31 assets. All 31 are classified CRITICAL and all 31 have an empty owner field. Twenty two of them sit in PROD-CORE-VLAN-100.
| asset_id | asset_name | network_segment | last_review_date |
|---|---|---|---|
| ASSET-STO-0067 | SAN Frame Muenchen | DR-VLAN-400 | 2024-11-02 |
| ASSET-DB-0233 | Kartenzahlung DB Replica | PROD-CORE-VLAN-100 | 2025-06-14 |
| ASSET-NET-0104 | Core Switch Stack B | PROD-CORE-VLAN-100 | 2023-09-30 |
3 of 31 shown.
Where the answer came fromict_asset_inventory, bound to itsm_prod.SYS_INVENTARcriticality_classification = CRITICALKRITISCH → CRITICAL, accepted by a human on 2026-09-08owner is nullitsm_prod.SYS_INVENTAR.VERANTW, 7 percent null across the tableWritten in entities and fields, not tables and columns. The binding turns it into SQL, so a rename in the source breaks the binding rather than the question.
Partly. 214 critical functions are bound, from the spreadsheet rather than from a system. Of those, 131 resolve to named assets and 3 of the 131 depend on a provider with no second source.
critical_function_register, bound to Kritische_Geschaeftsprozesse_2026.xlsxSYS_INVENTAR.BEZEICHNUNG by name, 61 percent resolveict_third_party_register, one provider per asset, count of distinct providersThe suggested question is the interesting one: it looks answerable, and it is not.
An assistant over a warehouse can always produce something. The value of a bound ontology is that it knows the difference between a field that is empty and a field that does not exist, so it can decline the second case instead of quietly answering a nearby question. Every refusal above names the missing binding, which turns a wrong answer into a work item.