowledge produced by AI tools using professional judgement with their
knowledge and expertise. This professional auditor judgement, for example, is particularly impor-
tant in the audit of intangible assets, where auditors are required to evaluate management’s assump-
tions and challenge conflicting information (Griffith et al., 2015).
Pointing out the need for the AI tools to be transparent (to enable trust), a partner comments that

it’s more a tool that allows you to be intelligent . . . (when) it maps out a complete view so that we can
understand it rather than letting the software make decisions about what to look at . . . you’re still making
your judgement. (Auditor 7)

An IT Specialist concurs,

. . . once auditors understand kind of what’s happening inside the AI black-box, then I think being an
auditor would become much more satisfying. (IT Specialist 18)

Respondents in our study, while acknowledging AI’s usefulness in facilitating auditor judge-
ment to foster audit quality, are reluctant to trust. Our study found that auditors’ involvement in
their design and an explainable and transparent process in the AI ‘black-box’ are necessary for
effective deployment of AI to complement human judgement and inform decision making. Big4
audit partner observes that the new AI tools now ‘are not focusing on productivity, and they are
focusing on modern-day issues of data . . . to inform future decision making’ (Auditor 10).

4.3. Environmental context

Environmental factors drive organisations to adopt technologies and vary for different contexts.
For example, as shown in Table 2, competitive pressure influenced the adoption of CRM (customer
relationship management) systems (Cruz-Jesus et al., 2019) while accounting standards and part-
ner support strongly influenced the adoption of SaaS ERP systems (Seethamraju, 2015). In our
study, we found standards (plus regulation readiness), and clients’ readiness to be significant in
influencing adoption, extending previous research (Bakarich and O’Brien, 2021; Kend and
Nguyen, 2020).

4.3.1. Auditing standards and regulation readiness. Current standards were developed when client
database sizes were small, processing power was low and compliance demands were fewer (Issa
et al., 2016). In our study, while audit firms covet the audit standards to drive AI adoption, regula-
tors consider standards and guidelines for emerging technologies, such as AI, as useful but not
essential. The following observation illustrates this:

the working party on technology has stalled . . . they don’t know how to address problems relating to AI
tools and their use . . . more of a nice to have rather than a must-have. (Standard Setter 24)

Standard setters are waiting for audit firms to adopt AI tools so their experience can inform the
standard-setting process. Audit firms, on the other hand, believe that, without relevant audit stand-
ards to guide the use of these technologies, they will not be able to demonstrate the process to regu-
lators, particularly to the Australian Securities and Investments Commission (ASIC) who undertake
inspections of audit quality. As noted by a senior manager in a mid-tier firm, ‘you could do the AI
794 Australian Journal of Management 48(4)

stuff, but how do you then talk to a regulator and say you’ve done that testing, we don’t have an
auditing standard on that’ (Auditor 3). An IT specialist also concurs that ‘it is a serious misconcep-
tion to think the standards don’t inhibit the progress of auditing’ (IT Specialist 18).
Developing and implementing standards is a complex process carried out in collaboration with
practitioners and regulatory bodies such as ASIC. It requires a full understanding of the inputs,
processes and analysis related to AI tools as a Big4 partner explains,

auditing standards board need to understand what AI is . . . and how do they write standards around that?
How then does regulator ASIC understand what the auditing standard is trying to do? . . . You then need
to train auditors . . . how does an audit done without AI differ from an audit done with AI? (Auditor 11)

A manager in a professional body contends that it is a costly, complex process requiring consid-
erable expertise and time with ‘a lot of practitioners blaming standards and the regulators . . .
because the standards are not clear . . . development of audit standards is quite costly, requires a lot
of time, a lot of expertise’ (Professional Body 21). Standards, a Big4 auditor suggests, should
impose restrictions on AI use:

you need to assign appropriate senior members to particular tasks because they have the knowledge and
experience . . . they might even put in a standard saying where AI should not work on these high-risk
areas. (Auditor 12)

Implementation of the relevant standards may also necessitate changes to audit methodology
including auditing of the AI tools as illustrated below:

. . . AI tools should also go through a rigorous process because if we don’t audit these tools ourselves, can
[we] rely on these? (Auditor 5) and,

we need to sign off accounts that meet auditing standards [which] have the force of law. So, does AI meet
the requirements of the auditing standards? (Auditor 3)

