Microsoft Corp. is stepping up its artificial intelligence capabilities within its Azure Cloud platform with the launch of a new service called Azure Percept that’s designed to enable a wealth of new AI capabilities for devices at the network edge.

Announced today at the virtual Microsoft Ignite event, Azure Percept is designed to enable futuristic scenarios that can be made possible now by combining AI with computing at the network edge. They include elevators that can respond to voice commands and video cameras that can notify store managers when a particular product needs restocking.

These possibilities can be enabled simply by adding AI capabilities such as voice recognition or image classification to standalone edge devices. The new Azure Percept service is set to become Microsoft’s way of doing exactly that.

Azure Percept, available in preview starting today, is actually an entire platform of hardware and services that helps to make it easier to enable AI capabilities at the edge. More specifically, it’s “an end to-end system, from the hardware to the AI capabilities, that ‘just works’ without requiring a lot of technical know-how,” said Roanne Sones, corporate vice president of Microsoft’s edge and platform group.

The platform (pictured) includes a development kit and an intelligent camera called Azure Percept Vision, as well as a “getting started” experience called Azure Percept Studio that guides users on how to develop, train and deploy AI within edge devices. Azure Percept Vision makes it possible to connect these newly AI-enabled devices to services within the Azure cloud.

There’s also something called Azure Percept Audio, a collection of embedded hardware-accelerated AI modules. They enable speech and vision AI at the edge during times when the device isn’t connected to the internet.

Azure Percept will connect to Azure AI Cognitive Services and Azure Machine Learning models as well as AI models available from the open-source community that have been designed to run on the edge. In addition, Microsoft is working with various outside silicon and device makers to create what it says will be an ecosystem of intelligent edge devices that are certified to run on Azure Percept.

“We’ve started with the two most common AI workloads, vision and voice, sight and sound, and we’ve given out that blueprint so that manufacturers can take the basics of what we’ve started,” Sones said. “But they can envision it in any kind of responsible form factor to cover a pattern of the world.”

Constellation Research Inc. analyst Holger Mueller told SiliconANGLE that with Azure Percept, Microsoft is giving enterprises both a platform and a toolset that will allow them to run AI at the edge. “Given it includes Azure Percept Audio, it is clear that Microsoft has more than just the industry edge case in mind, it is aiming to target consumer use cases too,” he said.

Azure Machine Learning

Microsoft’s renewed focus on AI extends to its other platforms too, notably the Azure Machine Learning service that’s used by developers to build and deploy AI models from the Azure Cloud. Azure Machine Learning gains several new capabilities, the most prominent being a new integration with the Azure Synapse Analytics service that’s being made available in preview starting today.

Azure Synapse Analytics brings together data integration, enterprise data warehousing and big data analytics in one service. It makes it possible to ingest, explore, prepare, manage and serve data for immediate business intelligence and machine learning needs. The integration with Azure Machine Learning makes it possible for developers and data science teams to use Azure Synapse Analytics Spark clusters to enable large-scale, interactive data preparation tasks from within Azure Machine Learning notebooks, Microsoft said.

In addition, Azure Machine Learning is now supported within Azure Arc, which is a platform that makes it possible to run Azure services in any Kubernetes environment, whether it’s on-premises, multicloud or at the edge.  Arpan Shah, general manager of Microsoft Azure, said in a blog post that Azure Arc’s support for Azure Machine Learning essentially extends the company’s machine learning capabilities to any hybrid or multicloud environment.

By doing this, he said, “customers can run training models where the data lives while leveraging existing infrastructure investments. This reduces data movement and network latency, while meeting security and compliance requirements.”

Mueller told SiliconANGLE that these were important new capabilities for enterprises that want to re-use their private code assets in public cloud, on-premises and edge environments. “Microsoft enables this with Azure Arc, and given the relevance of AI and ML for enterprise success, it is now adding AI portability to Arc’s supported platforms,” Mueller said. “That is a key step forward to enable enterprises to bolster their next generation applications with AI.”

