Understand the work.
Make the next
decision clearer.

Choose the account and period, then collect the bank and ledger records covering that same period. Keep the original files so the imported lines can be traced back to their source.

Check opening and closing balances, currency, date conventions and whether amounts are signed or split into debit and credit columns. Review the column interpretation before proceeding.

Inspect the records on both sides, their references and the matching reason. A group match can link several records when the combined amounts agree.

A confidence score helps prioritise review; it is not a guarantee. Account policies determine which eligible suggestions can be confirmed automatically. Differences must still be resolved.

Start with the source record, amount and date. Brillrec can surface possible counterparts, grouped amounts, charges, duplicates, reversals and timing patterns within the imported records; AI can summarise those findings.

Use them as a starting point. Check further supporting evidence and establish the cause before deciding how to treat the difference.

Record the explanation and supporting context, assign investigation where needed and route sensitive decisions for approval. Keep the explanation connected to the reconciliation.

Explain differences and complete required approvals. Review the statement, outstanding items and supporting evidence before sign-off. Whether items may remain open depends on your organisation’s policy and any required explanation.

Sign-off locks the reconciliation. Preparation and approval responsibilities should follow the organisation's configured controls; subsequent reopening follows a controlled workflow.

Inspect suggested matches against their source records, confirm proposed file mappings and check draft commentary before saving it. Open-item findings identify possible explanations for your team to assess.

Your organisation controls external AI use and its review mode. Discuss processing and data-handling requirements with Brillrec as part of your evaluation.

An enabled account policy applies confidence tiers, difference allowances and amount limits. The organisation also controls whether external AI review suggestions can be included. A match can still have a difference needing explanation.

A qualifying manual match can propose a name link. It is used only after an administrator accepts it, and amount and date evidence still matter. Repeated explanations can also generate rule proposals for review.

Confirmations and rejections inform a learned score used to order suggestions. The original confidence and account policy still govern automatic matching. Previously rejected pairings are remembered within the organisation.

Source handling: Can your formats be interpreted correctly, with the account, currency, dates and balances preserved?

Matching: Can reviewers explain why records match, inspect suggestions and identify differences that remain open?

Controls: Who can prepare, approve and sign off? What happens when an action requires another person or a reconciliation needs reopening?

Outputs: Does the resulting statement or schedule answer your reconciliation question, with outstanding items visible?

Evidence: Can your reviewer trace the outcome to source records, explanations, commentary and decisions?

Deployment: Which connections, hosting arrangements and external providers would your organisation use?