Page 23 - CMA Journal (May-Jun 2026)

F oc us S ec t ion
Transforming Fiscal Oversight: Digital
and AI Solutions for FBR & AGP
Introduction with monitoring rather
than veri cation being
Digital technologies are now becoming part of the the priority. The current
emerging economies' scal architecture in public operational framework
nancial management. The high level of tax evasion has been characterized
and low tax to GDP ratio in Pakistan has made it all the by a lack of integrated
more important to have e ective tax administration data governance, with
and audit mechanism (Rasheed et al., 2024). The data from customs,
Federal Board of Revenue (FBR), and the Auditor inland revenue and
General of Pakistan (AGP) are two cornerstones of scal provincial tax authorities
governance, but their isolated institutional not being easily linked
frameworks, manual approaches, and absence of a together and missing
common data environment puts the traditional way of elements in Abdul Qayyum, FCMA
work at risk (Ayyadurai et al., 2026). Many businesses comprehensive taxpayer Group Finance Manager,
are now transitioning to a digital- rst strategy, and one pro ling (Rasheed et al., Lootah Group, UAE
way to achieve more transparency, increase 2024).
compliance and improve revenue collection is to use
Arti cial Intelligence (AI). The academic literature in Challenges in Traditional Tax and Audit
recent years has emphasized the potential of Mechanisms
implementing AI-powered systems for better
detection, to reduce the risk of mistakes by humans in The existing tax and audit framework in Pakistan is
the process, and to make it conducive to algorithmic plagued with numerous challenges, jeopardizing scal
responsibility, following the OECD Tax Administration integrity. One of the main challenges is that tax evasion
3.0 blueprint (Belahouaoui & Alm, 2025). The article is common and the informal economy is signi cant,
analyzes both tax and audit systems in Pakistan and outside the formal regulatory framework to a large
their current state, as well as the transformative power extent. Manual auditing processes are ine cient,
of AI and digitalization, and suggests policy prone to human error and subject to discretion
recommendations for incorporating these exercised by tax o cials (unde ned & Shepel, 2024),
technologies to boost and enhance public nance and empirical evidence in Pakistan shows that
governance in Pakistan. traditional tax audits have contributed little to direct
tax collection and have not been successful in having a
Overview of Existing Tax and Audit signi cant impact on the deterrence of tax evasion
Systems in Pakistan because of their limited scope and reactive nature. In
addition, data is fragmented across di erent
The Federal Board of Revenue (FBR) administers the departments of the government, which makes it hard
federal taxes such as income tax, sales tax and customs for the FBR to have a single and complete picture of
duties. The AGP, on its part, is responsible for the taxpayer activities, and make it di cult to compare
auditing of public sector entities for scal income sources and nd any discrepancies. Financial
accountability and legislative oversight (Usman, 2023). data is not monitored in real-time, which makes it
Traditionally, these institutions have not been di cult to detect anomalies in nancial transactions
technologically interoperable. The FBR has adopted years after they happen, making them less e ective at
digital measures like introducing an online portal for correcting issues (Ahmed, 2024).
taxpayer registration, ling of returns, and processing
of payments called IRIS (Mohammad A an Habib et al., Role of Digitalization in Tax
2024). IRIS has also enabled the basic e- ling functions, Administration
and the system is primarily used as a repository of les Digitalization is the building block for modernizing the
rather than a smart, predictive analytics tool (Kinanti et tax administration. Reduction of manual work through
al., 2025). Likewise, the AGP audit processes are largely the use of digital tools such as e- ling, e-payment and
manual and are mostly conducted after the automated data processing, cuts down on
expenditure has been made, rather than in real-time, administrative work.
ICMA’s Chartered Management Accountant, May-Jun 2026 21
Examples of digital reforms such as Faceless Customs any unusual patterns that might constitute a tax Barriers to Implementation in Pakistan must be addressed to achieve a successful
Assessment FCA demonstrate the potential impact of evasion/money laundering event, something that a However, there are several challenges that hinder the implementation, including the development of policy
digital reforms in Pakistan, where reducing physical
and institutional capacity. In this regard AI must be
conventional tax detection system couldn't detect. In
implementation of AI-driven tax and audit systems in
interactions and paperwork could enhance trade and addition, Large Language Models LLM like TaxTajweez Pakistan. Lack of good digital infrastructure, such as seen as a strategic reform that aims to establish a more
increase revenue e ciency (Nasir, 2026). Digital record
show the potential to increase the e ectiveness of tax
accountable, responsive, and e cient system of public
the poor quality of broadband internet access, and
nancial governance, based on international best
keeping systems are more accurate and immutable for
compliance by o ering precise and relevant guidance,
tax data collection, increasing transparency and which in turn would decrease compliance pressures outdated hardware, hinders the scalability of practices. Further studies are needed on the ongoing
minimizing the risk of tampering with tax data to and boost voluntary compliance (Mohammad A an cutting-edge technological solutions (unde ned & AI initiatives in Pakistan to assess their long-term
defraud the tax system. Digital systems can be used to Habib et al., 2024). Shepel, 2024). One of the critical issues is cybersecurity impacts on tax compliance, audit e ciency, and scal
optimize and digitalize routine compliance monitoring and data privacy, as the inclusion of nancial results.