Respondents also raised a general lack of regulation readiness and expressed doubts about
ASIC’s capacity to review audits (and standards compliance) undertaken with AI technology: ‘not
sure there are people from ASIC who would be a specialist in using AI in audits and in inspecting
Big4 firms using AI tools . . . only 2 people in Australia’ (Standard Setter 23). This view is also
echoed by senior audit partner (Auditor 11): ‘ASIC who are looking at your files may not neces-
sarily agree with the tools providing the evidence . . . and may question did the computer even do
what you [auditor] said it would do?’.
Our study observed that although it is a costly and complex process, the development of rele-
vant auditing standards in consultation with practitioners and regulatory bodies is necessary for AI
adoption. It includes considerations of legal requirements, restrictions on the use of AI tools in
high-risk areas and the audit of AI audit technology itself. Our study supports Issa et al.’s (2016)
and Manita et al.’s (2020) findings that it is challenging for the profession to adopt any technology
if not required by standard setters and the absence of relevant standards, their slow development
and lack of regulation readiness are constraining the adoption.

4.3.2. Clients’ readiness. Auditing is a unique specialist profession. To provide independent assur-
ance, auditors must understand their client’s IT systems and processes (Salijeni et al., 2019). Cli-
ent’s readiness, including greater IT connectivity and accessibility of data, allows auditors not only
Seethamraju and Hecimovic 795

to enhance the efficiency and rigour of the audit process, but also forces them to offer a new suite
of assurance services to address risks emanating from the adoption of new emerging technologies
(Pedrosa et al., 2020).
The adoption of AI by audit firms tends to be dependent on the clients’ needs and the value they
attribute to AI. As clients prefer to interact with auditors rather than with the tools (Chan et al.,
2012), concerns clients may not see value in modern AI tools are observed in our study. A senior
partner in Big4 observed, ‘we’ll put it in an audit proposal saying we’ve got this analytic AI soft-
ware, but they’re not screaming out for it . . . what’s the value?’ (Auditor 11).
Although some clients want the auditor to use AI, they must invest in their own IT and human
resources to derive the benefits. A manager suggests, ‘you need to have infrastructure in place in
clients, where you can get access to data that is useful’ (Standard Setter 22). Similarly, a manager
at a mid-tier firm asserts, ‘it comes down to how sophisticated are the clients and would clients pay
for it?’ (Auditor 3).
Despite clients’ perceived lack of understanding of AI audit processes and outputs, they may,
however, still engage the audit firm (with AI tools) as the risk remains with the auditors. As a man-
ager points out, ‘though clients are not ready in terms of their own data, IT infrastructure . . .
they’ll push ahead anyway as it’s all about being ahead of the game, and risk resides with the audi-
tors because they sign-off’ (Standard Setter 24). Both auditors and their clients should ‘not see AI
technology as a “silver bullet” and must understand the capabilities of AI’ (Standard Setter 22)
before the standard setter considers developing relevant standards.

5. Conclusion and discussion

AI, if deployed smartly, will transform the audit by creating opportunities and challenges. This
study using the TOE framework identified several technological, organisational and environmental
factors affecting the adoption of AI in auditing.
From the technological perspective – the perceived benefits, compatibility of AI tools with
existing audit tools and client accounting systems, and maturity of AI available for use in auditing
were key factors influencing the adoption of AI. Our study observed that by enabling and accelerat-
ing the automation of audit tasks, AI would help optimise auditor’s time and allow deeper analysis
of client data. AI adoption may also raise client expectations on audit with the pressure to extend
the traditional financial audit to the provision of more forward-looking real-time services such as
continuous auditing and value-adding consulting services.
Within the organisational context, organisational readiness, data quality and trust in AI tools are
significant factors affecting the adoption. Human bias that could be built into AI solutions, poor
data quality and lack of trust in AI tools due to its ‘black-box’ nature, legal implications of financial
auditing and belief that such tools cannot support tasks that require auditor judgement and scepti-
cism are barriers to adoption, according to our study. Respondents while recognising AI’s value in
enhancing audit quality are concerned about technology dominance. Auditors, with an AI tool
deployed in the audit process, tend to surrender their decision-making by relaxing their own judge-
ment. Given that audit quality and practice are under constant scrutiny, this has implications not
only for adoption but in possibly making the audit process even more hidden from view. While our
study found the Big4 are investing in training, skills development and development of AI tools and
getting ready to apply more AI in their audits, mid-tier firms are resource-constrained and do not
see it as imperative to adopt unless clients or regulation demand it. Our study cautions that while
AI may automate many low-level tasks previously undertaken by junior auditors, consideration
needs to be given to the potential deskilling effect. This has implications for higher education
796 Australian Journal of Management 48(4)

