Merge pull request #4975 from laujan/kate-4954-4949-overview-overview

Kate 4954 4949 4972 overview overview

Author: AnnaMHuff

articles/ai-services/content-understanding/concepts/standard-pro-modes.md

@@ -75,18 +75,28 @@ Not sure which mode is right for your scenario? The following charts compare sta
 | Scenario | Standard mode | Pro mode|
 |----|----|----|
 | **Invoice analysis** | Extract insights on invoice data at scale and enable RAG search and further data analysis and visualization. Answer questions like: <br> &bullet; Extract purchase order number, total, due date, and line items for entry into database. | Analyze invoices and contractual agreements with clients and apply multi-step reasoning to draw conclusions on that data. Answer questions like: <br> &bullet; Does this invoice fulfill the contractual agreement we have in place with this client? <br> &bullet; Does this invoice need further review? |
-| **Call center transcript analytics** | Extract insights on large volumes of call center data to gain valuable insights on sentiment, understand customer issues, and create targeted training to address major pain points. Answer questions like: <br> &bullet; What are the main issues customers are calling about? <br> &bullet; What is the average length of calls made about x issue? | Analyze call center transcript data and apply multi-step reasoning to understand how call center employees are addressing customer needs, and if they're following guidelines. Answer questions like: <br> &bullet; Did the call center employee introduce themselves? <br> &bullet; Did this answer "pass" certain criteria? |
+| **Call center transcript analytics** | Extract insights on large volumes of call center data to gain valuable insights on sentiment, understand customer issues, and create targeted training to address major pain points. Answer questions like: <br> &bullet; What are the main issues customers are calling about? <br> &bullet; What is the average length of calls made about x issue? | Analyze call center transcript data and apply multi-step reasoning to understand how call center employees are addressing customer needs, and if they're following guidelines. Answer questions like: <br> &bullet; Did the call center employee introduce themselves? <br> &bullet; Did this answer *pass* certain criteria? |
 | **Mortgage application processing** | Extract the key values from mortgage application data and make it searchable and more easily accessible. Answer questions like: <br> &bullet;  What year was the mortgage application submitted? <br> &bullet; What are the names on the application? | Analyze supplementary supporting documentation and mortgage applications to determine whether a prospective home buyer provides all the necessary documentation to secure a mortgage. Answer questions like: <br> &bullet;  Do the names and social security numbers on the mortgage application match the supporting documentation? |
 
 ## Try pro mode
 
 You can try out the features of both Content Understanding standard and pro modes using the [Azure AI Foundry](https://ai.azure.com/explore/aiservices/vision/contentunderstanding). The service enables you to bring your own data and experiment with all the functionalities of both modes in a lightweight, no-code approach to help you find the best fit for your unique scenario.
 
-### Pro mode known limitations
+### Pro mode known limitations and best practices
 
 * Content Understanding pro mode currently doesn't offer confidence scores or grounding. It currently supports generative and classification of your fields but doesn't support extraction only.
+
 * Content Understanding pro mode is currently only available for documents.
 
+* The system operates in *lookup mode* when referencing documents. As a result, comprehensive information retrieval shouldn't be expected. If exhaustive recovery of data is required, we recommend that you incorporate the document into the input set.
+
+* Schemas should be designed with the highest level of specificity possible. For instance, instead of presenting a generalized list of inconsistencies, it's advisable to create distinct fields for each type of inconsistency, accompanied by detailed descriptions. Additionally, wherever feasible, references to specific sections of relevant documents that should be reviewed should be included.
+
+* Reference documents should be concise and focused. Prioritize essential documents and ensure they're as brief as possible to enhance retention and recall.
+
+
+
+
 ## Next steps
 
 For more information on document processing, see [Document processing overview](../document/overview.md).

articles/ai-services/content-understanding/media/overview/component-overview-updated.png

@@ -24,7 +24,7 @@ Azure AI Content Understanding is a new Generative AI based [**Azure AI Service*
 
 Content Understanding offers a streamlined process to reason over large amounts of unstructured data, accelerating time-to-value by generating an output that can be integrated into automation and analytical workflows.
 
