Checks are scattered.
Invoice, PO and receipt data often sit in separate exports. A status or total alone does not tell a reviewer which row disagrees or why.
Accounts payable / Evidence-led reconciliation
Find the discrepancy. Follow the evidence.
Compare supplier invoices with purchase orders and goods receipts. See which records disagree, which rule triggered, and the source values behind each finding—so your team can review exceptions with evidence instead of chasing spreadsheets.
SAMPLE WORKFLOW · READY TO EXPLORE
Start with a working reconciliation. Open a finding to inspect the source values, matching rule and evidence trail—then decide whether the workflow fits your review process.
LEDGERTRACE / PRODUCT SYSTEMFROM RAW RECORDS TO REVIEWABLE EVIDENCE
ILLUSTRATIVE WORKFLOW · SAMPLE RECORDSTRACEABLE FINDINGS · HUMAN REVIEW
Evidence, not a black-box scoreEach finding points to source fields and the rule that triggered it.
Designed for AP reviewInspect exceptions, compare records and export a review worklist.
Browser-side sampleUploaded CSVs are processed locally in this version.
02 / Working sample
See the reconciliation, not just the promise.
Load the illustrative records or choose your own three CSV exports. Run reconciliation to generate matched lines and flagged exceptions, then select a result to inspect the invoice, PO and receipt evidence.
Supplier invoicesSOURCE A / CSV
Invoice ID, supplier, PO reference, item, quantity and unit price.
Purchase ordersSOURCE B / CSV
PO line, supplier, item, ordered quantity and agreed unit price.
Goods receiptsSOURCE C / CSV
Receipt identifier, PO line, item and quantity received.
Choose all three CSV files, or load the sample, then run reconciliation.
03 / Who it serves
For finance teams that need to explain a mismatch—not just flag one.
Designed for accounts-payable leads, controllers, bookkeeping firms and finance service providers reviewing supplier invoices against purchase orders and goods receipts. Most relevant where spreadsheet checks recur and reviewers need a defensible trail back to source rows.
One reviewable trail.
Compare configured fields, see the triggered rule, inspect source-row references and export exceptions for human review.
Keep the decision with your team.
LedgerTrace organizes rule-triggered discrepancies and supporting records. Your reviewer determines whether a finding needs correction, explanation or no action.
What you can inspect
Every flagged result has something concrete to review.
- Compared values from the supplied records.
- The rule that triggered each finding.
- Source-row references and related PO/receipt rows.
- An exportable exception list for human review.
Data handling / Current build
CSV processing stays in this browser tab.
The app reads selected CSV files with browser file APIs and runs its reconciliation logic locally. Its current code does not upload those CSV contents to a LedgerTrace server. Review notes are held in page memory for this session. The site loads web fonts from Google; this is not a claim that the entire browsing session makes no network requests. Use synthetic or approved anonymized records for first evaluation.
A bounded evaluation
Start with one workflow. Validate every finding.
Use approved anonymized CSV exports to test one invoice-to-purchase-order-to-receipt workflow. The evaluation compares tool output with a human-reviewed reference set and documents what was caught, missed, or incorrectly flagged.
- Agree the workflow.Choose one AP process, supported CSV columns and the exception types that matter.
- Approve the boundaries.Confirm sample size, exclusions, data handling and acceptance criteria in writing.
- Run the evaluation.Process up to 500 combined rows after clean inputs and scope approval.
- Review the evidence.Receive an exception CSV, findings report and 30-minute handoff within five business days.
A scoped paid evaluation · Terms agreed privately
01Scope
One AP workflow, up to 500 combined CSV rows, and agreed checks.
02Agree terms
Confirm the fee, timeline, deliverables and acceptance criteria with LedgerTrace before work begins.
03Payment
Written scope and acceptance criteria first. Payment through an agreed, legitimate route before work begins.
04Evidence
Exception CSV, findings report, false-positive review and a 30-minute handoff.
The evaluation is scoped with each buyer before work begins. Use synthetic or client-approved anonymized data. LedgerTrace does not claim production-grade security, ERP integration or guaranteed detection. Fees, timeline, deliverables and acceptance criteria are agreed in writing; scope changes require written agreement.
04 / Buyer assurance
Proof before promises.
No inflated savings claims or black-box verdicts. Inspect the sample, define what a correct result looks like, and discuss a bounded paid evaluation with explicit deliverables and limitations.
