How LegalInk Auto-Extracts GST and Income Tax Notice Details

By LegalInk Editorial ·

Every tax litigation file begins with the same friction. A notice arrives — a scrutiny notice under Section 143(2) of the Income Tax Act, 1961, a show-cause notice under Section 74 of the CGST Act, 2017, a demand under Section 73 — and before any substantive drafting can begin, someone has to parse it manually for the PAN or GSTIN, the assessment year or tax period, the specific sub-sections invoked, the demand quantum, and the response deadline. For an advocate running a mixed dossier of GST and direct-tax matters, this parsing stage is invisible labour: low-skill, high-stakes, and entirely unscalable. LegalInk's notice extraction feature is designed to remove that stage. This post walks through what it extracts, how it fits into the drafting workflow that follows, which notice types it currently handles, and where it deliberately stops short.

The Problem With Manual Notice Reading at Scale

Every Notice Is Structured Differently

GST and income-tax notices follow broadly standardised departmental formats, but the variation in practice is considerable. Notices generated through the GSTN portal differ in layout from those issued manually by a Range Officer. Income Tax notices issued under the Faceless Assessment Scheme (Section 144B of the Income Tax Act, 1961) carry a National Faceless Assessment Centre header and a Document Identification Number (DIN), while older notices may still carry jurisdictional officer details. Layer on state GST authority notices, which often adapt the central template to local conventions, and you have a document type that resists clean template-matching.

The consequence for advocates is that each notice demands a fresh read — not because the legal issues are necessarily novel, but because the factual scaffolding (dates, demands, periods, provisions) has to be re-identified before any substantive analysis can begin. Multiply that across twenty open files and the cost compounds.

The Stakes of Missing a Detail

The pressure is not merely inconvenient. Under Section 75(3) of the CGST Act, an adjudicating authority must issue an order within three years of the due date for furnishing the annual return for the relevant financial year, or within five years where fraud or suppression is alleged under Section 74. Missing a response deadline, or misstating the demand period in your written submission, can prejudice your client's procedural position. Under the Income Tax Act, scrutiny timelines are equally strict: failure to respond to a notice under Section 142(1) in time can expose an assessee to best-judgment assessment under Section 144.

When these consequences are on the table, a parsing error at the intake stage is not a minor inefficiency. It is a professional risk that travels through every later stage of the file.

Structured Output From Unstructured Documents

LegalInk's extraction engine accepts a notice — uploaded as a PDF, an image scan, or pasted text — and returns a structured data output covering the fields a tax advocate actually uses. The system is built for the documents Indian practitioners receive in practice: GSTN portal PDFs, DIN-bearing income-tax notices, state commercial tax authority communications, and department-generated show-cause notices.

The extraction layer is calibrated to the actual field patterns that appear in CGST, IGST, and SGST notices and in Income Tax Act notices. It recognises, for instance, that what follows "Period of Dispute" in a GST show-cause notice and what follows "Assessment Year" in an Income Tax scrutiny notice are structurally analogous fields, even though they sit under different labels and in different positions on the page.

Fields Extracted Across Notice Types

For GST notices (Central and State), extraction covers:

  • GSTIN of the noticee
  • Legal name and trade name
  • Tax period (month, quarter, or financial year)
  • Notice reference number and date of issue
  • Provision(s) invoked — for example, Section 73 or Section 74 of the CGST Act, or the corresponding IGST Act or State GST Act provision
  • Nature of demand (tax, interest, penalty, or a combination)
  • Demand quantum broken down by head where the notice specifies it
  • Reply-by date and designated forum for the response
  • Officer designation and address (for non-faceless matters)

For Income Tax notices, extraction covers:

  • PAN and assessee name
  • Assessment year
  • Section under which the notice is issued — for example, Section 131, 142(1), 143(2), or 148 of the Income Tax Act
  • DIN (Document Identification Number) where present
  • Nature of query or addition proposed
  • Date of notice and compliance date
  • Faceless unit reference or jurisdictional officer details

The structured output is presented in a summary panel alongside the original document. You retain full access to the source text while working from the extracted fields.

