Benford's Law + period-over-period swing checks on every GL account — trial balance anomaly detection in under a minute per close.
| Code | Account Name | Opening | Debits | Credits | Closing | Pr. Closing |
|---|---|---|---|---|---|---|
| 1100 | Cash & Equivalents | 12,40,000 | 8,50,000 | 6,20,000 | 14,70,000 | 13,10,000 |
| 2200 | Trade Payables | 9,80,000 | 4,30,000 | 5,10,000 | 10,60,000 | 9,90,000 |
| 5210 | Travel & Conv. | 4,80,000 | 1,25,000 | 15,000 | 6,40,000 | 5,20,000 |
| 6001 | Legal & Prof. Fees | 6,75,000 | 21,60,000 | 0 | 28,35,000 | 6,75,000 |
| 9999 | Suspense Account | 0 | 1,87,500 | 0 | 1,87,500 | 0 |
Mid-market finance teams process 200–500 active GL accounts every month-end close, running trial balance anomaly detection the same way they have for years — by scrolling through a flat spreadsheet.
A ₹50,000 keying error in account 5210 Travel & Conveyance — where Opening ₹4,80,000 + Debits ₹1,25,000 − Credits ₹15,000 should equal ₹5,90,000 but the Closing column carries ₹6,40,000 — survives visual review and surfaces only when the auditor questions the workpaper under SA 520 (Analytical Procedures), often three to four weeks after close. Every day of delay converts a ₹50,000 keying error into a restatement risk.
A Legal & Professional Fees account that jumps from ₹6,75,000 to ₹28,35,000 — a 320% MoM increase — appears as just another row in a flat spreadsheet export. Without an automated swing threshold, no standard Excel formula differentiates a material variance from seasonal noise. Under CARO 2020 para 3(viii), unexplained material movements in cost accounts can attract a comment in the statutory auditor's report if the controller cannot produce documented flux commentary at fieldwork.
Suspense and clearing codes (account 9999 and similar) are used for entries that could not be classified at booking time. In practice, they are frequently forgotten: a ₹1,87,500 debit posted to Suspense in March rolls into April with no clearing entry and no flag. Under Ind AS 1 Para 54 and the ICAI Guidance Note on Preparation of Financial Statements, every suspense account must carry a zero balance before the period is locked — a control that most mid-market teams enforce manually, if at all.
Benford's Law predicts that in a naturally occurring set of financial balances, roughly 30.1% of leading digits will be 1 and only 5.8% will be 7. When a Marketing Agency Spend account shows closing balances that cluster on digit 7 across multiple periods — Opening ₹7,70,000; Debits ₹7,77,000; Closing ₹14,77,000 — the chi-square deviation is a recognised red-flag indicator per the AICPA Audit Guide: Analytical Procedures. Without automated statistical testing, this pattern is invisible to a controller reviewing the TB row by row and surfaces only during a forensic engagement.
The typical time a mid-market controller spends on manual trial balance review per close cycle — not because the volume is enormous, but because the five checks (math, identity, suspense, swing, Benford) are repetitive, error-prone, and produce no documented audit trail. A single correcting journal entry filed post-close because a math break was missed adds another half-day and, if material, triggers a restatement conversation with the auditor under SA 560 (Subsequent Events).
The trial balance is the foundational document of the period-close process under Ind AS, US GAAP, and IFRS. Four regulatory anchors govern the specific validation checks this workflow runs.
SA 520 requires auditors to apply analytical procedures at the planning stage (to understand the entity and identify unusual items) and as a final review (to corroborate conclusions). Period-over-period flux analysis and TB math checks are explicitly cited examples. Controllers who cannot produce a documented flux commentary trail face direct workpaper scrutiny — the working-paper export from this workflow satisfies that requirement without a manual write-up.
Ind AS 1 requires that the financial statements present a true and fair view, with no unclassified items parked in suspense or clearing accounts. Para 60 requires current/non-current segregation of every item — impossible if transactions remain unclassified at period end. A non-zero suspense balance at the time books are locked means indeterminate-value entries have leaked into one or more financial statement line items at an unknown classification.
CARO 2020 para 3(viii) requires the statutory auditor to report whether any transactions are not recorded in the books of account with proper descriptions or classifications. A Legal & Professional Fees account that triples month-on-month with no supporting documentation is precisely the kind of classification irregularity that attracts a CARO para 3(viii) comment — and an adverse or qualified CARO opinion has downstream consequences for bank covenant compliance and board filings.
The AICPA Audit Guide: Analytical Procedures designates Benford's Law chi-square testing as a recognised analytical procedure for identifying potential manipulation of numeric data in accounting records. A chi-square statistic exceeding 15.51 (8 degrees of freedom, 95% confidence) indicates a first-digit distribution unlikely to arise in naturally occurring accounting data — a finding that triggers escalation under ICAI SA 240 (The Auditor’s Responsibilities Relating to Fraud) when detected during the close process.
Seven deterministic passes from trial balance XLSX upload to a prioritised exception inbox and audit working paper — in under a minute per close cycle, with every check traceable back to the source row.
