The Many Use Cases for Cognitive Services

7 min readJoe Hanson on Jul 5, 2018
Try PubNub Today

Free up to 1MM monthly messages. No credit card required.

Subscribe to our newsletter

By submitting this form, you are agreeing to our Terms and Conditions and Privacy Policy.

Examples of apps and businesses that are being transformed by cognitive services, and some future use cases as well, to see just how much they’re changing, and will continue to change, the technology ...

Previously, we showed you how companies are building smarter businesses with cognitive services and continue to implement intelligence at the edge. Now let’s walk through some examples of apps and businesses that are being transformed by cognitive services, and some future use cases as well, to see just how much they’re changing, and will continue to change, the technology landscape.

Thanks to cloud giants like AWS, IBM and Microsoft, 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

At the beginning of 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 are one of the earliest forms of AI. While unlikely to pass the Turing test any time soon, they represent the natural evolution of voice-enabled applications. Where once you’d call a support line and press 1 for Accounts Payable, now you can speak in full 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.”

APIs like Watson Assistant or Amazon Lex, however, 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 delays, and change the sequence of options to prioritize rescheduling data. Or they may see that calls from a particular geography tend to be conducted in a different language, and change the default accordingly. They may even be able to identify grammatical patterns that indicate customers who should be immediately forwarded to a supervisor.

Intelligent conversational interfaces, using speech recognition, text-to-speech and machine learning can provide highly engaging experiences and lifelike conversations for any number of purposes. And even better, they’ll learn from those experiences.

Chatbots will change the way we bank, shop and learn, making recommendations, understanding abstract concepts and get 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 an intelligent chatbot 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.

        {"text":"Put off"},
        {"text":"Switch off"},
        {"text":"Turn off"}
      "description":"Turn on intents"
        {"text":"Put on"},
        {"text":"Switch on"},
        {"text":"Turn on"}
      "description":"Turn off intents"

Natural Language Processing

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

This 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 smashed 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 is responding to the product.

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

Natural Language Processing

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:

Watson analyses

Brands already spend massive 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.


Though online shopping has completely changed the way 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 assist shoppers with their questions, make recommendations and even check out.

Nordstrom dominated the 2017 holiday season 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 are also saving us from the dreaded customer support phone call, waiting an hour for a representative to deal with a simple problem. Amazon, with who-knows-how-many orders per day, has deployed chatbots equipped to solve the minor issues most customers have when they need help with their order.
Now that we’ve looked at a couple examples of intelligence in the real world today, let’s peer into the future and see how cognitive services will change the world we live in.

Smart Cities

Cities of the future will rely on a wide variety of integrated intelligent services to make them safer, more efficient and more environmentally conscious. Image recognition and machine vision will play a critical role in this transformation, processing and taking action on the innumerable images that present themselves in urban spaces.

To look at just one potential impact, let’s consider something we’re all passionate about fixing: traffic. The following example is already being tested on a single city block, but there’s no doubt technology like this will soon spread far and wide.

The NYC Department of Transportation partnered with a cognitive API provider called Intelliscape’s image recognition and machine learning technology enables the system to detect traffic jams, weather patterns, parking violations and more, sending real-time alerts to city officials along the way. The intersection is rigged with cameras that capture activity, process it and stream back findings and actionable data in real time (like sending traffic enforcement for illegally parked vehicles, or calling service for broken or malfunctioning equipment).

In addition, the system instantly combines weather, demographic and location-specific data with the original image-recognition data to provide highly enhanced real-time feeds for analytic dashboards.


Global populations continue to grow, and feeding those billions of people will be a huge 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 will incorporate as many useful data points as they can to make intelligent agricultural decisions, even ones that seem counterintuitive. For example, by aggregating real-time weather data, remote sensor data and historic performance, cognitive services will be able to perfect the individual irrigation plan and update it for every day’s unique combination of circumstances.

Data Security

As we grow more connected, and our digital lives overshadow our physical ones, data privacy and security are transforming 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 right 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. Acting as an intelligent question-answering chatbot, cognitive services enable the delivery of valuable insights into data security, from relevant regulations and laws to news and updates.


Innovation typically moves slower in the healthcare industry than others for a number of reasons, including tight margins, heavy regulation and siloed research and development. Cognitive services offer the opportunity to lift those 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 can and act on a comprehensive view of factors that influence health: socioeconomic status, environment, access to healthcare and so on. Supporting the physician, cognitive services can recommend better, more targeted patient care, health plans and wellness programs.

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

Intelligence Now

This is just a small sample of the many ways cognitive services will change how we think about business—and the role that applications can play. In the past, software followed instructions. With cognitive services, it can adapt and 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 already, there is little doubt the impact on business will be profound, positive—and here before you know it.

More from PubNub

Can Empathy Exist in the Metaverse
News May 16, 20221 min read

Can Empathy Exist in the Metaverse

A roundtable discussion led by PubNub’s COO, Casey Clegg, exploring the topics of what it means to be human in a virtual world.

PubNub Staff

PubNub Staff

How to Advance Telehealth and Virtual Care Technologies
News May 2, 20221 min read

How to Advance Telehealth and Virtual Care Technologies

Dr. Joe Kvedar, Chair of the Board for the American Telemedicine Association, joins our COO, Casey Clegg, to discuss why...

PubNub Staff

PubNub Staff

Another Step Towards Data Security: ISO-27001 Implementation
BuildMay 2, 20221 min read

Another Step Towards Data Security: ISO-27001 Implementation

Today, we are glad to announce that we are currently in the process of implementing ISO-27001 security standards.

PubNub Staff

PubNub Staff