How SMBs Can Prepare for an AI-First Service Model

by:
on February 6, 2017

Intelligence—it’s been in the headlines a lot lately. Well not intelligence itself, rather, the word intelligence. For whatever reason, many people have had intelligence both on and (perhaps to a lesser degree) in their minds.

Artificial intelligence (AI), in particular, is an especially hot topic, and it’s clear the business world is turning a corner with its assessment of AI and its potential.

We see evidence of this in everything from discussions at this year’s World Economic Forum to declarations made by in-the-know executives, such as IBM’s CFO, who said that “the debate about whether artificial intelligence is real is over.”

Unlike the flying cars we were promised, AI technology—useful, productive AI technology—is already on shelves—much of it available for free or at a very low cost—and the benefits it can bring businesses are just as available to the smallest mom-and-pop retailers as they are to multinational enterprises. It’s there for the taking; however, there is a steep learning curve.

This report will help you get your AI feet wet by explaining:

What Artificial Intelligence Is & Where Its Service Value Lies
Cloud AI Engines You Can Work With Today
How SMBs Can Prepare to Capitalize on AI

What Artificial Intelligence Is & Where Its Service Value Lies

Artificial intelligence (AI) is a collection of technologies designed to help computers process information more intelligently and, to varying degrees, more autonomously.

Rather than requiring that the user instruct the computer what to do at multiple stages of a process, an AI-enabled computer could simply be told the end goal. It would then determine how to reach the goal with little human assistance.

AI is not the answer to all your customer service challenges, at least not in its present form. But, within the larger field of AI, there are some very compelling applications that have the potential to add immediate value to your service strategy.

Natural language processing (NLP) is one such specialized field within AI, and it’s already being used effectively to power a type of virtual customer assistants (VCAs) known as chatbots.

While forecasts vary, many expect the use of VCS technologies will grow rapidly in coming years. According to research and advisory firm, Gartner:

“By 2020, 25% of customer service and support operations will integrate virtual customer assistant technology across engagement channels, up from less than 2% in 2015.”

From Gartner’s Market Guide for Virtual Customer Assistants (Available to subscribers here).

Chatbots and AI in general have great potential to improve the experience on both sides of the customer service equation.

  • For companies, they can remove some of the monotony faced by first-tier service agents. They can reduce the volume of incoming service requests by helping customers make better use of automated services.
  • For customers, they improve the service experience by facilitating more rapid resolutions. Chatbots respond immediately. They also provide a more seamless service experience because, as digital tools, they are easier to integrate with other digital services.For example, if you’re online and ask a chatbot a question, it can open a web page to show you the answer, something not as easily accomplished by agents on the phone.

Cloud AI Engines You Can Work With Today

Whether you’re a tech enthusiast, a curious early adopter or just a motivated business owner who wants to see what all the hype is about, there are some impressive AI tools you should know about.

They’re impressive because they’re from some of the biggest names in technology and use cutting-edge software to process, understand and even learn from interactions.

These are online tools, which means you can access them from the web browser of any internet-connected computer. On the back end, the AI engine is powered by the specialized hardware that has made recent AI advancements possible.

Further, because these AI tools don’t require local installation, they can be integrated with a wide variety of tech deployments, from mobile apps to online forums and social media.

The bottom line: At this moment when AI tech seems poised to start making real progress, we’re nearly all on the same, even playing field.

So what tools are available and what can they do? Let’s take a look:

Note: All of the below options offer free trials. However, you will need to pay if you deploy a VCA created with one of their services. Costs vary and most are based on frequency of use.

IBM Bluemix

IBM Bluemix lets users create VCAs that are powered by IBM’s Watson artificial intelligence engine.

To create a VCA, users must first train the app on the variety of contexts and questions it will be expected to handle.

IBM Watson VCA

An IBM Watson-powered VCA in the process of being trained

 

After basic training, users can go back and refine the model by adding specific words and phrases, or as IBM calls them, entities:

“An entity is a portion of the user’s input that you can use to provide a different response to a particular intent. Adding values and synonyms to entities helps your bot learn and understand important details that your users mention.”

Entities could be, for example, specific names, models or versions a company has for its products.

Amazon’s Lex Bot

One of Amazon’s many web services is the Lex Bot engine, named after the company’s popular home assistant Alexa. It lets users design and deploy chatbots in a nearly limitless variety of digital contexts.

Example chatbot interaction. On the right is a list of the elements that need to be pre-determined

Questions during registration for Amazon Lex Preview showing deployment and integration options

One nice thing about these chatbot AI engines is that most of them have a per-use pricing model. Compared to the cost of having live agents handle requests, the per-use costs are exceedingly low. For example, here are costs for Amazon Lex:

“You are charged based on the number of text or voice requests processed by your bot, at $0.004 per voice request, and $.00075 per text request. For example, the cost for 1,000 speech requests would be $4.00, and 1,000 text requests would cost $0.75. Your usage is measured in ‘requests processed,’ which are added up at the end of the month to generate your monthly charges.”

Microsoft Azure

Microsoft’s AI offering is powered by the same AI that’s behind the company’s flagship voice assistant, Cortana. Like the options mentioned above, Azure is part of a larger web-computing platform, and chatbots created with it can be integrated with the full of spectrum of common digital services.

Microsoft Bot Framework

Configuration screen for Microsoft’s Bot Framework showing a variety of possible bot deployment channels

 

Microsoft’s chatbot solution has gained a lot of traction, and there are many online examples where you can see it in action. Click here to see their Bot Directory.

designing chatbots

A small selection showing the creativity that goes into designing the wide variety of chatbots

 

Keep in mind, however, that all of these DIY options require some degree of programming knowledge. Not only will you need to train your DIY VCA by familiarizing it with the types of questions to expect and which answers might be appropriate, you’ll also need to know how to deploy it on your website.

In most cases, these tasks can be handled by a company’s in-house IT or web development staff, but more challenging integrations may require outside help.

How SMBs Can Prepare to Capitalize on AI

Chatbots have huge potential to add a variety of efficiencies and customer experience (CX) improvements to a host of standard service interactions. Even if you don’t expect to make use of them in the near-term, there are some customer service technology best practices you can (and should) implement today, that will make your eventual chatbot deployments much easier.

For starters, effective knowledge management is essential to a service organization’s efficacy, from the points of view of both the customer and the company balance sheet. Remember, when you begin training your chatbot, you’ll need to feed it a bunch of text-based questions and answers. This will be a much simpler task if you have that information already well organized, easily accessible and in standardized formats.

Consider the materials you already use to train your service representatives, everything from question banks to the scripts your phone or live chat staff follow. These can provide much of the text and context you’ll need when training a chatbot. They’ll also give indications of what types of inquiries will be best handled by chatbots, and which are best left to human staff.

Lastly, remember that chatbots aren’t independent agents. They can perform some specialized roles, but they typically need to fit into a larger, cohesive service strategy. In other words, if your company tracks service requests on spreadsheets, communicates with email and lacks a centralized system tying the two together, chatbot integration will be challenging.

But if you already have a centralized customer service platform in place, then the chatbot will fit right in! If you’d like help finding the right customer service platform for your company, its customers and its future chatbot denizens, then give us a call at 855-998-8505 for a free, 100 percent human-powered no-obligation consultation.

You may also like

Beating Your Competition With Better Mobile Customer Support

5 Cost-Saving Benefits of a Customer Self-Service Knowledge Base

Guide: How SMBs Can Improve Customer Service With Virtual Customer Assistants

Compare Customer Service Software