institutions and professional bodies in terms of developing strategies and opportunities for upskill-
ing graduates seeking to enter audit practice.
From the environmental perspective, auditing standards and clients’ readiness, plus regulation
readiness were key factors impacting AI adoption. Our study provides evidence that practitioners
are hesitant to adopt and are apportioning some blame to regulators and standard setters, while they
in turn are wanting to learn from practice to inform their standard setting process. Audit clients’
inadequate technology readiness towards AI tools, often blamed for the low level of adoption, and
incompatibility of AI tools with existing accounting systems are supported by our study.
Our study makes two important contributions. First, it is the first comprehensive study that
uncovers technological, organisational and environmental factors that affect AI adoption in audit-
ing, providing evidence to support the suppositions of prior studies. Second, our study extends the
scope of the TOE framework to an emerging technology, AI and to the auditing context, contribut-
ing empirical evidence to the technology adoption literature. Given the significant investment
required to acquire AI tools, build skills and capabilities, understanding the identified TOE factors
influencing successful adoption can help firms/practitioners prioritise resources based on their
individual strengths and weaknesses to improve solution development and implementation. Our
findings also have implications for regulatory bodies and standard setters, as they develop stand-
ards and guidance for the use of AI technologies on audit engagements. Adoption of AI technolo-
gies may lead to significant restructure of audit process with change in focus from the historical
nature of financial statement audits to one of assurance of information in real time.
Our study has some limitations opening up opportunities for future research. The use of a cross-
sectional qualitative exploratory study on AI adoption in audit, still in a nascent stage, suffers from
the limitation of drawing definitive conclusions that are embedded in an Australian context.
Because our exploratory study used the TOE framework to identify some of the salient factors
influencing adoption, it may have excluded other important factors. This risk would, however,
exist irrespective of the specific research approach and theoretical framework, considering the
under-explored nature of AI in the audit domain. It will be beneficial to investigate AI’s impact on
deskilling, auditor scepticism and organisation size – concerns identified as problematic in our
study. Further in-depth case studies and empirical survey studies to understand the extent of AI
adoption and its impact on the audit profession are necessary.
Auditing is entering a brave new world in which technologies such as AI have the transforma-
tive potential to improve efficiency and audit quality, create value-adding opportunities and erase
the boundaries between audit phases. Consequent to the new opportunities for continued innova-
tion, it is important for audit firms to apply professional scepticism, their professional trait, to
understand the impact of AI adoption. Only then will they be better positioned to understand how
‘value’ can be generated from such adoption and be in a position of control.

Acknowledgements
The authors would like to thank the research participants who generously shared their perceptions, views and
knowledge and the reviewers for their insightful and constructive comments.

Funding
The author(s) received no financial support for the research, authorship and/or publication of this article.

ORCID iD
Ravi Seethamraju https://orcid.org/0000-0002-1419-3806
Seethamraju and Hecimovic 797