Accepts the current-period trial balance as a single .xlsx file with seven columns: Account Code, Account Name, Opening Balance, Period Debits, Period Credits, Closing Balance, and Prior Month Closing. Every non-header row is parsed as a distinct account entry without manual column mapping — the format matches the standard export from Tally Prime, Zoho Books, NetSuite, and SAP F.01 transactions.
Verifies that the sum of all Period Debits across every account equals the sum of all Period Credits — the algebraic identity that must hold for any valid set of double-entry books. A mismatch raises a FAIL severity flag immediately, reporting the net imbalance amount in rupees so the controller can locate the errant journal entry before the rest of the validation run proceeds.
For every row, the workflow computes Opening + Debits − Credits and compares the result to the Closing Balance column. Any discrepancy — whether ₹1 or ₹50,00,000 — raises a FAIL severity flag on that specific account with the exact rupee break surfaced in the exception inbox, enabling the controller to identify the keying or export error at row level rather than hunting through the full spreadsheet.
Every account code is checked against a configurable suspense and clearing pattern list (codes matching 9999, 9998, SUSP, CLEAR, and similar conventions). Any suspense-coded account carrying a non-zero Closing Balance receives a FAIL severity Must-Clear badge that blocks the working-paper export from being stamped as final — enforcing the zero-balance-at-close control required under Ind AS 1 and standard close procedures.
Computes the MoM change for every account as (Current Closing − Prior Closing) / |Prior Closing| and raises a WARNING severity flag when the change exceeds 50% or ₹5,00,000 in absolute rupees — whichever threshold is breached first (OR condition, not AND). Both the percentage swing and the absolute rupee movement are reported, and accounts with a prior-period zero balance are evaluated on the absolute threshold alone to avoid division-by-zero distortion.
Extracts the first significant digit from every non-zero Closing Balance in the trial balance, computes the observed frequency distribution for digits 1–9, and compares it against Benford expected probabilities (digit 1 → 30.1%, digit 7 → 5.8%). A chi-square statistic exceeding 15.51 (8 degrees of freedom, 95% confidence level) triggers a WARNING severity flag. Accounts with fewer than 30 non-zero balance observations are excluded because the sample is too small for statistical reliability.
All flagged accounts are assembled into a severity-ranked inbox — FAIL items listed first (must resolve before close), WARNING items below (require explanation or acceptance). The full working paper is exported to Excel with per-row columns for account code, account name, validation name, severity, computed values (swing %, rupee swing, chi-square statistic, digit frequency table), and a PASS / WARNING / FAIL outcome badge — formatted for direct attachment to the statutory audit close package under ICAI SA 230 (Audit Documentation).
Five synthetic test scenarios that exercise every failure mode observed in real-world trial balance data. Each produces a deterministic outcome an auditor or controller can verify in seconds against the bundled sample files.
tb_math_ties for account 5210; ₹50,000 break surfaced in the exception inbox; close sign-off blocked until corrected.period_swing_threshold; swing reported as +320% / +₹21,60,000; controller prompted to attach a drill-down explanation before the working paper is exported as final.suspense_zero; red Must-Clear badge on account 9999; working-paper export blocked from final stamp until balance is cleared.benford_first_digit; chi-square value and observed vs. expected digit-frequency table reported; account tagged for auditor review in the working-paper export.entries_list_non_empty; workflow surfaces "No data rows extracted" and routes the file for manual review — preventing a silent pass on an empty dataset.Five seeded trial balance files — each engineered to isolate one exception class. Together they cover every validation pass the workflow runs.
Use to verify zero-flag behavior. All five validation passes return PASS — math ties, debits equal credits, no suspense balance, swings within threshold, and Benford distribution normal. Correct baseline for regression testing after configuration changes.
Demonstrates tb_math_ties FAIL and the exact break amount in the exception inbox. Opening ₹4,80,000 + Debits ₹1,25,000 − Credits ₹15,000 should equal ₹5,90,000 but Closing reads ₹6,40,000 — a ₹50,000 discrepancy consistent with a keying or export error in the ERP.
Demonstrates the period_swing_threshold check and drill-down swing calculation. A Legal & Professional Fees account that more than quadruples in a single month requires documented flux commentary — the controller must attach a rationale before the working paper can be exported as final.
Demonstrates the suspense_zero FAIL and Must-Clear badge. Account 9999 opened at nil, received ₹1,87,500 in debits during the period with no offsetting credit — the standard symptom of an unclassified entry parked at booking and forgotten before month-end lock.
Demonstrates the benford_first_digit WARNING and the observed vs. expected digit-frequency table. Account 5340 shows Opening ₹7,70,000; Debits ₹7,77,000; Closing ₹14,77,000 — a concentration on leading digit 7 that is statistically improbable in naturally occurring accounting data and requires auditor review before the period can be closed.
Replacing the manual trial balance review with this automated workflow compresses what is typically a two-to-three-day process into a single-pass run that completes in under a minute — catching five distinct exception classes that visual review routinely misses.