Handling Scanned and Handwritten Notices

A real operational constraint in Indian tax practice is that not every notice arrives as a clean digital PDF. Older department notices, or notices issued by state-level authorities, may arrive as scanned images — sometimes photographed on a mobile phone by the client and forwarded over WhatsApp. LegalInk's extraction pipeline includes an optical recognition layer that pre-processes image-based inputs, recovering legible text from moderate-quality scans.

The system flags low-confidence extractions explicitly. You know which fields to verify against the original rather than discovering a parsing gap at the drafting stage, when the cost of correction is higher.

How the Extraction Fits Into Your Drafting Workflow

From Notice to Reply in a Single Session

Once extraction is complete, LegalInk maps the output directly into the drafting environment. If you are preparing a reply to a GST show-cause notice, the extracted fields — GSTIN, tax period, provision invoked, demand quantum — populate the reply template automatically. You are not re-typing the assessee's GSTIN into a document header or re-reading the notice to confirm which financial year is under dispute.

This matters because the cognitive overhead of context-switching between a notice PDF, a draft response, and your notes is where small errors typically enter a file. When the extraction output is already in the drafting pane, your attention can go directly to the substantive legal argument: whether the limitation period under Section 73(10) of the CGST Act has expired, whether the demand conflates two distinct taxable supplies, or whether the addition proposed under the Income Tax Act is supported by the material on record.

Take a representative example. A client forwards a Section 74 show-cause notice covering FY 2018-19, alleging suppression of outward supplies and proposing a demand of ₹42 lakh in tax plus equivalent penalty. Manually, the intake involves locating the GSTIN, confirming the period, isolating the alleged suppression amount, and checking whether the five-year limitation under Section 74(10) has been observed. With extraction, those four data points are surfaced in the summary panel within a single review pass, and the limitation calculation is computed against the relevant annual return due date.

For advocates handling a volume of notices across multiple clients, the LegalInk compliance framework provides a consolidated view of all extracted notices, their deadlines, and their current drafting status — so nothing slips through because it was sitting in an email thread.

Integration With the Brief Generator

Not every notice response is a standalone reply. In matters that have progressed to appeal, or where the underlying transaction is complex, you may need to produce a structured brief alongside the written response — covering the facts, the legal provisions in issue, the department's position, and your counter-arguments.

The brief generator pulls directly from the extraction output. The assessee's details, the notice particulars, and the provisions under challenge are already populated. You are building legal argument on top of a pre-populated factual foundation rather than assembling the foundation from scratch.

Version Control and Audit Readiness

Tax litigation files often span multiple rounds: a show-cause notice is followed by a personal hearing, a hearing by an order, an order by an appeal. LegalInk maintains an extraction record for each notice in a matter, so you can see at a glance what was extracted at intake, how the drafted response addressed each extracted field, and what the timeline of communications looks like. For advocates working alongside CAs or in-house finance teams, this audit trail is directly useful — it is the kind of documentation that supports a transparent, defensible client file.

Notice Types Currently Supported and Known Limitations

Supported Notice Categories

LegalInk currently supports extraction across the following categories, which together cover the large majority of notices Indian tax advocates encounter in practice:

  • CGST Act, 2017: Notices under Sections 46, 61, 63, 65, 73, 74, and demand-cum-show-cause notices issued by GST authorities
  • IGST Act, 2017 and State GST Acts: Notices substantially mirroring the CGST format, with state-specific header adaptations handled
  • Income Tax Act, 1961: Notices under Sections 131, 133(6), 142(1), 143(2), 148, and 156 (demand notice)
  • Faceless Assessment Scheme notices bearing DIN, issued under Section 144B
  • TDS-related communications issued under Sections 200A and 201

Limitations Worth Knowing

The extraction layer performs reliably on notices where the department has followed standard templates — which is true for the overwhelming majority of centrally issued GST and income-tax notices. A few categories require closer review.

