Iām thrilled to announce a follow up on our goal to provide visibility into the consumed AI Units, specifically to Document Understanding
Pre-requisites
Required activity versions:
- DocumentUnderstanding.Activities >= 2.9.1
- IntelligentOCR >= 6.19.1
- DU.ML >= 1.28.2
- Activities.MLServices >= 2.0.2.0
- Robot version >= 23.6
One can find the (Preview) Document Understanding AI Units consumption in Insights - in preview, awaiting your feedback before going GA, so do reach out!
The dashboard displays all data, be it either reported from APIs, cross-platform & IOCR activities , for modern & classic projects (reported part of an AI center project), pretrained specialized & generative models, as well as custom models - for you to clearly understand how AI Units are being consumed.
However, it may be the case to see entries as
- N/A - where they do not apply (for example, when reporting the infra cost of AI center proects, an Orchestrator Folder does not apply, as it is not assigned to)
- or Unknown when
ā old model, robot or activities version are used - in this case, make sure to update to the latest versions
ā the data is in the tenant for the region it resides in (e.g. the automations run in the US, so they can appear in the US tenant, assuming it is configured for the US region) ā this is something we work on fixing
ā consumption happens from hybrid deployments
ā consumption happens using API Keys as authentication method, for example via the Machine Learning or Form Extractor from the IntelligentOCR package ā this is something we work on addressing
Do try them out, do let us know your thoughts!