Try the browser demo first
The included sample runs locally in your browser. Review the inputs, triggered rules and exported exceptions before sharing any business data. No account or upload is required to explore it.
Define the expected answer
Before evaluation, agree the sample/reference cases, supported CSV columns, checks in scope and acceptance criteria. Ambiguous records stay reviewable; they are not silently treated as resolved.
Report misses as well as catches
The report should distinguish confirmed findings, false positives and missed exceptions against the agreed reference set. If no suitable reference set exists, we define it with you before making accuracy claims.
Commercial honesty
LedgerTrace is an early, static browser-based pilot—not a certified accounting system. There are no claimed customer results, accuracy rate, savings figure, security certification or production readiness. A paid evaluation covers the agreed assessment, not a guaranteed saving or positive finding. You receive the agreed outputs even if the tool misses cases or produces false positives.
05 / Methodology
A finding should explain itself.
LedgerTrace applies explicit comparisons to the records provided. It does not infer intent, decide whether an invoice is fraudulent, or replace an AP review process.
Connect the source rows
Normalize identifiers and compare invoice lines with purchase-order and receipt rows using PO number and item code. Missing or ambiguous relationships must remain visible.
Show the rule and values
Flag configured conditions such as duplicate invoice IDs, supplier differences, price deltas, quantity excesses, missing receipts and unmatched source rows.
Keep the human decision
Inspect the evidence, record review notes and export the worklist. A finding is not a payment instruction, and opening it does not resolve it.
06 / Product details
Clear answers before you evaluate.
The important details about matching, inputs, evidence and review are spelled out before you decide whether to use LedgerTrace.
Which records can I compare?
Three CSV exports: supplier invoice lines, purchase-order lines and goods-receipt lines. Each file should include the relevant identifiers, item code and quantity; invoice and PO files should include unit price for price checks.
How are records matched?
The current matching key is purchase-order number plus item code. LedgerTrace then checks configured conditions such as duplicate invoice identifiers, supplier differences, invoice-versus-PO price differences, quantity excesses, receipt gaps and unmatched PO lines. Confirm column mappings and edge cases before using a new data shape.
What can I inspect in a finding?
Select a result to review its invoice row, related PO and receipt rows, compared values, triggered rule and explanation. The workbench supports review notes and CSV export so a person can investigate and record the next step.
Where is uploaded data processed?
In the current version, selected CSV files are processed in your browser; they are not uploaded to a LedgerTrace application server. Use synthetic or approved anonymized data for evaluation, and do not provide confidential records until data handling has been agreed.
What needs separate validation?
Currency conversion, tax and credit-note treatment, split or merged invoice lines, and complex supplier-resolution cases are not covered as general capabilities. A “matched” status means no configured rule triggered; it is not a certification that the invoice is correct or approval to pay.
How do I know the results are useful?
Use a human-reviewed reference set with expected outcomes. Compare findings against that set, record both missed exceptions and false positives, and agree the scope and acceptance criteria before a paid evaluation. Illustrative sample records demonstrate the workflow; they are not customer results.
06 / Paid evaluation
Evaluate it against a known answer.
Five-business-day AP reconciliation evaluation. One invoice-to-PO-to-receipt workflow, up to 500 combined CSV rows, and a fixed fee quoted in writing before kickoff. The five-day clock starts when the agreed files and scope are ready. No payment or accounting action is automated.
You receive: a source-linked exception CSV, a concise findings report, comparison against an agreed human-reviewed reference set, a false-positive and missed-case summary, and a 30-minute review handoff. Acceptance criteria are agreed before work starts: each agreed test case is processed, every reported finding includes its rule and source-row references where available, and the report records known misses and false positives against the reference set.
Agreed reference set with human-reviewed expected outcomes
Results separated into matches, false positives and misses
A traceable report and agreed review criteria
Written scope, agreed exclusions and acceptance criteria
Evaluation scope builderTERMS AGREED PRIVATELY
Step 01 / Source records
Step 02 / Checks to discuss
These are scope topics, not a guarantee that every check is fully supported in every data shape. Confirm edge cases and acceptance criteria before quoting.
Step 03 / Review the brief
Proposed evaluation briefDraft
Source types3 selected
Checks6 selected
DeliverablesFindings report + CSV
Commercial termsAgreed together
Scope, fee, timeline and acceptance criteria are agreed before work begins. Do not send confidential files before a written agreement.