Notices issued by state commercial tax authorities for legacy VAT periods may use formats that depart significantly from GSTN or CBIC templates. Extraction accuracy for these is lower, and the system flags them with a manual-review prompt rather than returning a clean-looking but unreliable output.

Composite notices — a single document consolidating demands across multiple tax periods and multiple heads with differential interest calculations — may return partially structured output. The extraction will identify what it can with confidence and will not invent values for fields it cannot resolve from the document. This is a deliberate design choice: a partial, honest output is more useful than a complete, unreliable one.

Handwritten or typewriter-issued notices from smaller Range Offices, while processable through the optical recognition layer, show higher variability in output quality depending on scan resolution.

The system's response in all three cases is the same: surface uncertainty explicitly rather than return an apparently complete output that contains an error. For an advocate whose professional standing depends on accuracy, that design choice matters more than a superficially clean interface.

How This Sits Within the Broader Statutory Framework

Why the Invoked Section Matters

When a notice arrives under Section 74 of the CGST Act alleging fraud or wilful misstatement, the stakes at the extraction stage are especially high. Section 74 allows the department to demand tax, interest, and a penalty of up to one hundred percent of the tax due, and extends the limitation period to five years. The provision invoked — Section 73 versus Section 74 — therefore determines the applicable limitation, the quantum of penalty exposure, and the legal standard the department must meet to sustain the demand.

An advocate who misreads the invoked section at intake and prepares a response calibrated to Section 73 when the notice is actually issued under Section 74 has, in effect, given substantively wrong advice on penalty exposure. The reverse error — treating a Section 73 notice as a Section 74 matter — wastes argumentative bandwidth on a fraud rebuttal the department was never required to make out.

Similarly, under the Income Tax Act, the distinction between a notice under Section 143(2) (scrutiny) and a notice under Section 148 (reassessment) is legally significant: the procedural rights available to the assessee, the time limits applicable to the authority, and the burden of proof each party carries differ materially. The Supreme Court's reasoning in Union of India v. Ashish Agarwal on the reassessment regime under the amended Section 148 / 148A framework illustrates just how much turns on correctly identifying which provision is in play. Extracting the invoked section accurately is therefore not a drafting convenience. It is a safeguard at the point where a file's legal strategy is first framed.

Faceless Proceedings and the DIN Requirement

The Faceless Assessment Scheme has added a mandatory procedural layer: every notice, order, or communication issued under the scheme must carry a computer-generated DIN, in line with CBDT Circular No. 19/2019. A notice issued without a valid DIN is, under the circular, to be treated as invalid and non-est — a point any response should record at the threshold, before engaging with the merits. LegalInk's extraction output includes a DIN field and flags its absence, prompting the procedural objection at intake rather than after the substantive reply has been drafted.

The same logic extends to GST notices issued through the common portal. A REG, ASMT, or DRC-series reference number is structurally required, and its absence or malformation is itself a ground worth flagging in the reply.

Why This Matters

Notice extraction is only useful if it is accurate, if it integrates with the drafting steps that follow, and if it is honest about what it cannot resolve. LegalInk's tool is built around those three constraints — not as a general-purpose document scanner, but as a purpose-built intake layer for Indian tax litigation practice. Advocates running volume practices, CA firms coordinating with external counsel, and in-house tax teams managing a notice register will find that the extraction feature addresses the most labour-intensive part of the notice lifecycle: the point before any legal work has started.

The broader point is structural. The first ten minutes spent on a notice — identifying the section, the period, the quantum, the deadline — are the ten minutes most likely to determine whether the rest of the file is built on accurate foundations. Automating that step is less about speed than about consistency: every file in the practice starts from the same structured baseline, regardless of which junior opened it or how the notice arrived. The extraction tool, the compliance framework, and the brief generator sit alongside each other at legalink.co.in, and together they cover the notice lifecycle from intake to appellate brief without the file passing through a typing or re-keying stage.

Related posts

← Back to all posts