tasks, and help tax authorities focus their resources on Comparative Analysis: Traditional vs information means it needs to be secured to ensure it is
more high-risk cases (Ahmed, 2024). Also, the AI-Based Tax and Audit Systems not compromised and maintains public trust (Umbet References
reduction of human touch in the normal processes will et al., 2025). Entangled in a web of institutional • Ahmed, E. R. (2024). Impact of Digital Procedures in Accounting and Auditing on Tax
contribute to minimize the risk of corruption and The table below enumerates the major di erences interests and without strategic vision, institutional Collection E ciency. Acta Academiae Beregsasiensis. Economics(6), 300-306.
https://doi.org/10.58423/2786-6742/2024-6-300-306
enhance the level of public trust in tax system fairness between the conventional and AI tax and audit resistance to change is another obstacle to progress. • Ayyadurai, R., Maamra, O. E., Praseptiawan, M., Kanimozhi, R., & Soundarraj, K. (2026).
(Pali & Rama, 2025). systems and the advantages of digital tax and audit Moreover, there is a shortage of data science and Developing a technology-driven simulation model for integrating nance, auditing,
accounting, and taxation in emerging economies. African Journal of Science,
systems in improving transparency and e ciency. AI-related skilled human resources in the public sector. Technology, Innovation and Development, 1-17. https://doi.org/10.1080/
Application of Arti cial Intelligence in International Best Practices Another signi cant concern is data incompleteness, • 20421338.2026.2659633
Azenzoul, A., Mahouat, N., Vandapuye, S., Slimane, S. N., Jbene, M., & Mokhlis, K. (2026).
Tax and Audit Systems where disparate data silos make it di cult to perform a From Predictive Accuracy to Algorithmic Justice: Mapping the Multidimensional Impact
of AI in Tax Auditing. Journal of Risk and Financial Management, 19(5), 354.
AI has been widely adopted by the tax administration thorough analysis needed for e ective AI application. https://doi.org/10.3390/jrfm19050354
The digitization revolution is not the only revolution industry, with the OECD countries and emerging Policy Recommendations • Belahouaoui, R., & Alm, J. (2025). Tax Fraud Detection Using Arti cial Intelligence-Based
Technologies: Trends and Implications. Journal of Risk and Financial Management,
that's taking place with Arti cial Intelligence (AI) in tax economies having a high level of AI adoption. The 18(9), 502. https://doi.org/10.3390/jrfm18090502
and audit. Machine learning algorithms, deep learning OECD Tax Administration 3.0 framework is focused on The following policy recommendations will address • Fajardo, B., & Neiva, S. (2025). ‘Malha Fina de Convênios’: an AI application for auditing
in the Brazilian federal government. Public Money & Management, 1-12.
networks and natural language processing (NLP) are adaptable AI monitoring to improve compliance and these issues, and will help put the advantages of AI and https://doi.org/10.1080/09540962.2025.2499704
detection processes (Belahouaoui & Alm, 2025). The
examples of AI technologies that assist with proactive adoption of AI-powered analytics in India has not only digital systems into practice. First, integrating the FBR • Hassan, B. (2018). Do Tax Audits Contribute to Tax Revenue and Deter Tax Evasion? – An
Exploratory Study of Tax Audits in Pakistan. Asia-Paci c Tax Bulletin, 24(6).
