Document Automation settings
- 최종 업데이트2024/11/11
Document Automation settings
The Document Automation settings enable users to use 생성형 AI provider and extraction provider services in learning instances.
주: 생성형 AI 모델은 오류가 발생하거나 생성하는 정보를 잘못 표시할 수 있습니다. AI 모델이 생성한 콘텐츠는 정확성, 신뢰성, 완성도를 검증하는 것이 좋습니다.
주: A valid 클라우드 license is required to configure
these settings.
Users with administration permissions can enable or disable the 생성형 AI provider and extraction provider services for users to use in their learning instances. See 역할의 기능 권한.
생성형 AI provider
To use the 생성형 AI provider option when you create a learning
instance, select a provider and a region to send and receive data. Selecting a
region depends on following factors:
- Data residency and compliance: Different countries, regions, and organizations have legal or regulatory requirements that govern how data is stored and processed. Selecting an appropriate region ensures that you comply with such requirements.
- Latency and performance: Selecting a region closer to you or the users using these services might reduce latency, leading to faster response times and improved performance.
주:
- Support for additional new regions (Australia, Canada, Japan, and India) were added in v.34 release. To use these new regions, you must use 문서 추출 package version 3.34.xx or later.
- Learning instances that are using the 생성형 AI and external
provider options will fail when processing documents and the
illegalArgumentException error will be seen on
the Automation Co-Pilot requests page in the following scenario:
- If you had disabled the external connections option in v.33 or earlier release
- If you are using an older version of the 문서 추출 package (3.33.xx or earlier) after updating to v.34 or later release
You must enable the 생성형 AI and external provider options for your learning instance to fix this error.
- When you update your Control Room to v.34 or later release, you will see the
following behavior:
- If you previously enabled external connections manually:
- 생성형 AI and external provider options will be enabled.
- The region will be set to the region you had selected in your previous release.
- If you previously disabled external connections: 생성형 AI and external provider options will be disabled.
- If you did not configure the external connections and used the default settings: 생성형 AI and external provider options will be enabled.
- If you previously enabled external connections manually:
- When the 생성형 AI provider is disabled after creating learning instances, you will see an error when you process documents. You need to enable the 생성형 AI provider or configure BYOL for the 생성형 AI provider in the Extract data action for the learning instance to process documents. See 데이터 추출 작업.
To enable 생성형 AI providers, see 생성형 AI 및 기타 Document Automation에 대한 외부 연결 활성화.
Extraction provider
To use Google Document AI extraction provider or Google Vision OCR, select the extraction provider option and select an appropriate region.
주:
- You will not be able to create a learning instance using the Google Vision OCR when the extraction provider option is disabled. Users will see an error when they try to create a learning instance using Google Vision OCR.
- If you are using Google Vision OCR in a learning instance and if this option is disabled after creating the learning instance, you will see an error when you process documents. You need to enable this option to process documents.
- If you are using Google Document AI extraction provider in a learning instance and if this option is disabled after creating the learning instance, you will see an error when you process documents. You need to enable this option or configure BYOL for the 생성형 AI provider in the Extract data action for the learning instance to process documents. See 데이터 추출 작업.
Regions support matrix
The following table provides the regions supported in the Control Room for 생성형 AI and extraction providers:
| 생성형 AI and extraction providers | Supported regions |
|---|---|
| Microsoft OpenAI |
|
| Anthropic Claude |
|
| Google Document AI and Google Vision OCR |
|