Digital Technology
Aug 03, 2017

Beyond home devices: Conversational AI for business

August 3, 2017 - In today’s device-driven landscape, the idea of speaking to a machine and having it reply doesn’t sound like something dreamt up in a Hollywood studio. Anyone with an iPhone can say, “Hi Siri” to know this is true. And Amazon Alexa owners know how easy it is check the weather, play their favorite Spotify playlist or hail an Uber all through voice commands. This human-machine interaction is called conversational artificial intelligence (AI).

What is conversational AI?

Since the beginning of computing, machines have been very literal devices. Users had to input precise codes or commands to make queries or elicit a desired response, everything was binary – black and white. But now, with breakthroughs in AI, like natural language processing (NLP), machine learning and cognitive computing, machines can understand, process and respond to natural human speech and text.

So whereas before, if someone wanted to check the weather, they would go to a weather provider, type in a specific zip code or address, and define a period of time. Now, they can simply ask Siri or Alexa, “Will it rain tomorrow?” These systems will understand they’re asking about the weather, use their location information, parse the timing information, and retrieve the relevant data to provide an answer using natural language. Then, the user might ask subsequent questions about other events on their calendar or current news and receive responses – much like a normal conversation.

How does this human-machine interaction work?

In short, it requires three parts:

  • Intent analysis: How an individual speaks or expresses themselves is often different from how someone else would. Intent analysis processes the different nuances of natural, human language and derives the intended meaning behind it – no matter who the user is.
  • Corpus mining  and knowledge graphs: The machine will absorb knowledge from all the information available, convert it to a knowledge graph – a web of connected information and meanings with weighted probability of accuracy to understand the context so it can intelligently reply.
  • Conversation thread management: Rather than plain A-to-B requests and responses, there is a back and forth in common conversation. Now, machines can track the threads of a discussion and carry the context through the progression of its interaction with a user for a more fluid, organic experience.

How can I make this work for my business?

Conversational AI is transforming how people not only interact with machines and devices, but also businesses. One of the most promising areas that companies can apply conversational AI is customer service via chatbots.

Customers can now reach out to a business using a messenger – on the organization’s website or through a popular platform like Facebook Messenger – and ask common questions, get updates on their personal information or even make purchases. But they’re never actually interacting with a human staff member. They’re talking to a chatbot just like if they were communicating with a person. Customers get their requests and questions fulfilled quickly, which adds to a better experience. Furthermore, it frees employees up to work on more complex customer cases and inquiries that require analyses.

One of the front-runners in chatbot technology is WeChat, a Chinese messenger application developed by Tencent that is now integrating conversational AI with payment and other services to increase the power of the platform. On WeChat, users can chat with their friends and family like any other messaging app. But they can also chat with businesses. Many of the 10 million official business accounts on WeChat have a mixture of bots and real-live customer service agents. In fact, restaurants in Beijing can routinely have WeChat bots take customer orders and bill for payments right inside the messenger application.

This is where the future of customer service is heading. AI will become more prevalent as a means to drive better customer experiences, especially as companies look to digitize their business processes.

According to a Forrester report titled “Artificial Intelligence Revitalizes BPM,” published on February 10, 2017, by Rob Koplowitz, "Digital transformation is the strategic goal. As processes traverse the customer engagement experience all the way to the back office, the ability to drive end-to-end digital process automation moves from vision to reality. Other emerging trends, like low-code development, lower traditional barriers to automation and allow the long tail of processes to be digitized, furthering the goal of true digital automation.”

About the author

Sanjay Srivastava

Sanjay Srivastava

Chief Digital Officer

Sanjay Srivastava is Chief Digital Officer, where he runs Genpact’s growing Digital business, overseeing the Genpact Cora platform and all Digital products and services.