VAT-06 · FlowCount VAT

VAT return review checks

An independent review of a prepared return, 30 to 50 checks deep.

مراجعة مستقلة لإقرار ضريبة القيمة المضافة قبل تقديمه

Design partners wantedFor: Tax & accounting practices · In-house VAT teams

30 to 50

configurable checks on every prepared return

Basis: property of the method

The problem

The reviewer gets the prepared return, the listings and a deadline. They test a sample, tick the summary and sign. Sign conventions differ by ERP and by client, so half the review notes are about minus signs rather than tax.

“Negative sales are not an error. Inconsistent signs are.”

What FlowCount does

Run a configurable library of 30 to 50 control checks over the prepared return and its source data, score every exception high, medium or low, and produce reviewer comments with the evidence attached.

AI and rules used
Rules; LLM drafting of review notes grounded in check results
Builds on
VAT return ↔ GL ↔ e-invoice reconciliation

How it works

Who does each step

  1. 01Rules

    Detect the sign convention

    From account type, tax code, return box and history. Raw values are kept; tests run on the economic sign.

  2. 02Rules

    Run the checks

    Payable movement, output and input VAT, negative values, zero-rating, prior-period adjustments, reverse charge, the VAT control account and e-invoice differences.

  3. 03AI

    Draft review notes

    One comment per exception, citing the lines and the check that fired.

  4. 04Reviewer

    Close and sign

    A named reviewer closes every high-risk item and signs the review.

What you send

  • Prepared VAT return
  • GL and TB
  • Sales and purchase registers
  • E-invoice population
  • Prior returns

What you receive

  • AI reviewer report
  • Control checklist with results
  • Prioritised exceptions with evidence

Controls built in

  • Every check versioned with an effective date
  • High-risk items closed by a named reviewer
  • Check library configurable per client and per firm

KPIs we report

  • Review hours per return
  • High-risk items found after filing (target 0)

Sample input → expected output

What the result looks like

Basis: property of the method

Checks drawn from real Saudi VAT working files (excerpt)
CheckWhat it testsTypical finding
Advance paymentsVAT is due when an advance is received; each advance is tracked to the invoice that later settles itAdvance declared twice: on receipt, and again on the final invoice
Export evidenceEach zero-rated export matched to its customs export declaration, with the invoice converted to SAR at the declaration dateDeclared value differs from the invoice; an export with no declaration on file
Late input claimsInvoice date against posting date and the claim windowSupplier invoices from November claimed two quarters later
Prior-period credit notesCredit notes linked to the original invoice and its periodA credit note reduces this quarter for an invoice declared last year
Intra-group reverse chargeServices from group entities abroad self-assessed at 15%Credit memos from a foreign affiliate not reversed in the reverse charge
Sign conventionConsistency of signs by box, not the sign itselfA credit note entered with positive values
Payable movementVAT payable on the return against the movement on the VAT control accountA manual journal to the control account with no invoice behind it
Rate by supplyThe rate applied against the nature of the supply5% on a standard-rated domestic sale

Portfolio scorecard

Where it sits on the roadmap

Value

4.30 / 5

Feasibility

4.20 / 5

Priority index

9.03 · rank 4 of 27

Quadrant

Quick win

Delivery

Wave 1 (0-3 m)

T-shirt size S: 2 to 4 weeks with 1-2 engineers + 0.5 tax SME, indicative build 50k to 120k. New rule set on the existing data model, one new input. How we score and sequence →

An independent review of a prepared return, 30 to 50 checks deep.

Design partners shape the rulebook and review console, and get first access on pilot terms.