-:::image type="content" source="media/overview/component-overview.png" alt-text="Screenshot of Content Understanding overview, process, and workflow.":::
+:::image type="content" source="media/overview/component-overview-updated.png" alt-text="Screenshot of Content Understanding overview, process, and workflow." lightbox="media/overview/component-overview-updated.png" :::
 
 ## Why process with Content Understanding?
 
@@ -56,12 +56,14 @@ Common applications for Content Understanding include:
 |Media asset management| Software and media vendors can use Content Understanding to extract richer, targeted information from videos for media asset management solutions.| [**Media asset management quickstart**](concepts/analyzer-templates.md#modality-templates) |
 |Tax automation| Tax preparation companies can use Content Understanding to generate a unified view of information from various documents and create comprehensive tax returns.| [**Tax automation quickstart**](concepts/analyzer-templates.md#modality-templates) | 
 |Chart understanding| Businesses can enhance chart understanding by automating the analysis and interpretation of various types of charts and diagrams using Content Understanding.| [**Chart understanding quickstart**](concepts/analyzer-templates.md#modality-templates) |
+|Mortgage application processing|Analyze supplementary supporting documentation and mortgage applications to determine whether a prospective home buyer provided all the necessary documentation to secure a mortgage.| [**Content Understanding Pro quickstart**](concepts/standard-pro-modes.md#apply-standard-or-pro-mode-to-your-scenarios)|
+|Invoice contract verification|Review invoices and contractual agreements with clients carefully. Apply a multi-step reasoning process to analyze the data. Ensure that conclusions, such as validating the consistency between the invoice and the contract, are accurate and thorough.| [**Content Understanding Pro quickstart**](concepts/standard-pro-modes.md#apply-standard-or-pro-mode-to-your-scenarios)|
 
 See [Quickstart](quickstart/use-ai-foundry.md) for more examples.
 
 ## Components
 
-:::image type="content" source="media/overview/cu-components.png" alt-text="Screenshot of Content Understanding components.":::
+:::image type="content" source="media/overview/pro-components.png" lightbox="media/overview/pro-components.png"alt-text="Screenshot of Content Understanding pro components.":::
 
 |Component|Description|
 |:---------|:----------|
@@ -71,6 +73,8 @@ See [Quickstart](quickstart/use-ai-foundry.md) for more examples.
 |Field extraction|Field extraction allows users to define the structure and schema of the desired fields to extract from input files. See [service limits](service-limits.md) for a complete list of field types supported. Fields can be generated via one of the following methods:</br></br> &bullet; **Extract**: Directly extract values as they appear in the input content, such as dates from receipts or item details from invoices.</br></br>&bullet; **Classify**: Classify content from a predefined set of categories, such as call sentiment or chart type.</br></br>&bullet; **Generate**: Generate values freely from input data, such as summarizing an audio conversation or creating scene descriptions from videos.|
 |Grounding source| Content Understanding identifies the specific regions in the content where the value was generated from. Source grounding allows users in automation scenarios to quickly verify the correctness of the field values, leading to higher confidence in the extracted data. |
 |Confidence score | Content Understanding provides confidence scores from 0 to 1 to estimate the reliability of the results. High scores indicate accurate data extraction, enabling straight-through processing in automation workflows.|
+|Reference dataset (offered in Pro mode)|The service can reference documents at inference time to aid in providing context. For example, if you're looking to analyze invoices to ensure they're consistent with a contractual agreement, you can supply the invoice and other relevant documents (for example, purchase order) as inputs, and supply the contract files as reference data.|
+|Multi-step reasoning (offered in Pro mode)|Multi-step reasoning takes data analysis a step further than extracting and aggregating structured data and allows you to draw conclusions on that data, minimizing the need for human review.|
 
 ## Responsible AI