Starting today, Azure Synapse brings the capabilities of Azure AI directly to the Synapse Studio. Customers can now apply AI to analyze structured, semi-structured, and unstructured data like forms, audio, video and text documents without having to write a single line of code or be a data scientist. 

“Some of the most useful insight often comes in the form of audio, text, or even images. Without learning ML, or even writing any code, Azure Synapse customers can use AI models built by Microsoft to ‘schematize’ these unstructured or semi-structured document types and analyze them in a structured format," says John Macintyre, Director of Product for Azure Synapse Analytics.

Being able to use AI to infer structure from data sources previously unavailable for analytics enables customers to expand the analytics landscape, apply common data lineage and management concepts to their unstructured data, and easily derive insights from it. Inferring latent structure from audio, video, and text data structure enables Azure Synapse customers to normalize their data landscape and discover powerful insights that were previously hidden or difficult to analyze.

These AI-infused pipelines automate the task of starting with unstructured data in different formats and languages, normalizing, annotating, and extracting insights.

Getting started

To get started, customers can browse the gallery of pipeline templates in the Knowledge Center in Azure Synapse and select Customer Feedback Analytics. This template enables users to analyze customer feedback by simply adding an AI transformation directly into the data pipeline. It can easily be edited to fit a different scenario like analyzing medical forms, documents, or business forms.

Figure 1: Knowledge Center in Azure Synapse

A common pattern in most organizations is to mine customer feedback for product improvements, churn analysis, and generate business metrics. For example, a hotel chain may want to understand how routine customer experiences eventually impacts sales. However, most customer engagement data is locked up inside phone calls, texts, online reviews and emails.

Using the integration of Azure Synapse and Azure Cognitive Services, B2C calls can be transcribed, and product or service mentions and their associated sentiment is transformed into structured tables directly in Azure Synapse. Joining insight from reviews or phone calls with historical revenue data enables customers to understand how specific experiences were good or bad for business. This is all possible today within Azure Synapse, no code or machine learning required.

The screenshot below shows the entire data pipeline that makes this possible—including the AI transformation in the red box where opinion mining infers sentiment from reviews.

Figure 2: “Customer Feedback Analytics” Pipeline Template in Azure Synapse

This next screenshot is a double click on the AI transformation infused into the data pipeline. These activities invoke the Azure Cognitive Services APIs to identify key phrases, entities and perform sentiment analysis on the reviews.

Figure 3: “Opinion Mining” copy data activity in the data pipeline to invoke API for sentiment analysis

Configuring each of the Cognitive Services endpoints to enrich each review within the dataset.

 Figure 4: Activity in the data pipeline calling the API in Azure Cognitive Services

The final step in the pipeline is to generated a structured representation of the data the enables BI and ad-hoc analytics on the data.

Figure 5: Databases created by the “Customer Feedback Analytics” already infused with sentiment analysis results

Conclusion

To try the customer feedback analytics sample, visit the Knowledge Center in Azure Synapse, select Pipelines, and then launch Customer Feedback Analytics. The sample demonstrates how you can build data pipeline infused with code-free AI to answer questions like, “What were the most common descriptors or key phrases associated with negative reviews?”

This new feature allows you to leverage pre-built and custom AI models to infer structure from your unstructured data and leverage the same tools to analyze structured and unstructured data. You can now apply AI directly to your data integration workloads in Azure Synapse to harness the latent insights from unstructured data in your organization’s data analytics processes.

Get Started Today

Customers with *qualifying subscription types can now try the serverless SQL and Apache Spark pool resources in Azure Synapse using free quantities until July 31^st, 2021.

• If you don’t have one already, follow this tutorial to create a Synapse workspace
• Use this step-by-step tutorial for a quick tour of Azure Synapse

*Free quantities apply only to the following subscription types: Pay-As-You-Go, Microsoft Azure Enterprise, Microsoft Azure Plan, Azure in CSP, Enterprise Dev/Test. These included free quantities aggregate at the enrollment level for enterprise agreements and at the subscription level for pay-as-you-go subscriptions.