How Does Chatbot-Driven Digital Self-Service Work?

Chatbots and virtual assistants create a conversational experience. They help people find information and answers to their questions. They also help people get things done by using automated tasks and workflows.

These tools achieve this digital-self service by using Natural Language Processing (NLP) to program computers to process large amounts of natural language (the way humans speak). NLP allows computers to recognize and generate speech and interpret sentiment when interacting with humans. 

NLU (natural language understanding), a subset of natural language processing, is essential to achieving NLP. It centers on machine reading comprehension by transforming human language into a “machine-readable" format.

The difference is that NLP processes language so machines and humans can have a “back-and-forth” conversation with each other, while NLU rearranges unstructured data so machines can understand it and analyze it.

Conversational AI needs both. NLP and NLU drive conversational AI to create a digital self-service experience that’s:

  • Immediate - Empowers customers and employees to self-serve and quickly get answers to their questions.
  • Conversational - Guides customers and employees to solve problems, complete tasks, increase productivity and decrease operational costs.
  • Automated - Enables employees and customers to quickly and easily complete tasks.

All of this means chatbots and virtual assistants are getting smarter. That’s a good thing.

What Tools Create a Great Digital Self-Service Experience?

In an AWS environment, Amazon Lex allows you to quickly and easily build conversational interfaces. You can embed them into major messaging platforms, websites, search and mobile applications using NLU to drive chatbot responses that are lifelike and helpful.

It’s the technology used in the popular virtual assistant Amazon Alexa, which not only can hold chatbot conversations but also “assists” users by answering questions and performing basic tasks. This includes sharing what time it is, delivering a weather report or finding the perfect song by a favorite rock star upon request.

NLU and high-quality, automated speech recognition (ASR) combine to create a speech-language understanding (SLU) system. This allows the Amazon Lex-driven chatbot to understand the intent behind a user’s text or speech input.

In chatbot terms, an intent is an action a user wants the bot to perform. Each bot can support one or more related intents. The SLU system accepts the natural-language speech and/or text input, understands the input’s intent and fulfills the intent by providing the most relevant response.

Developers can easily create an Amazon Lex chatbot without any machine-learning expertise. All they need to do is specify the basic conversation flow in the Amazon Lex console. Lex takes care of the rest. It builds the natural-language model that the bot uses to interact with text and voice. It manages the dialogue to dynamically adjust conversational responses and actions based on questions and answers provided by the user. It gets smarter over time by using deep-learning models.

How does search fit into all of this? It is quite simple. Search is made smarter by taking deep-learning models and applying them to information as it flows in and out of the indexes. This is where Amazon Kendra comes into the picture.

Amazon Kendra, a natural language enterprise search service from Amazon Web Services and Smart Answers, uses native language to answer questions. This approach isn’t the case with all machine-learning-based search solutions. Kendra is different.

Conversational AI to build voice or text-based bot interfaces. Uses deep-learning techniques for automatic speech recognition and for natural-language understanding of the users intent. Its natural-language ability changes the game, providing real-life answers to user questions. Kendra also provides and ranks documents featuring essential answers using deep-learning language models.

This deep learning ability enables Kendra to integrate with an Amazon-Lex-driven chatbot. This means users can easily ask questions using natural language and send those questions through the same deep-learning model. Kendra will easily match the analysis of their question, regardless of how they asked it. The analysis is stored in the index and the system will deliver the answer in a few milliseconds, surfacing it to the Lex-powered chatbot conversation. Digital self-service is that simple.

How Can Employees Use Digital Self-Service During the Workday?

For employees, digital self-service helps them get their jobs done without fuss or time-consuming wild goose chases to get the information they need when they need it. The pandemic has created many more cloud workers who are doing their jobs apart from each other and on-site company resources.

A February 2018 study from Forrester Consulting defined a “cloud worker” as one who:

  • Uses a laptop and/or tablet for work purposes
  • Uses cloud apps daily
  • Spends three or more hours per workday using a web browser

Most of these workers were in IT, product development, engineering and marketing. In 2020, the scope changed rapidly, expanding to encompass all industries. Now, most business environments have quickly become mobile out of necessity and will continue to do so into the future as far-flung teams connect effortlessly and virtually.

Companies such as Nationwide Insurance, Quora, Facebook, Twitter, to name just a few, are moving to partial, remote-first or optional remote work environments. It’s becoming the norm, rather than the exception. We predict this model will only increase across all business sectors, turning increasingly more employees into “cloud workers.”

This transformation to a more remote, digital work experience has revealed a hard truth; the old ways of doing things no longer work in the new now.

A 2020 report from IDG and CIO reveals that 54% of IT executives indicate this shift to remote work has forced digitization and automation of existing workflows and processes. Sixty-four percent add they’ve been seeking to improve operational efficiency, a notable result of successful digital transformation.

It’s crucial to examine how today’s workers do their jobs and find the information they need to do it successfully. A dispersed workforce requires new ways of communicating and locating information anywhere, at any time. The acceleration of digital transformation due to the pandemic necessitates faster self-service and enterprise search for work-from-home employees.

