Google was one of the first cloud providers to offer services in the Artificial Intelligence and machine learning space, with a multitude of services in this domain catering to different types of users from developers to data scientists. Also, most of these services/APIs have originated from Google's internal usage and existing product offerings, which have now been exposed as APIs/services under the Google cloud umbrella. These services cover a variety of AI/ML use cases, like the following:

• Image and Video Analysis: One of the common requirements for most of the organizations starting with their AI/ML journey is to classify and understand the context and metadata around their images and video streams. In the image analysis space, Google offers a Cloud Vision API (https://cloud.google.com/vision/), which can help classify images into thousands of categories, detect individual objects and faces within images, and find and read printed words contained within images. You can also detect any explicit content, identify logos, landmarks, or even search the web for similar images. Similarly, for video analysis, Google offers Cloud Video Intelligence (https://cloud.google.com/video-intelligence/), which can search every moment of every video file in your catalog, quickly annotates videos stored in Google Cloud Storage, and helps identify key entities within your video and when they occur within the video. Using this video analytics, you can generate appropriate content recommendations for end users and even show more contextual advertisements in lines with the content itself.

• Speech and text-related AI services: There are multiple AI-related aspects when it comes to text and speech analysis as well as conversions between one form to another. For the same reason, Google offers the Cloud Speech API (https://cloud.google.com/speech/), which can convert audio to text and can recognize over 110 languages and variants and thereby help transcribe audio content. Likewise, to convert written text-to-speech, Google offers the Cloud Text-To-Speech API (https://cloud.google.com/text-to-speech/) to synthesize natural-sounding speech with 30 voices, which is available in multiple languages and variants. Apart from this, Google also offers a service to detect a particular language in the text and also to convert text from one language to another using the Cloud Translation API (https://cloud.google.com/translate/). Another service which Google offers in this space is for the deep analysis of text by extracting information about people, places, events, and much more, which is mentioned in text documents, news articles, or blog posts. This service is a Cloud Natural Language (https://cloud.google.com/natural-language/), which can help understand the sentiment and intent (like positive or negative reviews) from social media content or customer conversations in call center type of text message exchanges.
• Chatbots: One of the most common uses of AI these days is to create conversational interfaces (or chatbots) for websites, mobile applications, messaging platforms, and IoT devices, that are capable of natural and rich interactions between users and your business. These chatbots can understand the intent and context of conversation to provide highly efficient and accurate responses. For this, Google has a service called DialogFlow Enterprise Edition (https://cloud.google.com/dialogflow-enterprise/), which has multiple out-of-the box templates and supports 20+ languages and integration with 14 different platforms to provide rich experience to the end users.

• Custom Machine Learning: In most cases, advanced users (such as data scientists) like to have more control of their algorithms, ML models, and the way the system generates results. So, for such scenarios, services like the ones discussed previously don't provide that level of deeper control and configurability, and that's why Google offers a set of services like Cloud AutoML (https://cloud.google.com/automl/) and Cloud Machine Learning Engine (https://cloud.google.com/ml-engine/), which provide greater options. As an example, if a retailer would like to classify images of various dresses as per their colors, shape, and design, then they can use Cloud AutoML to feed some sample data to train custom ML models, which can then be used to generate image recognition for actual images. Likewise, if data scientists would like to create their own custom models to perform predictive analytics, then they can use a framework like TensorFlow along with a managed service like the Cloud Machine Learning Engine, which is already integrated with many other Google services and also provides a familiar interface like Jupyter notebooks to create custom models. These services, clubbed with Cloud TPUs (TensorFlow Processing Unit, https://cloud.google.com/tpu/), provide up to 180 teraflops of performance, providing the computational power to train and run cutting-edge machine learning models at scale.