References
Abed SS (2020) Social commerce adoption using TOE framework: An empirical investigation of Saudi
Arabian SMEs. International Journal of Information Management 53: 1–11.
Abou-El-Sood H, Kotb A and Allam A (2019) Exploring auditors’ perceptions of the usage and importance
of audit information technology. International Journal of Auditing 19: 252–266.
Ahmi A and Kent S (2013) The utilisation of generalized audit software (GAS) by external auditors.
Managerial Auditing Journal 28: 88–113.
Association of Chartered Certified Accountants (ACCA) (2019) Audit and technology, ACCA and CAANZ
(Chartered Accountants Australia and New Zealand). Available at: https://www.accaglobal.com/us/en/
professional-insights/technology/audit-and-tech.html
Australian Financial Review (2022) Big four consulting firm data tracker. Australian Financial Review,
16 March Available at: https://www.afr.com/companies/professional-services/big-four-consulting-
firm-data-tracker-20200207-p53yp4
Bakarich KM and O’Brien PE (2021) The robots are coming . . . but aren’t here yet: The use of artificial
intelligence technologies in the public accounting profession. Journal of Emerging Technologies in
Accounting 18: 27–43.
Bierstaker J, Janvrin D and Lowe D (2014) What factors influence auditors’ use of computer-assisted audit
techniques? Advances in Accounting 30: 67–74.
Bizarro PA and Dorian M (2017) Artificial intelligence: The future of auditing. Internal Auditing 32: 21–26.
Bose R and Luo X (2011) Integrative framework for assessing firms’ potential to undertake Green IT initia-
tives via virtualization – A theoretical perspective. The Journal of Strategic Information Systems 20:
38–54.
Braun RL and Davis HE (2003) Computer-assisted audit tools and techniques: Analysis and perspectives.
Managerial Auditing Journal 18: 725–731.
Brennan B, Baccala M and Flynn M (2017) How auditing will incorporate AI. CFO Auditing, 2 February.
Available at: https://www.cfo.com/auditing/2017/02/artificial-intelligence-audits
Byrnes PE, Al-Awadhi A, Gullvist B, et al. (2018) Evolution of auditing: From the traditional approach to the
future audit. Continuous Auditing: Theory and Application 21: 285–297.
Chan HL, DeBoskey DG and Hee K (2012) Audit fee patterns of Big4 and non-Big4 firms. The CPA Journal
82: 32–36.
Clohessy T and Acton T (2019) Investigating the influence of organizational factors on blockchain adoption:
An innovation theory perspective. Industrial Management & Data System 119: 1457–1491.
Cohen M and Rozario A (2019) Exploring the use of robotic process automation (RPA) in substantive audit
procedures. The CPA Journal 89: 49–53.
Cooper L, Holderness DK, Sorensen T, et al. (2019) Robotic process automation in public accounting.
Accounting Horizons 33: 15–35.
Cruz-Jesus F, Pinheiro A and Oliveira T (2019) Understanding CRM adoption stages: Empirical analysis
building on the TOE framework. Computers in Industry 109: 1–13.
Deloitte Insights (2019) Using AI to unleash the power of unstructured government data. Deloitte Insights, 16
January. Available at: https://www2.deloitte.com/insights/us/en/focus/cognitive-technologies/natural-
language-processing-examples-in-government-data.html
Dowling C and Leech SA (2014) A Big 4 firm’s use of information technology to control the audit process:
How an audit support system is changing auditor behavior. Contemporary Accounting Research 31:
230–252.
Duan Y, Edwards JS and Dwivedi YK (2019) Artificial intelligence for decision making in the era of Big
Data – Evolution, challenges and research agenda. International Journal of Information Management
48: 63–71.
Ernst H (1997) The use of patent data for technological forecasting: The diffusion of CNC-technology in the
machine tool industry. Small Business Economics 9: 361–381.
EY (2018) How artificial intelligence will transform the audit. Available at: https://www.ey.com/en_gl/assur-
ance/how-artificial-intelligence-will-transform-the-audit
798 Australian Journal of Management 48(4)

