Free Download · Excel workbook
Bank Statement Narration Parser Workbook
Paste a raw HDFC, ICICI, SBI, Axis, or Kotak bank statement into one tab and get back a parsed transaction table with UTR, UPI VPA, NACH mandate ID, and cheque number extracted into their own columns from the narration. XLOOKUP-based, no VBA, every formula visible and editable. Ships with illustrative data so you can see it working — replace with your statement to run your own bank leg.
What's inside — 7 components
Six narration-structure recognisers
One dedicated recogniser per payment rail: NEFT, RTGS, IMPS, UPI, NACH, and cheque. Each detects the rail from its narration signature (prefix tokens, delimiter shape, presence of a UTR vs a VPA vs a mandate ID) and routes the row to the right extraction column before any parsing runs.
UTR extraction — TEXTBEFORE + TEXTAFTER
A 12/16/22-character reference is lifted out of NEFT, RTGS, and IMPS narrations using TEXTBEFORE / TEXTAFTER against the rail-specific delimiters your bank prints. No hard-coded character offsets — the formula is readable and holds up when the bank shifts its narration format by a few characters.
UPI VPA extraction — MID + FIND
The virtual payment address (something@bank) is parsed out of UPI narrations with a MID + FIND pair that locates the "@" and reads backward to the last space or delimiter. Works across UPI/CR, UPI/DR, and the p2m/p2p variants HDFC, ICICI, and SBI each format differently.
NACH mandate ID + batch reference parsing
The mandate UMRN and batch reference number are pulled out of ACH/D and ACH/C narrations for reconciling debit/credit runs against the sponsor bank return file. Both fields land in their own columns so the workbook feeds directly into a NACH bounce and success reconciliation.
300+ column-variant tolerance across HDFC / ICICI / SBI / Axis / Kotak
Column-name normalisation table maps the many header variants each bank prints (Narration vs Description vs Particulars vs Remarks; Withdrawal Amt vs Debit vs Dr; Deposit Amt vs Credit vs Cr; Value Date vs Txn Date; Reference No. vs Ref No vs Chq/Ref No) into a single canonical schema before parsing runs.
Error-trap IFERROR for unrecognised patterns
Every extraction formula is wrapped in IFERROR so a narration the parser cannot classify lands in an "Unrecognised" bucket rather than crashing the sheet with #VALUE! or #N/A. The Unrecognised tab is your review queue — sort by frequency, add a rail signature to the recogniser table, and the pattern is picked up on next refresh.
Named ranges + all formulas visible and editable
Every recogniser, delimiter, and column mapping lives in a named range on the Config tab — no hidden constants, no VBA, no macros. If your bank prints a UTR after "TRF-" instead of "UTR-", you edit one cell. Anyone comfortable with Excel can maintain this without unwinding nested INDIRECTs.
Illustrative data included.
The workbook ships pre-populated with a sample two-week paste from a fictitious HDFC
current account — mixed NEFT, RTGS, IMPS, UPI, NACH debits, and one cheque — so every
recogniser, delimiter, and extraction column is visibly wired up when you first open
it. A REPLACE WITH YOUR STATEMENT DATA note sits on the input tab —
clear the sample rows before pasting your own statement.
Get the Workbook
The download appears immediately below — no wait, no follow-up call.
Your Workbook is ready
Download bank-narration-parser-workbook.xlsx →Submission failed. You can still download the Workbook directly.
Where this workbook fits
This is the operational asset for the first leg of a monthly bank close. Read the full formula-by-formula recipe, the runbook it slots into, or the pillar the whole methodology sits under.
Bank Narration Parsing in Excel — Formula-by-Formula Recipe
The step-by-step article the workbook operationalises: which TEXTBEFORE / MID / FIND combination extracts what from which rail, with worked examples for each of the five banks.
Bank Reconciliation Runbook — Days 1–5
The parent monthly runbook this workbook plugs into: how narration parsing sits inside the first five days of a bank leg close, and what the parsed columns feed into next.
Monthly Close Playbook — India Pillar
The umbrella methodology: how the bank leg fits alongside the GST, TDS, and platform-settlement legs of a monthly close.
Bank Statement Narration Patterns — Reference Library
The lookup library the parser is built against: narration signatures, prefix tokens, and delimiter shapes for HDFC, ICICI, SBI, Axis, and Kotak across NEFT / RTGS / IMPS / UPI / NACH / cheque.
Three-Way ITC Reconciliation Workbook
Sibling free workbook: purchase register vs GSTR-2B vs IMS actions in one Excel file, with dropdown categorisation and a Section 16(4) countdown.
Controller's Toolkit
Companion free bundle: close checklist, variance-classification reference, CSV templates, calculator index.
Reconciliation Software (India)
The automated equivalent — what live narration parsing across every bank account looks like when a workbook stops keeping up with volume.
Frequently Asked Questions
Is the workbook really free?
Yes. There is no charge and no trial period. The download link appears immediately on this page after you submit the form — nothing is emailed to you first, and there is no waiting period.
Which banks does it work with?
The recogniser tables ship with narration patterns for HDFC, ICICI, SBI, Axis, and Kotak — the five banks that between them cover the current-account footprint of most Indian mid-market companies. Every recogniser is a named range on the Config tab, so if your bank is not one of the five (or your account uses a legacy narration format), you add a row with its rail signature and delimiter, and the parser picks it up on next refresh. No VBA, no add-in — just Excel formulas.
What happens after I submit the form?
The download link appears immediately on this page — nothing is emailed to you first, and there is no waiting period. We use your details to send an occasional update on Indian reconciliation (statutory changes, new calculators, worked examples); you can ignore or unsubscribe from those at any time without losing access to the download you already have.
What if the parser cannot recognise a narration?
Every extraction formula is wrapped in IFERROR, so an unrecognised row lands in an "Unrecognised" bucket rather than crashing the sheet with #VALUE! or #N/A. Sort the Unrecognised tab by frequency, look at the top offenders, add their rail signature to the Config tab's recogniser table, and the pattern is picked up on next refresh. This is how you tune the parser to your bank's specific narration quirks.
Do I need VBA or macros enabled?
No. The workbook is pure Excel formulas — XLOOKUP, TEXTBEFORE, TEXTAFTER, MID, FIND, IFERROR, and named ranges. No VBA, no macros, no add-ins, no external data connections. Opens and runs identically in Excel 365, Excel for Mac, and LibreOffice Calc (a couple of the newer text functions require the 2021-or-later formula engine).
Disclaimer.
The pre-loaded rows in the workbook are illustrative only — invented account numbers,
invented UTRs, invented VPAs, invented mandate IDs — so every recogniser, extraction
column, and error trap has something to compute against on first open. Replace with your
statement data before running an actual bank leg. Narration formats change over time as
banks migrate systems; if a rail signature stops matching, add the new pattern to the
Config tab's recogniser table.
When the workbook stops scaling
Excel parses one bank account for one month cleanly and five accounts slowly. Past that — or once narration parsing needs to run continuously as statements land daily rather than as a paste-and-refresh — a live reconciliation platform makes sense. TransactIG configures in 2–4 weeks. ISO 27001:2022 certified.