Data and analytics

Data and AI are inextricably linked, so the second major focus for Azure at Ignite is on enhancing “data agility” for users so they can quickly gain insights at any scale.

To that end, Microsoft said, it’s helping customers migrate to the Azure Synaptics Analytics service more easily with a new tool called Azure Synapse Pathway. Rohan Kumar, corporate vice president of Azure Data, said in a blog post that Azure Synapse Pathway makes it possible for users to scan their source systems and automatically translate existing scripts into Transact-SQL, a proprietary extension to the Structured Query Language that’s used to interact with relational databases, with just a few clicks.

What that means is that it’s now possible to migrate data from databases and data warehouses such as Teradata, Snowflake, SQL Server, Google BigQuery and Amazon Redshift into the Synapse service in more quickly. There it can be quickly analyzed for insights and used by other Azure services.

Another related update pertains to Azure Purview, which is Microsoft’s unified data governance service that was launched in December to give users a more holistic view of their data wherever it resides. Kumar said Azure Purview has already proved to be a big hit with customers, and has been used to automatically scan, discover and classify more than 14.5 billion data assets.

Kumar said it’s now possible to use Azure Purview to scan Azure Synapse workspaces across both serverless and dedicated SQL pools. “Starting today, customers can now automatically scan and classify data residing in Amazon AWS S3, as well as data residing in on-premises Oracle DB and SAP ERP instances,” Kumar said. “This is in addition to Teradata, SQL Server on-premises, Azure data services and Power BI, which have been supported data sources since Azure Purview’s debut.”

Mueller said enterprises have always been a little bit stranded when it comes to data governance, and that it was good to see Microsoft addressing this problem. “Azure Purview not only scans Azure Synapse Analtics databases, but also AWS S3, SAP ECC, S/4HANA and Oracle’s database, giving enterprises a compete view of what data is where and how it may need to be governed,” Mueller said.

Microsoft Azure users are getting some AI-enabled tools to search their data too. They’re coming via an update to Azure Cognitive Search, which offers built-in AI capabilities that enrich all kinds of data so users can easily identify and explore relevant content within individual applications. Azure Cognitive Search gains some new “semantic search” capabilities that will enable a more relevant contextual search experience within their apps.

Finally, Kumar announced a new service called Azure Migrate, which is meant to make it easy for companies to migrate their data to the Azure SQL database. “Customers can now easily understand all the technical and financial aspects of their cloud journey, by assessing their SQL source and destination landscapes at scale, with integrated SKU recommendations and cost estimates included,” he said.

New vertical clouds

Microsoft also announced the launch of three new vertical clouds at Ignite, for manufacturing, financial services and nonprofits.

Microsoft launched its first vertical cloud, for healthcare, in October last year before following up with a second edition for retail customers in January. The vertical clouds bundle an array of tools and services that Microsoft deems to be most useful for those particular industries.

The retail cloud, for example, included Microsoft Azure, Microsoft 365, Dynamics 365 and the Power Platform, as well as the Microsoft Advertising suite, plus various templates and connectors to help customers in that industry deal with structured and unstructured data.

All three of the new vertical clouds offer a similar package, together with various “industry-specific capabilities via Microsoft 365, Dynamics 365, Microsoft Power Platform and Azure.” They also support industry interoperability and compliance standards, Microsoft said.

“The vertical cloud trend that was started early in 2020 goes on strong in 2021 with Microsoft building on its launch with three additional new clouds,” Mueller said. “This is great news for enterprises because anything that addresses their vertical needs will allow them to adopt cloud processes faster and with that, adopt enterprise acceleration too. What is key is that any vertical cloud should have software assets as integral part of the offering, and that these assets evolve following a roadmap for the next years.”

The company said the new Financial Services and Retail clouds will start accepting customers for public preview at the end of this month, while the Nonprofit and Manufacturing clouds will open up by the end of June. Microsoft’s Healthcare cloud is already available in preview.

Image: Microsoft