Why Machine Learning Can Improve Customer Service

(Image: panuwat phimpha/Shutterstock)

(Image: panuwat phimpha/Shutterstock)

AI is changing our everyday interactions. What once required a human rep can now be handled by a virtual assistant whose programming allows customer problems to be solved more quickly.

A recent Venturebeat article declared, "AI chatbots are the next big shift in customer service." Those of us of a certain generation expect to wait on a line or on the phone for a person to take care of our customer service issues. But the generation that favors texts to calls has come to have different expectations.

A survey of 1,500 millennials in North America found that 65% preferred not to have to engage a human in person or on the phone. That's actually good news for businesses that embrace the technology because they can improve customer satisfaction at lower costs by implementing chatbots and virtual assistants powered by AI.

One of the companies that enables businesses to harness the power of AI in that way is Rul.ai. Yi Zhang, co-founder and CTO of Rul.ai also teaches data mining. She spoke with All Analytics about data mining works for virtual agents to allow for personalized responses.

Just like Amazon and Netflix learn more about customers through their use of the sites and then can offer more on targeted recommendations, interactions with virtual assistants enable companies to "know more about you and provide more" help in finding what you seek, Zhang explains. "We use AI technologies in natural language processing, deep learning to build the next generation of virtual assistants."

[Learn more about AI and Machine Learning. Sign up now for the next edition of the A2 Academy.]

Should organizations reveal that customer queries are being handled by chatbots? Zhang said that it is better for businesses to do so. She concedes that there are some observable differences between "human to human conversations and human to machine conversations." She also explained "Studies show that it would cut back on social chat like 'how are you?'" While this is considered part of polite human interaction, it becomes irrelevant in the context of explaining the problem to a virtual assistant.

An efficient solution to the problem is what virtual assistant are all about. That's the main appeal for the customer, that they can get right to it "rather than waiting 30 minutes" for a human rep to come on. In cases in which the virtual assistant can't solve the problem, the "customer service experience manager can see what's going on and intervene" as necessary, providing guidance for "the chatbot to serve the customer better."

What makes the chatbot effective is that even before it gains the knowledge from its own interactions with customers, it incorporates "a large amount of historical data," as well as the domain-specific knowledge of whoever put in the specifications for it. Zhang points out that "this is very different from the government platforms you have seen in which the engineers write code to build the bot." Theirs is a "self-service solution in which people who have the domain knowledge own and control the bot."

She explained that while "engineers don't like to change the wording of the bot or add" in additional uses cases, customer service managers have no such hesitation and really know what they are doing. "That's why we think they should be in control."

That fits with her view of the democratization of AI and the empowerment of "nonengineers to utilize AI without writing any code." Zhang envisions a future in which AI will be adapted not just to customer service for online sellers but for a range of organizations. "I think AI will be everywhere" she says.

She offered several possibilities for AI enabled automation in the future. One is universities that can use it to automate profile reads for student applications. Another is putting virtual assistants in government offices like the DMV to cut down on waiting time by allowing some tasks to be completed without a human representative.

In Zhang's view, government employees often operate like customer service representatives. Accordingly, the same kind of efficiencies that are realized by utilizing chatbots in the retail space can apply to the government sector.

It's just a matter of thinking through what task needs to be accomplished. As she puts it "Think about the application you want to work out," and then the "machine learning can be used for just about any aspect of the customer experience."

Ariella Brown,

Ariella Brown is a social media consultant, editor, and freelance writer who frequently writes about the application of technology to business. She holds a PhD in English from the City University of New York. Her Twitter handle is @AriellaBrown.

AI, Machine Learning Power Transformation

AI and machine learning are transforming how companies work with big data. Here's how one organization is implementing the technology.

Car Makers Drive Hard Towards AI Advances

One of the most visible applications of AI and machine learning is in the area of connected cars. Here's how data and analytics are driving these advances.