Fan W, Liu J, Zhu S, et al. (2020) Investigating the impacting factors for the healthcare professionals to
adopt artificial intelligence-based medical diagnosis support system (AIMDSS). Annals of Operations
Research 294: 567–592.
Furneaux B and Wade M (2011) An exploration of organizational level information systems discontinuance
intentions. MIS Quarterly 35: 573–598.
Goh C, Pan G, Seow PS, et al. (2019) Charting the Future of Accountancy with AI. Research Collection
School of Accountancy. Available at: https://ink.library.smu.edu.sg/soa_research/1806
Gray G, Chiu V, Liu Q, et al. (2014) The expert systems life cycle in AIS research: What does it mean for
future AIS research? International Journal of Accounting Information Systems 15: 423–451.
Griffith EE, Hammersley JS and Kadous K (2015) Audits of complex estimates as verification of management
numbers: How institutional pressures shape practice. Contemporary Accounting Research 32: 833–863.
Harding N, Azim MI, Jidin R, et al. (2016) A consideration of literature on trust and distrust as they relate to
auditor professional scepticism. Australian Accounting Review 26: 243–254.
Hossain MA and Quaddus M (2011) The adoption and continued usage intention of RFID: An integrated
framework. Information Technology & People 24: 236–256.
Issa H, Sun T and Vasarhelyi MA (2016) Research ideas for artificial intelligence in auditing: The formaliza-
tion of audit and workforce supplementation. Journal of Emerging Technologies in Accounting 13: 1–20.
Janssen M, Weerakkody V, Ismagilova E, et al. (2020) A framework for analysing blockchain technology
adoption: Integrating institutional, market and technical factors. International Journal of Information
Management 50: 302–309.
Kachroo P, Saiewitz A, Raschke R, et al. (2020) A new language and input-output hidden Markov model for
automated audit inquiry. IEEE Intelligent Systems 35: 39–49.
Kapoor M (2020) Big4 invest in billions in tech, reshaping their identities. Available at: https://news.bloomb-
ergtax.com/daily-tax-report/
Kend M and Nguyen LA (2020) Big Data analytics and other emerging technologies: The impact on the
Australian audit and assurance profession. Australian Accounting Review 30: 269–282.
Khan R, Adi E and Hussain O (2021) AI-based audit of fuzzy front end innovation using ISO56002.
Managerial Auditing Journal 36: 564–590.
Kokina J and Davenport TH (2017) The emergence of artificial intelligence: How automation is changing
auditing. Journal of Emerging Technologies in Accounting 14: 115–122.
Kordzadeh N and Ghasemaghaei M (2021) Algorithmic bias: Review, synthesis, and future research direc-
tions. European Journal of Information Systems 1: 1–22.
KPMG (2020) AI: The new power generation. Available at: https://home.kpmg/xx/en/home/insights/2020/02/
ai-the-new-power-generation
Li H, Dai J, Gershberg T, et al. (2018) Understanding usage and value of audit analytics for internal auditors:
An organizational approach. International Journal of Accounting Information Systems 28: 59–76.
Lian JW, Yen DC and Wang YT (2014) An exploratory study to understand the critical factors affecting the
decision to adopt cloud computing in Taiwan hospital. International Journal of Information Management
34: 28–36.
Lillis AM and Mundy J (2005) Cross-sectional field studies in management accounting research – Closing
the gaps between surveys and case studies. Journal of Management Accounting Research 17: 119–141.
Lin, H-F (2014) Understanding the determinants of electronic supply chain management systems adoption:
Using technology-organization-environment framework. Technological Forecasting & Social Change
86: 80–92.
McKnight DH, Carter M, Thatcher JB, et al. (2011) Trust in a specific technology: An investigation of its
components and measures. ACM Transactions of Management Information Systems 2: 1–25.
Mahzan N and Lymer A (2009) Examining the adoption of computer-assisted audit tools and techniques.
Managerial Auditing Journal 29: 327–349.
Manita R, Baudier P and Hikkerova L (2020) The digital transformation of external audit and its impact on
corporate governance. Technological Forecasting & Social Change 150: 119751.
Martinov-Bennie N, Roebuck P and Soh D (2016) Auditing and Assurance: A Case Studies Approach. New
York: LexisNexis.
Seethamraju and Hecimovic 799

