Exploring Use Cases for Cognitive Services

Let’s walk through some examples of apps and companies transformed by cognitive services and some future use cases to see how much they’re changing the technology landscape.

Thanks to cloud giants like AWS, IBM, and Microsoft Azure's cognitive services, developer teams of all sizes now have access to cognitive services of staggering power. Delivered through APIs, these services make it easy to inject next-generation intelligence into applications.

Chat and Social Interaction

In 2015, monthly active users on chat apps surpassed those on social networks, and the chasm continues to widen. Indeed, messaging has become an essential feature of social networks themselves. And with this rapid growth, messaging apps have evolved from simple tools for sending and receiving short, text-based messages to innovative, full-featured experiences boasting surprising and delightful features. And driving that innovation are cognitive APIs.

Chatbots and Cognitive Computing

Chatbots are one of the earliest forms of AI algorithms. While unlikely to pass the Turing test soon, they represent the natural evolution of voice-enabled applications. Where once you would call a support line and press 1 for Accounts Payable, now you can speak in complete sentences to a system that can discern your intent.

Whether you’re aware or not, chatbot adoption has exploded as companies seek to reduce wait times, improve customer experience and minimize the cost of human telephone operators. Right now, they’re mainly used to handle simple tasks: understanding basic requests and responding based on predefined rules, answering questions like, “Where is my order?” or “Chatbot, turn on mood lights.”

However, APIs like Watson Assistant or Amazon Lex make it easy to build services that can apply logic to observed patterns in those natural-language requests. These services may, for instance, observe a sudden rush of calls from an airport suffering take-off delays and change the sequence of options to prioritize rescheduling flights. Or they may see that calls from a particular country or region tend to be conducted in a different language and change the default accordingly. They may even identify grammatical patterns that indicate customers to immediately forward to a supervisor.

Intelligent conversational interfaces using speech recognition, text-to-speech, facial recognition, and machine learning models can provide highly engaging experiences and life-like conversations for various purposes. And even better, they’ll learn from those experiences.

Chatbots will change how we bank, shop, and learn: making recommendations, understanding abstract concepts, and getting to know individuals based on prior engagements. Eventually, they’ll get so good you won’t even know if you’re talking to a human.

Code Example: Home Automation Chatbot

Using Watson and PubNub ChatEngine, you can easily spin up a chatbot with artificial intelligence that controls your smart home.

This tutorial shows you how to build a chatbot that accepts text commands, parses them, and takes action based on them. For example, a user types “turn on the lights in the living room,” and the bot will trigger the lights.

{

  "homeauto_intents": 
  [
    {
      "intent":"turnOFF",
      "examples":
      [
        {"text":"Put off"},
        {"text":"Switch off"},
        {"text":"Turn off"}
      ],
      "description":"Turn on intents"
    },
    {
      "intent":"turnON",
      "examples":
      [
        {"text":"Put on"},
        {"text":"Switch on"},
        {"text":"Turn on"}
      ],
      "description":"Turn off intents"
    }

Natural Language Processing

Another hugely impactful area is data science and natural language processing (NLP), the umbrella term for AI solutions that can fruitfully process large amounts of natural language data. NLP can not just gauge words and grammar from a semantic perspective but can divine sentiment and emotion, unearthing how users feel about a topic or subject through message-by-message analysis.

NLP is a huge benefit for brands, public figures, and organizations that need to understand and respond to user opinions at a time when reputations can be made or unmade in a matter of minutes. Imagine a brand launches a new commercial for a product. Using the right cognitive services, it can tap into a social media stream on a specific hashtag or the product name and have its NLP API analyze all relevant messages and provide feedback on how the public responds to the product.

Below is an example of an app designed to analyze and gauge how people felt about US politicians on Twitter. It monitors specific keywords and phrases and can then plot the emotion of users in defined geographical regions.

