Treasury Data - Pain Points & Challenges¶

Pain Points¶

Data quality issues discovered too late — downstream analytics polluted before anyone notices; no pre-ingestion validation
Source: Feature file
Can't trace data back to source — when numbers don't match, no audit trail to investigate
Source: Feature file
Kyriba rejects data if bank account not assigned — integration rules cause silent data drops; flows without assigned bank accounts are not integrated
Source: ON (2025-10-02) — "Kyriba rejects [data]. So for example, we have rules based on Kyriba bank accounts. If any of the flows don't have a bank account assigned, it will not be integrated."

Looker/ERP not real-time — Dashboard data depends on bank reconciliation, not live feeds
Source: On (2024-09-26)
Source: On (2024-10-24) — "Right now until the end of the month, we don't really know how much money we have"
External reconciliation creates delays — Bank reconciliation done by Deloitte center in Madrid, can take days for spots to appear
Source: On (2024-09-26)
Monthly reconciliation cadence — Reconciliation only done once a month, would prefer daily
Source: On (2024-10-24) — "It will be much more often and not only once a month, but maybe every day"
No visibility into data freshness — teams don't know when data is stale or missing until someone checks manually
Source: Feature file

Different banks, different formats — MT940, CAMT, BAI2, CSV all need harmonization into a single schema
Source: Feature file
Bank connectivity projects take months — integration projects are heavyweight and slow; each new bank is a new project
Source: Feature file
Single GL account per currency — ERP structure prevents drilling down to individual bank account balances
Source: On (2024-09-26)

Kyriba deletes ALL forecast data daily — No smart updates; daily wipe and replace causes missing data during training sessions
Source: ON (2025-10-02) — "We're training the regions now... and the data is not there because that day because we delete"