Page 25 - GSTL_16 April 2020_Vol 35_Part 3

2020 ] FROM DATA MINING TO ARTIFICIAL INTELLIGENCE J45
is no secret that some of the mapping Apps used by him in his mobile like loca-
tion map is provided to the service provider who in turn uses that data for their
marketing. Now it is not strange that every mobile phone user provides or com-
pelled to provide certain personal information including his/her location to var-
ious service providers through their mobile Applications.
Use of data analytics and AI in GST risk analysis
The real time data of people visiting the premises of say ‘X’ a popular
food chain or ‘Y’ a high-end night club may be collected through these Apps.
Total footfall during a month or within a particular period vis-a-vis retail bills
generated and the revenue paid in that period could be compared. The same
might not directly lead to the revenue realisation but may play an important role
in gauging the quantum of business and tax paid. The modalities to tap the in-
formation like how many messages or notifications were generated on the basis
of these Apps may be ascertained by asking the information under Section 151 of
the CGST Act, 2017. In fact, the Section supra empowers the Commissioner of
CGST to ask for a specific information. Thus, the companies issuing the push no-
tifications on behalf of their customers by whatever medium can be asked to di-
vulge the details like number of messages sent, discarded, failed, viewed and
clicked etc. of the targeted customers.
Whereas the AI can be applied in many ways depending on the need, in
revenue matters its scope may be divided in two broad categories :
(1) Directions/enforcement by the department; and
(2) The compliance/procedural adaptation by the assessee.
Further, department can use the AI in better compliance of returns, reports, re-
funds and in similar fields where nudging can improve the overall rate of approv-
als. At the same time AI based actions are an effective and probably the only viable
solution for enforcement. In the era of GST, where almost the entire gamut of
scheme is technology centric and all the communication whatsoever is strictly
driven through the electronic medium, it gains enhanced significance. So much so
that the failure of GSTN (Goods and Services Tax Network) is stated to be the pri-
mary cause of poor compliance and revenue leakage of GST.
It is the case that more than 12 million of assessees who not only file their
complete periodical returns and issue their invoices in electronic mode but the
entire transportation of goods with mandatory E-way Bill system has taken a
paradigm shift from the previous regime. It is but obvious that enforcement
ought to be technologically driven as the old method of information collection,
crystallization, converting it into intelligence and then actionable/credible input
is a long drawn one. Forensic signature or cyber world-oriented information is
now the most credible and authentic workable source as far as the enforcement
of law (revenue law) is concern.
Though both the categories i.e. compliance and enforcement of law are
equally important yet education to taxpayers to help them comply more easily
are equally significant for a successful GST. Exploring the use of advanced ana-
lytics and AI to meet these two main objectives in the new set up is of utmost
importance. Any non-technology driven approach not only be time taking and
unviable given the sheer number of registered taxpayers but it also involves the
GST LAW TIMES  16th April 2020  145