fraud detection and determining which risks to audit. enhanced the e ciency of GST compliance but also and AGP databases is essential to ensure that there is https://doi.org/10.59403/nnpxkz
Supervised machine learning models such as random helped mitigate fraud, providing Pakistan with lessons one scal database which enables the tax-payers to be • Kinanti, K. A., Chandra, Y. A., & Suyono, A. P. (2025). Digital Integration In Tax Systems:
forest models can help improve e ciency and reduce
Literature Review On Enhancing E ciency And Ensuring Compliance. Assets : Jurnal
processing expenses in the audit while irregularity in on how to seamlessly connect various tax systems pro led as a whole and for audit purposes (Ayyadurai Ilmiah Ilmu Akuntansi, Keuangan dan Pajak, 9(1), 30-54. https://doi.org/10.30741/
assets.v9i1.1508
nancial transactions can be detected. The more the (Rathi & et al., 2021). Likewise, Brazil's ‘Malha Fina de et al., 2026). Second, there is a need to invest in the • Mohammad A an Habib, u., Shehryar Amin, u., Muhammad Oqba, u., Sameer Jaipal, u.,
tax authorities are able to predict tax compliance Convênios’ algorithm uses supervised machine adoption of AI and big data analytics, including Muhammad Junaid Khan, u., & Abdul Samad, u. (2024). TaxTajweez: A Large Language
Model-based Chatbot for Income Tax Information In Pakistan Using Retrieval
patterns and the revenue streams the more quickly learning to audit federal transfers, resulting in marked predictive modelling, machine learning tools to detect Augmented Generation (RAG). The International FLAIRS Conference Proceedings,
and e ectively they can take action to identify e ciencies and cost savings. The highlights from these fraud and assess risk. Third, there is a need for extensive • 37.https://doi.org/10.32473/ airs.37.1.135648
Nasir, J. (2026). Impact of Digitalization on Trade Facilitation and Revenue: Evidence
potential tax compliance concerns before they even foreign cases demonstrate the need for strong digital data science and cybersecurity capacity building and from Pakistan. Global Trade and Customs Journal, 21, 71-82. https://doi.org/10.54648/
become issues, and take appropriate action. infrastructure, inter-agency data exchange, and training to bolster existing sta as well as hire of data • gtcj2026008
Nasira Perveen, u., & Ashfaq Ahmad, u. (2022). Tax Technology, Fairness Perception and
Furthermore, the anomaly detection systems can investment in AI technologies for impactful scal science and cybersecurity specialists. To safeguard Tax Compliance among Individual Taxpayers. Audit and Accounting Review, 2(2).
process a vast amount of data in real-time, and warn of reforms. sensitive information and promote responsible use of • https://doi.org/10.32350/aar.22.05
Pali, A., & Rama, S. (2025). Digital Transformation and Its Impact on Public Services: The
AI, it is essential to adopt more stringent cybersecurity Case of the Tax Administration Services. Interdisciplinary Journal of Research and
Development, 12, 40. https://doi.org/10.56345/ijrdv12n1s106
measures and set up data governance protocols. Last • Rasheed, S., Dr. Milhan Afzal Khan, u., Ijaz, I., Musa Khan, M., & Arshad, M. N. (2024).
but not least, the public-private partnership (P3) ANALYSIS OF CORDA IMPLEMENTATION IN TAXATION SYSTEM OF PAKISTAN: A
approach can be used to speed up the digital PERMISSIONED BLOCKCHAIN FRAMEWORK. Pakistan Journal of Scienti c Research,
4(1), 61-67. https://doi.org/10.57041/pjosr.v4i1.1019
transformation and to tap the private sector's expertise • Rathi, A., & et al. (2021). A Study on Application of Arti cial Intelligence and Machine
Learning in Indian Taxation System. Psychology and Education Journal, 58(2),
and innovations to provide suitable solutions in • 1226-1233. https://doi.org/10.17762/pae.v58i2.2265
Pakistan's scal environment.
Umbet, M., Askarov, D., Rudžionienė, K., Christauskas, Č., & Alikulova, L. (2025).
Evaluating the Implementation of Information Technology Audit Systems Within Tax
Conclusion Administration: A Risk Governance Perspective for Enhancing Digital Fiscal Integrity.
Journal of Risk and Financial Management, 18(8), 422. https://doi.org/
The use of digital technologies and Arti cial 10.3390/jrfm18080422
• unde ned, u., & Shepel, I. (2024). DIGITALIZATION OF TAX AUDITS: NEW APPROACHES
AND TECHNOLOGIES FOR IMPROVING CONTROL. Black Sea Economic Studies(90).
Intelligence (AI) in tax and audit procedures o ers a https://doi.org/10.32782/bses.90-12
improve
transparency,
to
valuable
chance
e ectiveness, and scal integrity in the public nancial • Usman, M. (2023). Evaluation of the Citizen Participatory Audit by the Department of
the
Dergisi(129),
201-233.
Auditor
Pakistan.
Sayıştay
General
of
management system in Pakistan. Institutions like the https://doi.org/10.52836/sayistay.1319323
Federal Board of Revenue (FBR) and the Auditor • Vistro, D. M., Farooq, M. S., Rehman, A. U., & Khan, M. A. (2021). Fraud Prevention in
Taxation System of Pakistan Using Blockchain Technology. Atlantis Highlights in
General of Pakistan (AGP) can increase revenue Computer Sciences. https://doi.org/10.2991/ahis.k.210913.074
collection, minimize tax evasion losses and boost
public con dence in government institutions by
moving away from the traditional manual process and
introducing AI-based, risk-based methods. However,
enabling policy measures, such as Data integration,
infrastructural, technical and organizational challenges