Skip to content

Categorization - Current Solutions & Workarounds

Tools & Systems

Cash Analytics

  • What it is: Treasury Management System
  • How they use it: Manual transaction classification interface
  • Limitations: No auto-classification, difficult to reclassify historical transactions (requires support tickets)
  • Used by: Euroports

Excel

  • What it is: Spreadsheet software
  • How they use it: Mapping tables between transaction types and budget categories; validation and review
  • Limitations: No integration, manual cross-reference
  • Used by: Euroports

Workday ERP

  • What it is: Enterprise resource planning system
  • How they use it: Source of reconciled transactions; journal entry data with memo fields for categorization
  • Limitations: Full reconciliation only available at T+11/T+12; memo text is free-form and inconsistent
  • Used by: Personio

Manual Workarounds

Manual Transaction Review

  • What they do: Review each transaction and manually assign budget category and GL account
  • Why: No auto-classification available in current tools
  • Source: Euroports (2025-10-27)

Support Tickets for Reclassification

  • What they do: Submit support tickets when historical transactions need to be reclassified
  • Why: Current tool doesn't allow self-service historical corrections
  • Source: Euroports (2025-10-27)

VLOOKUP on Memo Text

  • What they do: Use Excel VLOOKUP formulas to match bank memo fields against category lookup tables
  • Why: No automated categorization; manual matching required
  • Limitations: Free-form memo text has typos and inconsistent labeling; unreliable matches
  • Source: Personio (2024-10-03), On (2024-10-24)

SKG Payment Processor

  • What it is: Payment processor with Treasury Automation Suite
  • How they use it: Routes transactions using rule-based allocation
  • Limitations: Not smart - have to manually create all rules; many transactions fall through
  • Used by: On

Dynamics ERP "Process Bank Statement" Queue

  • What it is: Queue in Microsoft Dynamics for unprocessed/unposted transactions
  • How they use it: Transactions sit here until manually identified and booked to correct GL
  • Limitations: Transactions pile up; no visibility on actual balances until cleared
  • Used by: On

Kyriba Rule-Based Categorization

  • What it is: Rule-based transaction categorization in Kyriba TMS
  • How they use it: Creating matching rules to assign budget codes and cash flow codes
  • Limitations: Extensive manual setup and maintenance; not expected to be much better than basic rules
  • Used by: On (implementing)

"I don't think I expect you to be much better than Kyriba. So I don't know if I want to take the rule-based categorization as the source of truth" - Lucía Galán Cáceres, On

LLM-Based Categorization (Palm)

  • What it is: Machine learning approach using LLMs to categorize transactions
  • Advantages: Easier to maintain, works across languages/geographies, handles edge cases better than rules
  • Used by: On (piloting with Palm)