Martins R, Oliveira T and Thomas MA (2016) An empirical analysis to assess the determinants of SaaS dif-
fusion in firms. Computers in Human Behavior 62: 19–33.
Miller T (2019) Artificial intelligence: Insights from the social sciences. Artificial Intelligence 267: 1–38.
Mishra AN, Konana P and Barua A (2007) Antecedents and consequences of internet use in procurement: An
empirical investigation of US manufacturing firms. Information Systems Research 18: 103–120.
Munoko I, Brown- Liburd HL and Vasarhelyi M (2020) The ethical implications of using artificial intelli-
gence in auditing. Journal of Business Ethics 167: 209–234.
Oliveira T and Martins MF (2011) Literature review of information technology adoption models at firm level.
The Electronic Journal of Information Systems Evaluation 14: 110–121.
Oliveira T, Martins R, Sarker S, et al. (2019) Understanding SaaS adoption: The moderating impact of the
environment context. International Journal of Information Management 49: 1–12.
Panicker RO, Soman B, Gangadharan KV, et al. (2016) An adoption model describing clinician’s acceptance
of automated diagnostic system for tuberculosis. Health Technology 6: 247–257.
Pedrosa I, Costa CJ and Aparicio M (2020) Determinants of adoption of computer-assisted auditing tools
(CAATs). Cognition, Technology & Work 22: 565–583.
Prakash AV and Das S (2021) Medical practitioner’s adoption of intelligent clinical diagnostic decision sup-
port systems: A mixed-methods study. Information & Management 58: 103524.
Puklavec B, Oliveira T and Popovič A (2018) Understanding the determinants of business intelligence system
adoption stages. Industrial Management & Data Systems 118: 236–261.
PwC (2020) What is an audit? Available at: https://www.pwc.com.au/assurance/audit.html
Rozario AM and Thomas C (2019) Reengineering the audit with blockchain and smart contracts. Journal of
Emerging Technologies in Accounting 16: 21–35.
Rozario AM and Vasarhelyi MA (2018) How robotic process automation is transforming accounting and
auditing. The CPA Journal 88: 46–49.
Salijeni G, Samsonova-Taddei A and Turley S (2019) Big Data and changes in audit technology: Contemplating
a research agenda. Accounting and Business Research 49: 95–119.
Schmidt P, Biessmann F and Teubner T (2020) Transparency and trust in artificial intelligence systems.
Journal of Decision Systems 29: 260–278.
Seethamraju R (2015) Adoption of software as a service (SaaS) enterprise resource planning (ERP) system in
small and medium sized enterprises (SMEs). Information Systems Frontiers 17: 475–492.
Sidhu H (2019) Trust in technology: Raising the bar in the audit industry. Available at: https://www.accoun-
tingtoday.com/opinion/trust-in-technology-raising-the-bar-in-the-audit-industry
Sutton SG, Holt M and Arnold V (2016) The reports of my death are greatly exaggerated – Artificial intel-
ligence research in accounting. International Journal of Accounting Information Systems 22: 60–73.
Tapscott D and Tapscott A (2016) Blockchain Revolution: How the Technology Behind Bitcoin Is Changing
Money, Business, and the World. New York: Penguin.
Ting S and Vasarhelyi MA (2018) Embracing textual data analytics in auditing with deep learning.
International Journal of Digital Accounting Research 18: 49–67.
Tornatzky L and Fleischer M (1990) The Process of Technology Innovation. Lexington, KY: Lexington
Books.
Trotman KT, Tan HC and Ang N (2011) Fifty-year overview of judgment and decision-making research in
accounting. Accounting & Finance 51: 278–360.
Vasarhelyi MA, Bonson E and Hooitash R (2005) Artificial Intelligence in Accounting and Auditing:
International Perspectives, vol. 6. Princeton, NJ: Markus Wiener Publishers.
Vasarhelyi MA, Kogan A and Tuttle BM (2015) Big Data in accounting: An overview. Accounting Horizons
29: 381–396.
Venkatesh V and Bala H (2012) Adoption and impacts of interorganizational business process standards: Role
of partnering synergy. Information Systems Research 23: 1131–1157.
Weick KE and Quinn RE (1999) Organizational change and development. Annual Review of Psychology 50:
361–386.
Wiggins MW, Auton J, Bayl-Smith P, et al. (2020) Optimising the future of technology in organisations: A
human factors perspective. Australian Journal of Management 45: 449–467.
800 Australian Journal of Management 48(4)

World Economic Forum (WEF) (2015) Deep shift: Technology tipping points and social impact. Available at:
https://www3.weforum.org/docs/WEF_GAC15_Technological_Tipping_Points_report_2015.pdf
Xu W, Ou P and Fan W (2017) Antecedents of ERP assimilation and its impact on ERP value: A TOE-based
model and empirical test. Information Systems Frontiers 19: 13–30.
Yin RK (2009) How to do better case studies. In: Bickman L and Rog DJ (eds) The SAGE Handbook of
Applied Social Research Methods. 2nd ed. Thousand Oaks, CA: SAGE, pp. 254–282.
Zhu K and Kraemer KL (2005) Post-adoption variations in usage and value of e-business by organizations:
Cross-country evidence from the retail industry. Information Systems Research 16: 61–84.