What You Need to Know About Azure Cognitive Services

By - July 21, 2020

Microsoft’s Azure Cognitive Services (ACS) main goal is to bring artificial intelligence within reach of every developer, no matter their background or expertise in machine learning or AI. ACS does this by providing API capabilities to see, hear, speak, search, understand, and accelerate decision making into apps that you are building. One of Microsoft’s main goals as an organization is to democratize advanced technology so its capabilities can reach a wider audience that may not have the deep technical skills normally required to build with raw code. ACS is another example of this democratization at work.

How does ACS work?

To gain a better understanding of how ACS works, we’ll do so through the lens of a common deployment method, Microsoft’s Power Platform. The Power Platform provides organizations with little to no app development experience the ability to build applications that add value to their business. The Power Platform provides multiple ways to access, store, and interact with data in a mostly non-technical manner and provides no-code/low-code methods for building apps, reports, and workflows. ACS is delivered as a Platform as a Service, meaning all hosting and management is handled by Microsoft in the cloud and needs to be integrated with a deployment solution like the Power Platform to be used.

Utilizing these methods, one of your team members with little app-building, AI, and machine learning experience could deploy ACS to tap into the power of Artificial Intelligence. Let’s dive into the main functions of ACS:

Decision Making

  • ACS provides the ability to build apps that surface recommendations for informed and efficient decision-making. The three common types are:
    • Anomaly Detector: find irregularities in your data
    • Content Moderator: screen text, image, and video for inappropriate content
  • Personalizer: choose the content best suited for users in real-time. Personalizer has won awards for Microsoft in reinforcement learning.

Vision

  • Review pictures and video to recognize, identify, caption, index, and moderate using techniques such as:
    • Form Recognition
    • Ink Recognition (reviewing handwritten notes)
  • Computer Vision
    • Custom Vision
    • Face Recognition
    • Video Indexer

Language

  • Give your apps the power to process natural language with pre-built scripts, evaluate sentiment, and learn to recognize what users want:
    • Text Analytics
    • Translator
  • QnA Maker (examples include social media apps, chat bots, and speech enabled desktop applications)
    • Language Understanding
    • Immersive Reader

Speech

  • Turn speech into text and text into natural-sounding speech. Translate from one language to another and enable speaker verification and recognition with Speech Service.

Search

  • Enable Bing Search APIs to capture the ability to filter through billions of webpages, images, videos, and news with a single API call.

Azure Cognitive Services machine learning capabilities leverage the thousands of model iterations that have been built by Microsoft over the past decade. The fact that these models have been tried and tested thousands of times provides generically accurate models that can be deployed in a short period of time. By taking the learnings from the thousands of model iterations, Microsoft taps into model “transfer learning” that allows models to be implemented to a variety of scenarios by personnel who may not have a strong machine learning background.

How to implement ACS

Let’s take a look at a few examples of how ACS can be implemented at your organization:

  • A call center needs to equip customer service agents with better information to help customers in a more accurate and efficient manner. Using ACS speech recognition and sentiment analysis, the call agent could be given the top three support documents for the caller’s issue before getting on the phone. By addressing the issue correctly and in a timely manner upfront, the company is able to save money and deliver better customer service.
  • An organization spends thousands of hours each year with basic data entry tasks. ACS’s Form Recognizer can be used to scan, review, and document the data from forms to eliminate human error-prone activity. Form Recognizer has the ability to recognize and combine fields and continues to become more accurate the more forms you feed it.
  • A company realizes that the majority of customer support questions are similar and the website FAQ page is ineffective. ACS’s QnA Maker can provide chatbot capabilities to process customer questions and deliver the appropriate responses automatically without the supervision of a human. This cuts back on the need for a human customer service agent to step in to solve the frequently asked questions.
  • The finance department of a construction organization feels it has a strong understanding of its own safety data, but would like to see side by side comparisons with industry data to compare accident rates to make sure its insurance costs are appropriate. ACS’s Bing Search APIs would allow the company to search and scrape the Occupational Safety and Health Administration (OSHA) database to enable a side by side comparison.

If you are looking to tap into the world of artificial intelligence, Azure Cognitive Services may be the answer for you. Building intelligent and supported algorithms into apps, websites, and bots to see, hear, speak, understand, and interpret your customer needs is well within reach. RSM is a Microsoft Gold Partner and is familiar with how ACS can benefit an organization in a variety of ways. To learn more about how RSM can support your data analytics needs, please visit our website, call 800-274-3978, or contact us.

Receive Posts by Email

Subscribe and receive notifications of new posts by email.