For example, if a user submits the text “I am happy”…

{
   "session_id": 1,
   "text": "I am happy!"
}

Watson analyses the text and returns the following:

{
   "session_id": 1,
   "text": "I am happy!"
   "session_sentiment": {
      "overall": 0.879998,
      "positive": {
         "count": 1,
         "avg": 0.879998
      },
      "negative": {
         "count": 0,
         "avg": 0
      },
      "neutral": {
         "count": 2,
         "avg": null
      }
   },
   "score": 0.88006828
}

Brands already spend large amounts on market sentiment analysis. As these systems grow more intelligent, robust, and automated, they’ll be able to understand the public far better at a lower cost.

eCommerce

Though online shopping has completely changed how we buy goods, e-tail lacks one key component of a brick-and-mortar store: helpful employees. At the scale online stores operate, it isn’t economically viable to have actual people staff live chat.

As a result, many online stores are turning to intelligent shopping assistant bots to optimize the experience, assist shoppers with their questions, make recommendations and even check out.

Nordstrom has dominated previous holiday seasons with their Messenger chatbot, which went beyond simple predefined questions and answers and used cognitive services to truly understand what the customer was looking for and assist as needed. It offered gift recommendations and could even help fulfill the order.

Chatbots also save us from the dreaded customer support phone call, waiting an hour for a representative to deal with a simple problem. Amazon has deployed chatbots that can resolve minor issues that most customers have when they need help with their orders.

Now that we’ve looked at a couple of examples of intelligence in the real world today let’s peer into the future and see how cognitive services will change our world in the future.

Smart Cities

Cities of the future will rely on various integrated intelligent services to make them safer, more efficient, and more environmentally conscious. Image recognition, computer vision, and vision APIs will play a critical role in this transformation, processing and taking action on images within the urban space.

Agriculture

Global populations continue to grow, and feeding those billions of people will be a considerable challenge in the years to come. Cognitive services will play a critical role in managing fields and factories, allowing us to make intelligent decisions and control resources with a precision we’ve never had before.

Smart farms and IoT will incorporate as many valuable data points as possible to create intelligent agricultural decisions, even ones that seem counterintuitive. For example, by aggregating real-time weather data, remote sensor data, and historical performance, cognitive services can perfect the individual irrigation plan and update it for every day’s unique circumstances.

Data Security

As we grow more connected and our digital lives overshadow our physical ones, data privacy and security transform from something we’re vaguely aware of to a disconcerting, ever-present personal threat.

Regulations and rules—HIPAA, GDPR, SOC II—are one way to ensure that businesses and organizations have the proper guardrails in place. Implementing these complex regulations in detail can be a lot to handle, which is where machine learning comes into play.

Cognitive services can be trained to understand and make sense of rules and regulations, then suggest ways to achieve compliance. Cognitive services enable the delivery of valuable insights into data security, from relevant rules and laws to content moderation.

Healthcare

Innovation typically moves slower in the healthcare industry than others for several reasons, including tight margins, heavy regulation, and siloed research and development. Cognitive services offer the opportunity to lift barriers of innovation and improve the delivery system from organizations down to patients.

Decision-making in healthcare typically happens on a siloed patient-by-patient basis. Cognitive services, by contrast, analyze and act on a comprehensive view of factors that influence health: socioeconomic status, environment, access to healthcare, and so on. Cognitive services can recommend better, more targeted patient care to the physician, including health and wellness programs.

Cognitive services can drive the integration and connection of existing systems within healthcare organizations and unearth essential insights. Suddenly able to aggregate data and connect stakeholder needs, organizations can deliver better care while operating more efficiently.

Intelligence Now

This article has described only a tiny sample of how cognitive services will change the way we think about business and the role that applications can play. In the past, software followed instructions. With cognitive services, solutions can adapt, evolve, and accomplish things that might have seemed impossible just a few years ago. We can’t see all the implications, but from what we know, there is little doubt the impact on business will be profound, positive—and here before you know it.

How can PubNub help you?

This article was originally published on PubNub.com

Our platform helps developers build, deliver, and manage real-time interactivity for web apps, mobile apps, and IoT devices.

The foundation of our platform is the industry's largest and most scalable real-time edge messaging network. With over 15 points-of-presence worldwide supporting 800 million monthly active users, and 99.999% reliability, you'll never have to worry about outages, concurrency limits, or any latency issues caused by traffic spikes.