Re: Human touch
  • 5/25/2017 5:31:09 PM

@impactnow If you are brought into regular contact with teens and twentysomethings, you'd find this consistent with your observation of their general preferences. 

Human touch
  • 5/25/2017 5:16:22 PM

"A survey of 1,500 millennials in North America found that 65% preferred not to have to engage a human in person or on the phone."


I can understand the need for fast service but it's interesting that there are so many millennials that don't want to talk to a person. I do like email service and chat at times but I am also still okay with a phone call if needed. If the preferences continue on this trend we will see the customer service reps diminish but service levels will take a dip for more complex calls and issues. For simple tasks its easier to implement AI but for more complicated tasks a person is still required.

Re: More Chatbots
  • 5/25/2017 4:57:51 PM

@Lyndon_Henry I'm certain the people invovled were thining about tourist appeal in the design. Wouldn't they all want a selfie with a robocop?

Re: More Chatbots
  • 5/25/2017 4:27:25 PM


Ariella writes that "BBC reported today ... that Dubai is launching robot polic officers ...."

Here's a YouTube video clip of one of the robocops at a tech show in Dubai. Seems pretty rudimentary to me. Definitely fond of saluting and occasionally shaking a hand Turns out it also will pose with you for selfies ...

Re: More Chatbots
  • 5/25/2017 2:43:38 PM

@PC Ah but T-Mobile is also the company that had the worst human customer service reps I have ever had the misfortune of dealing with. One canceled my service without telling me (I was a customer in good standing, but she didn't like that I was signing up for new deal after my contract was fulfilled (back when all of them worked on contracts(. So I was on vacation away from home without a working phone. I resolved never to use them again and have been with AT&T ever since. It has much, much better customer service, and if I'm paying more for that, it's worth it.

Re: More Chatbots
  • 5/25/2017 2:36:37 PM

< People will keep paying a fee  -- until it all blows up in their faces with an expose on the news. >

Or until the competition offers a better alternative. T-Mobile seems to be doing this for cell phone service. No Fees. No taxes. Unlimited data. Add two lines for the price of one. Every time you turn around there is another new offer.

Stock in the "Uncarrier" is up 70% over the last two years. Meanwhile AT&T and Verizon have been mostly flat.

Re: More Chatbots
  • 5/25/2017 1:04:31 PM

< People will keep paying a fee that shows up on their mortgage statement for years without question because calling to find out what it is a battle. > 

@SaneIT isn't that also what enables other businesses like cell phone carriers get away with bogus charges for millions of customers? I think that some don't even notice, and those who do may not want to do battle over a few dollars a month. But multiply that over millions of customers, and it turns into a substantial unwarranted gain for the business -- until it all blows up in their faces with an expose on the news.

Re: More Chatbots
  • 5/25/2017 1:02:05 PM

@Michelle there are niche publications for more than you can imagine. I've been surprised by some of the titles out there.

Re: More Chatbots
  • 5/25/2017 1:00:22 PM

@Lyndon_Henry the life advice bot is also the one that most caught my eye here. The concept reminds me of a simple program that was around back when I was a kid. You could type in some expression of a problem, and the virtual psychiatrist (though I don't think they were using "virtual" in that way then) would respond "Tell me more about it" or "How do you feel about that?"

Re: More Chatbots
  • 5/25/2017 8:38:56 AM

Having managed a call center group in the past the telephony side of "bots" really hasn't changed in a long time.  They are very simple decision trees intended to do one of two things, gather information that SHOULD be passed to a human when you finally reach one or make you hang up because the frustration with the product you're calling about is lower than the frustration of calling to complain.  When companies grow quickly and don't update these decision trees or have hand offs between phone systems they lose the ability to pass data along but they never lose the ability to make you reconsider your desire to call and complain.  Surprisingly this isn't limited to what we may consider disposable items, this tactic works really well with complex and very expensive purchases/services.  People will keep paying a fee that shows up on their mortgage statement for years without question because calling to find out what it is a battle. 

Page 1 / 4   >   >>