Computing Opening + Debits − Credits = Closing for every account row deterministically surfaces arithmetic breaks — the ₹50,000 discrepancy in account 5210 that would survive visual review and surface only when the statutory auditor applies ICAI SA 520 analytical procedures at fieldwork, by which time a correcting JE and restatement conversation may be unavoidable.
Benford chi-square testing — a recognised analytical procedure in the AICPA Audit Guide — is applied automatically to every close, not just during annual forensic audits. The digit-7 cluster in account 5340 Marketing Agency Spend would be invisible to a controller reviewing the TB row by row; the workflow quantifies the deviation and surfaces it with the chi-square value and observed vs. expected frequency table for auditor review.
The Excel export — with per-account validation outcomes, swing calculations, chi-square values, and PASS / WARNING / FAIL badges — satisfies the analytical-procedure documentation requirement under ICAI SA 230 (Audit Documentation) and drops directly into the close package shared with statutory auditors, reducing the documentation step from a manual write-up to a single file attachment that requires zero reformatting.
The questions controllers, CFOs, and auditors ask most often before deploying automated trial balance scrutiny.
The TB math and suspense-account checks align with the ICAI Guidance Note on Audit of Internal Financial Controls and support documentation requirements under SA 315 (Identifying and Assessing the Risks of Material Misstatement) and SA 520 (Analytical Procedures). The Benford's Law test is a recognized analytical procedure documented in the AICPA Audit Guide: Analytical Procedures and is increasingly cited in forensic accounting literature as a first-pass fraud indicator. Under Ind AS and US GAAP engagements, the period-over-period flux analysis satisfies the flux commentary requirement that most auditors request during interim and year-end close package review.
Benford's Law predicts that in naturally occurring accounting data, the leading digit 1 appears ~30.1% of the time, digit 2 ~17.6%, down to digit 9 ~4.6%. The workflow extracts the first significant digit from every non-zero closing balance, computes observed vs. expected frequencies, and calculates a chi-square statistic with 8 degrees of freedom. A result exceeding the default threshold (chi-square > 15.51 at the 95% confidence level) triggers a WARNING flag; the threshold is configurable per engagement. Accounts with fewer than 30 non-zero balance observations are excluded from the Benford test because the sample size is insufficient for statistical reliability.
By default, the workflow flags any account whose closing balance changes by more than 50% or ₹5,00,000 MoM, whichever is breached first — an OR condition, not AND. Both the percentage and the absolute rupee threshold are configurable in the workflow settings, allowing controllers at different company sizes or with different materiality levels to calibrate sensitivity. The swing is computed as (Current Closing − Prior Closing) / |Prior Closing|; accounts with a prior-period closing of zero are evaluated solely on the absolute amount to avoid division-by-zero distortion.
Yes. The input is a single .xlsx file with seven columns (Account Code, Account Name, Opening Balance, Debits, Credits, Closing Balance, Prior Month Closing), a format that can be exported directly from Tally Prime's Trial Balance with Opening/Closing report, Zoho Books' Trial Balance export, NetSuite's Financial Statement Builder, or SAP's F.01 / S_ALR_87012284 transaction. No ERP connector or API integration is required; the controller downloads the standard export and uploads it to Cadel.
Every run produces a timestamped Excel working paper recording: (1) each account's raw input values, (2) the computed math-tie result, (3) the MoM swing percentage and absolute amount, (4) the Benford chi-square statistic and per-digit observed vs. expected table, and (5) the severity badge (PASS / WARNING / FAIL) for each of the five validations. The file meets the ICAI standard under SA 230 (Audit Documentation) and attaches directly to the close package shared with statutory auditors. The Cadel platform also logs the user, timestamp, and input file hash for each run, providing an immutable review trail.
The current workflow processes a single-entity, single-currency trial balance per run, with all amounts expected in Indian Rupees. For multi-entity consolidations, controllers should run one instance per legal entity before consolidation entries are posted — the recommended practice under Ind AS 110 (Consolidated Financial Statements) to catch entity-level anomalies before they are obscured in the consolidated view. Multi-currency support — where each row carries an original-currency amount alongside an INR equivalent — is on the Cadel product roadmap; in the interim, the workflow can be used on the INR-translated TB without any loss of math-tie or swing-check functionality.
Yes. The working-paper export — with per-account validation outcomes, swing calculations, and chi-square values — provides the documented analytical procedure trail that SOX Section 302 certifying officers and CARO 2020 para 3(viii) statutory auditors require. An unexplained large swing in a cost account with no supporting documentation is precisely the finding that triggers a CARO para 3(viii) comment or an internal control weakness under SOX; this workflow ensures the documentation exists before the auditor asks. For US-listed Indian entities filing both MCA forms and SEC disclosures, the same export serves both requirements without duplication.
Accounts with a prior-period closing balance of zero are excluded from the percentage-swing calculation to avoid division-by-zero distortion. They are still evaluated against the absolute rupee threshold (₹5,00,000 by default) — so a newly opened account that immediately accumulates a large balance will still be flagged on the absolute test. New suspense-coded accounts with any non-zero closing balance are always flagged under the suspense_zero check regardless of prior-period value, because a first-period suspense debit with no prior history is still an unclassified transaction that must clear before the books are locked.