Over the past few years, we have seen many examples of organizations applying conversational AI in meaningful ways. Accenture and Caesars Entertainment are making their employees more productive with enterprise bots. UPS and Asiana Airlines are using bots to deliver better customer service. And finally, BMW and LaLiga have built their own branded voice assistants, taking control of how customers experience their brand. These are just a few of the organizations that have built conversational AI solutions with Azure AI.
This week at Microsoft Ignite, we announced updates to our products to make it easier for organizations to build robust conversational solutions, and to deploy them wherever their customers are. We are sharing some of the highlights below.
Most popular open source SDK for accelerated bot development
We announced the release of Bot Framework SDK 4.6 making it easier for developers to build enterprise-grade conversational AI experiences. Bot Framework includes a set of open source SDKs and tools for bot development, and can easily integrate with Azure Cognitive Services, enabling developers to build bots that can speak to, listen to, and understand users.
- Bot Framework SDK for Microsoft Teams. Developers can build Teams bots with built-in support for Teams messaging extensions, proactive messaging and notifications, and more.
- Bot Framework SDK Skills (preview). Developers can create a reusable conversational skill and also leverage pre-built skills which come with the language models, dialogs, QnA, and integration code. Pre-built skills include calendar, email, task, point of interest, weather, news, and more.
- Adaptive Dialog (preview). enabling developers to build conversations that can dynamically handle interruptions and context switching.
- Language Generation (preview). enabling developers to define multiple variations on a phrase, execute simple expressions based on context.
- Python and Java updates (preview). enabling developers to build in the language of their choice.
“We compared several options, and in terms of flexibility, time-to-market, and interoperating with our existing infrastructure, we felt that Microsoft Bot Framework was by far the best way to go.” – Cole Dutcher, Associate Director of Engineering, Jet.com/Walmart Labs
Simplifying bot building with a low-code visual experience
In order to help developers get started quickly with bot development, we offered bot code samples and templates. To make it even easier to get started, we have launched Bot Framework Composer (preview), an integrated development tool built on top of the Bot Framework SDK. Rather than having to start coding completely from scratch, developers can start with Composer, a low-code visual experience enabling developers to create, edit, test, and refine conversational apps (bots), with the flexibility to extend the bot with custom code. Developers can also have one centralized place to incorporate Azure Cognitive Services, starting initially with language understanding, with more Cognitive Services to come.
Like Bot Framework, the Composer is an open source project. Get started and build a bot today.
Connecting your bot to your users
Azure Bot Service allows you to take the bot you’ve built using Bot Framework, and host it easily in Azure so you can connect your bot to your users across popular channels like Facebook, Slack, Teams, or even your own websites—wherever your customers are. We’ve announced some new integrations:
- Direct Line Speech. Voice-first conversational experiences continue to grow in popularity and importance, and today announcing the general availability of the Direct Line Speech. Direct Line Speech is a new channel that simplifies the creation of end-to-end conversational solutions for streamed speech and text bi-directionally from the client to the bot application using WebSockets on Azure Bot Service. Get started with this the step-by-step tutorial.
- New integrations and adapters. As Bot Framework adoption increases, so does the demand for integrations to new channels. We are pleased to announce a new integration with LivePerson, the provider of conversational platform LiveEngage, so that developers can build customer care scenarios that can escalate conversations to human agents. We also added a new WeChat adapter. If you are interested in connectors to other platforms, learn more about our additional channels and adapters. You’ll find we have a growing list of Bot Framework adapters, including community adapters for platforms like WebEx Teams, Google Hangouts, Google Assistant, Twitter, and Amazon Alexa.
- Direct Line App Service Extension (preview). Bots using Web Chat and Direct Line can now be isolated from other traffic on the Bot Service by running Direct Line on a dedicated Azure App Service. This also allows bots to participate in Azure VNET configurations. A VNET allows developers to create their own private space in Azure and is crucial to their cloud network as it offers isolation, segmentation, and other key benefits. Learn more about this capability.
“We used Microsoft Azure Bot Service and Cognitive Services to help cope with the complexity of launching Aura in six countries on four separate channels—and do it all seamlessly.” – Chema Alonso, Chief Data Officer, Telefonica
Integration with Azure Cognitive Services (Language Understanding, QnA Maker)
One of the key benefits of using Bot Framework and Azure Bot Service is the ability to also integrate powerful domain-specific AI models using Azure Cognitive Services. We made several new announcements for Azure Cognitive Services, such as new Speech capability like Custom Neural Voice, which allows users to create personalized voices.
Language Understanding, an Azure Cognitive Service, introduced new capabilities that enable developers to handle even more sophisticated language structures. These capabilities can better parse complex sentence structures into a hierarchical structure and better understand natural language. It also offers a new UI experience that enables you to more easily benefit from these new models by labeling sub-components and overlapping entities within the user interface. Finally, Language Understanding service now supports Hindi and Arabic, expanding its coverage of languages.
QnA Maker is a cloud-based API service that creates a conversational, question-and-answer layer over your data. With QnA Maker, you can build, train, and publish a simple question and answer bot based on FAQ URLs, structured documents, product manuals or editorial content in minutes. QnA Maker’s multi-turn capability is now generally available, so you can create multi-turn QnA bots without having to write any code. In addition, QnA Maker is also going big on global requirements with increased region availability and ranking feature support for 10 languages, and chit-chat support allows you to handle small talk for 8 languages. Finally, Qna Maker also supports batch testing to quickly test knowledge base with a standard set of test cases and validate the quality of the answers.
“By using Microsoft Azure Bot Service and Cognitive Services, we’ve been able to continue our own Progressive journey of digital innovation and do it in an agile, fast, and cost-effective way.” – Matt White, Marketing Manager, Personal Lines Acquisition Experience, Progressive Insurance
Power Virtual Agents
At Microsoft Ignite, we also announced Power Virtual Agents, a UI-based bot building experience built on the Bot Framework and made available on the Power Platform. It is designed for business users who are looking for a code-free bot building experience. Bots built using Power Virtual Agents can be extended by developers using the Bot Framework SDK and tools. In fact, this allows collaboration between business users who have subject matter expertise, and developers who have the technical expertise to build custom conversational experiences.
With these enhancements, we are delivering value across the entire Microsoft Bot Framework SDKs and tools, Language Understanding, and QnA Maker in order to help developers become more productive in building a variety of conversational experiences.
We look forward to seeing what conversational experiences you will build for your customers.
Azure. Invent with purpose.