When a payment arrives — and nobody knows what it is for
An invoice is overdue, several dunning notices are ready, then a payment appears on the bank statement but the purpose field only contains a short code or booking reference. No one has responded in the email inbox. Such everyday situations are common in many B2B companies and consume time: searching accounts, matching invoices, allocating partial payments.
A structured open‑items workflow with automatic payment matching significantly reduces this work — provided the rules are well thought out and keep clear human control. Below I show practically how Debentra PRO uses rule‑based matching templates, how trust scores are constructed and when accountants should approve manually.
What good automatic matching must achieve
- Detect unambiguous cases (e.g. exact invoice number + amount) and assign immediately.
- Detect partial payments, instalments or rounding differences and flag them with appropriate handling.
- Clear priorities: reference > amount > historical allocation; rules must be prioritized.
- Transparent trust values so accountants can intervene by exception.
How Debentra PRO builds matching templates
Debentra PRO works with rule‑based templates that perform several checks in sequence. Typical components of a template:
- Regex extraction: Pulls structured content from the purpose field (e.g. INV‑123456, KND/98765). Regex templates can be created for internal invoice number formats — e.g. \bINV-?\d{6}\b for invoice numbers with 6 digits.
- Amount patterns: Checks whether the paid amount exactly matches an open item, slightly differs (e.g. cent rounding) or is a partial payment.
- Reference patterns: Searches for customer numbers, order numbers or internal references, even when these do not follow the invoice format.
- Priorities: Rules are executed in a specific order; the first valid rule can dominate, others provide additional scores.
- Historical allocation: Recognizes recurring payer IBANs or purpose texts and increases trust accordingly.
Trust score: What it is and how it is created
Debentra PRO assigns a trust score to every automatic allocation. This score summarizes the results of the rule checks and makes the decision traceable. Important components of the score are:
- Reference hits: Direct hit on invoice number or unique customer number (high weight).
- Amount equality: Exact amount vs. expected invoice amount.
- Partial payments: Detection whether amount matches an installment plan or partial payment pattern.
- Temporal plausibility: Payment within the expected timeframe (e.g. after invoice date).
- Historical data: Match with known payers (e.g. IBAN, payment reference patterns).
The individual components in Debentra PRO have weights that are adjustable. Recommended approach: define conservative default weights and clear thresholds (e.g. Score > 90 = automatic assignment, 60–90 = proposed for review, <60 = manual).
Concrete examples — what happens in practice?
Example 1 — Clear hit: Payment with "INV-000123" and exactly 1.250,00 €:
- Regex extracts invoice number → reference hit.
- Amount matches exactly → high score.
- Historical payer known → additional bonus.
- Confidence score high → automatic assignment, accountant can optionally review.
Example 2 — Partial payment/rounding difference: Payment 624,99 € for an open item of 625,00 €:
- Amount deviates slightly → medium score.
- No clear reference hit → score in the middle category.
- Debentra PRO can detect partial payment patterns and mark the item as "probable partial payment"; proposal for manual approval.
Example 3 — Ambiguity: Payment reference contains only a customer number that nevertheless has multiple open items:
- Multiple possible allocations → low score.
- System flags the transaction for manual review and displays the relevant options with contextual elements (amount history, recent payments).
When should accountants approve manually?
Automation should save time, not relinquish responsibility. Typical situations for manual approval:
- Score below the defined threshold (e.g. < 60) — ambiguous cases.
- High amounts: payments above a defined threshold should always require manual confirmation.
- New customers or changed account details — increased caution with unusual payment behavior.
- Partial payments that cannot be clearly assigned to an installment plan.
- Discrepancies between payment reference and expected information (e.g. different customer number, differing currency).
Important: Debentra PRO keeps the approval in the hands of the user — automatic suggestions do not require full automated execution rights if a company does not want that.
Tips for building robust matching templates
- Start conservatively: Initially set restrictive rules with a low automation threshold and expand them only after gaining reliable experience.
- Use multiple small regex templates instead of one complex super‑regex. Smaller patterns are easier to test and maintain.
- Prioritize rules clearly: exact invoice number matches take precedence over amount equality.
- Normalize inputs: Remove spaces and special characters and unify date formats to reduce false matches.
- Log and audit decisions: Every automatically matched item should be documented and traceable.
Typical mistakes and how to avoid them
- Too many automatic rules without a review path → risk of incorrect allocations. Solution: score thresholds and review queues.
- Regex too broadly defined → incorrect extractions. Solution: test patterns on historical payment texts.
- No consideration of partial payments → lots of manual rework. Solution: include installment plans and partial sum patterns in rules.
- No escalation logic for high amounts → compliance and control risks. Solution: define amount limits that require mandatory manual approval.
Practical checklist for rollout in your team
- Identify common use cases and invoice number formats.
- Create and test regex templates on historical data.
- Define trust score weightings and thresholds.
- Set up review queues for medium and low scores.
- Train accountants: when to check, when to approve — and how to document cases.
Debentra PRO in practice: integration into the AP workflow
Debentra PRO integrates into the accounts payable flow: import invoice → detect open items → apply matching template → calculate score → automatic allocation or review queue → prepare and approve payment reminder → detect payment and document the transaction. Control remains with the user: approvals, thresholds and priorities are configurable so automation can be aligned precisely with your internal processes.
Conclusion
Rule‑based matching templates and transparent trust scores make automatic payment matching practical and secure. Debentra PRO provides the building blocks: regex extractions, amount and reference patterns, priorities and configurable scores. The key is a conservative rollout with clear review rules and defined escalation points — this way automation reduces workload without relinquishing control.
Act now: Try the editions and choose the right configuration for your team. Choose an edition: https://debentra.de
Next step
Try Debentra yourself
Central overview and process control for open items and payments while retaining full data sovereignty; reduces manual allocation work (PRO) and enables integrated accounting (PREMIUM); selectable installation and licensing models.