Chat is a powerful workplace tool. As more people work outside the office, they are rapidly discovering chat quickly facilitates many ad hoc interactions, such as pop-up conversations, people asking questions and expecting rapid answers and more.

What if a chatbot joined those conversations? Organizations that use chatbots move all of their FAQs, knowledge articles, project information, status reports and other kinds of vital data into a search index. That index makes the information searchable and integrates with backend databases and warehouses.

These tools, combined with automation, facilitate ad hoc queries and help workers find what they need, when they need it and get their jobs done faster and more efficiently.

chatbot answering questionsLet’s say a worker, Jill, wants to find information about her employer’s vacation policy. With a chatbot in place, she doesn’t need to send emails to various co-workers.

Instead, she goes to the company’s chat app that uses a bot and types in, “What is the latest vacation policy?” Once she does this, the bot will search through connected data repositories and deliver the answer back to her.

Integrated automation allows the chatbot to take things a step further. It asks Jill if she wants to submit a vacation request. If she replies, “Yes,” the chatbot will then walk her through the request process, asking pertinent questions to help her submit her vacation request through the chat app.

Another employee, Raj, needs to create sandboxes but can’t remember all of the steps. He hasn’t built one in a while. Instead of spending a huge amount of time tracking down information about how to do this, he goes to the organization’s chat app and asks the chatbot, “How do I set up a sandbox?”

The chatbot then searches for information related to Raj’s query and returns an answer that shares the procedure. It also knows that, because he made the request, Raj wants to set up a sandbox environment. It asks Raj if he wants to do that. If he says, “Yes,” the chatbot will guide him through each step of the process.

In both cases, Raj and Jill used digital self-service to get the answers to their questions and get things done without help from a human support agent or the people around them in the workplace.

Using Chatbots to Improve Customers’ Digital Experiences

The situation for customers isn’t all that different than that of internal employees. No matter where these customers are and what they do, they’re trying to accomplish many different things and sometimes need help doing it.

They need to find information, and the way they do has become increasingly digital. They need to pay bills, make insurance claims or even find a recipe and then locate where they can get the ingredients. They might need customer support to help them find a technician or troubleshoot a challenge.

COVID-19 has driven much of this to an online setting. But, these customers often struggle to find these answers online via public search engines or those on a company’s website. Consider the following example.

customer experiences are reimagined with a chatbotDave, a salon owner, discovers that his shop’s 15-year-old dryer broke down. It’s only blowing cold air, so it’s likely the heating element. Dave hopes that it’s something that can be repaired instead of being replaced, so he looks at the label on the dryer to get the make, model and serial number.

Armed with this information, Dave visits the manufacturer’s website to search the support portal and FAQs. He wants to find manuals or information about repairs. He doesn’t get what he needs, so he calls the contact number listed on the support site. The long wait to be connected to an agent frustrates him, so he hangs up and goes back to searching online.

He tries to locate a certified technician for this brand in his area but can’t find any information. Finally, he searches for appliance repair companies, locates a technician with good ratings and gets the dryer fixed after a rather time-consuming process of making it happen.

What would that same situation look like if that manufacturer offered an Amazon Kendra + Lex chatbot platform for its customers?

The structured, well-architected solution would index all of the pertinent information, from knowledge bases to online manuals to FAQs, and make it easily available to users when they arrive at the website seeking help.

Instead, Dave would visit the dryer manufacturer’s website just as before, but this time he would enter his issue into the chatbot. The chatbot would then use natural-language processing to ask Dave a few questions about the unit’s make, model and serial number.

Dave would input that information so the chatbot could process it and perform a Kendra search. It would respond that it would appear the dryer is having issues with its heating element.

The chatbot would then ask Dave to share his salon’s address and return the name of a certified technician in the area, asking if Dave would want to book an appointment. Once Dave says, “Yes,” and the chatbot would guide him through a conversation that ultimately results in a booked appointment.

In the end, digital self-service through chatbots allows customers like Dave to quickly find what they need. It’s also efficient for Dave’s dryer manufacturer to put a chatbot on its site. That solution doesn’t involve a lengthy phone call with its support staff and keeps the customer moving effortlessly throughout the workflow. Chatbots deliver a productive, efficient experience for everyone involved.

Post Your Comments


Search Blog

AI Chatbot Workshop

Meet the Author

Bryan McKay, Practice Lead - Enterprise Search in the Cloud PS Practice

Bryan McKay, Practice Lead - Enterprise Search in the Cloud PS Practice

Bryan leads Onix’s Enterprise Search practice, helping customers find what they are looking for faster. He implements enterprise search technology to help customers grow their business, leverage information in systems and silos across the organization, tap into the network of expertise in the organization to improve collaboration and innovation, and deliver an enhanced customer experience.

More Posts By Bryan McKay, Practice Lead - Enterprise Search in the Cloud PS Practice
 

Discover the Power of Conversational AI

Dive deeper into the benefits of digital self-service with a private Chatbot Workshop.

Learn More and Sign Up