At Microsoft Build on May 10, 2017, the company said its Cognitive Services catalog had grown to 29 offerings. The pitch was practical: developers could add ready-made AI capabilities with a few lines of code, or customize services without designing deep-learning models themselves. [story-s1]
The new offerings included Bing Custom Search, Custom Vision, Custom Decision Service and Video Indexer. Microsoft also opened Cognitive Services Labs, beginning with a gesture API intended to let people control and interact with devices through gestures. These were announcements and demonstrations, not evidence that every tool was generally available or ready for production. [story-s1]
Build also brought two different paths for teams seeking more control. Azure Batch AI Training entered private preview for developers who wanted to train neural-network models with frameworks such as TensorFlow, Caffe or Microsoft Cognitive Toolkit. Separately, Microsoft described Presentation Translator as a feature announced that day: it would add live transcription and translation to PowerPoint presentations when launched. [story-s1]
The distinction mattered. A broader menu of hosted services lowered the barrier to experimentation; private-preview training still required access, and a promised presentation feature was not yet a shipped product. For teams evaluating the announcements, the useful question was not simply how many AI tools Microsoft offered, but which capabilities they could actually test and deploy. [story-s1]
Source documents
- Microsoft Build 2017: Microsoft AI – Amplify human ingenuityhttps://blogs.microsoft.com/blog/2017/05/10/microsoft-build-2017-microsoft-ai-amplify-human-ingenuity/