Content authored by practitioners with experience at Amazon India, Intuit QuickBooks, and the Tata Group. Meet the team →
Finance teams assume reconciliation software reports are audit-ready, but auditors test completeness, accuracy, access and change trails before relying on them.
Map each auditor question to evidence the tool can produce: ingestion logs, edit logs, role lists, rule history and reproducible re-runs.
An evidence table of auditor question, tool evidence, owner and frequency, plus a 25-item sample walk-through before purchase.
Confidence that the auditor can rely on the tool's reports, with the evidence pack tested before contract signature.
Auditors do not trust a reconciliation report because software produced it. They test three things: whether the report is complete and accurate, who could change the underlying data, and whether every change leaves a trace. Before you buy, check that the vendor can help you answer all three with evidence, not assurances.
For what a trail must contain, see reconciliation audit trail; for how internal auditors use reconciliations, see internal audit reconciliation.
What will an auditor test?
Completeness and accuracy. SA 500 (Audit Evidence) requires the auditor, when using information produced by the entity, to evaluate whether it is reliable enough for the purpose, including obtaining evidence about its completeness and accuracy. A reconciliation report is information produced by the entity. Expect the auditor to ask whether every bank statement row and every ledger entry for the period made it into the run.
Access. Who could upload files, change matches, override exceptions or edit rules? A shared login fails this test.
Edit log. The proviso to Rule 3(1) of the Companies (Accounts) Rules, 2014, in force from 1 April 2023, requires accounting software to record an audit trail of each change, with an edit log, that cannot be disabled. Under Rule 11(g) of the Companies (Audit and Auditors) Rules, 2014, the auditor reports on whether that feature operated throughout the year. Section 128(5) of the Companies Act, 2013 sets an eight-year retention period for books of account. Whether your reconciliation tool itself falls within these rules depends on whether it holds books of account; your auditor will apply the same logic either way.
Change control. If a matching rule or tolerance changed mid-year, which months used which rule, and who approved it?
What can the vendor give you?
Ask for five extracts before you buy:
- Edit-log extract for one month: every manual match, unmatch and override, with user and timestamp.
- User and role list showing who can do what, including maker-checker separation.
- Rule-change history: every change to matching rules and tolerances, with approver and effective date.
- File-ingestion log with row counts per file, so completeness can be checked against the source.
- Re-run of a closed month producing identical output, which shows the result does not depend on when it was run. See deterministic reconciliation.
Which evidence answers which auditor question?
| Auditor question | What to give the vendor | What a pass looks like | Owner | Frequency |
|---|---|---|---|---|
| Were all source rows included? | One month’s bank statement and ledger export | Ingestion log row counts equal source row counts | Finance | Monthly |
| Are matches accurate? | 25 matched items chosen by your auditor | Each traces to statement row and ledger entry | Finance | Quarterly |
| Who changed what? | Request for edit log on overridden items | User, timestamp, before and after values | IT and finance | On request |
| Who can change rules? | Role list request | Rule edits restricted and approved | IT | Quarterly review |
| Did rules change mid-year? | Request for rule history | Each change dated with approver | Finance | Year end |
| Can a month be reproduced? | A closed month to re-run | Identical output | Vendor | Year end |
Outside the tool: bank balance confirmations, management representations and the auditor’s own sampling. The tool supplies evidence; it does not replace those procedures.
What does a sample walk-through look like?
Illustrative case: the auditor selects 25 matched items from Q2 of FY 2026-27 bank-to-books reconciliation. For each, the trail should run: report line, then bank statement row (date, amount, UTR or reference), then ledger entry (voucher number, amount), then match record (automatic or manual, user, timestamp). Suppose 22 are automatic matches and 3 are manual. The 3 manual items should show who matched them, when and the note given; if one was matched with a ₹500 difference, the tolerance rule in force that day and its approver should be visible.
If any of the 25 cannot be traced without asking finance for another file, the auditor will widen the sample. That costs more audit hours than any feature in the demo saves.
How do you test this before signing?
Ask the vendor to produce all five extracts on your own sample data during evaluation, and show them to your statutory or internal auditor before you sign. Controls testing under ICFR uses the same evidence, so one pack serves both.
What questions should finance ask the vendor?
- Can the edit log be switched off by anyone, including your own administrators? The answer should be no, matching the intent of Rule 3(1).
- How long is the log kept? Ask for at least eight years so it lines up with Section 128(5) retention of books.
- Is a tolerance a setting with history? If your rule auto-matches differences up to ₹100 and someone raises it to ₹1,000, the log should show who, when and from which date.
- What happens to a closed period? A period lock should block edits; any reopening should need approval and appear in the log.
- What share of matches are manual? If manual matches exceed, say, 10 percent of lines in a month, auditors will sample them more heavily, so the tool should report the split.
- Where is data stored and backed up? Rule 3(5) of the Companies (Accounts) Rules, 2014 requires backups of electronic books on servers physically located in India; ask how the vendor’s storage supports that.
Record the answers and the extracts in your evaluation file. If your statutory auditor reviews them before you sign, you avoid a year-end surprise.
How Terra Insight fits
TransactIG’s bank-to-books reconciliation and TDS reconciliation produce an exception queue and audit-ready exports, with the trail described in our machine-readable evidence trail article. Audit and evidence is one of the six dimensions on our reconciliation software comparison page; bring your auditor’s questions to a representative-sample fit benchmark.