Will Chat Bots Be Boon or Bust for Customer Experience?


Share on LinkedIn

One of the hottest topics today revolves around chat bots, “computer programs that mimic conversations with people using artificial intelligence [AI]” … to “transform the way you interact with the internet from a series of self-initiated tasks to quasi-conversation”1. What’s all the fuss? And will these chat bots be a boon, or a bust, for customer experience? Tell me what you’re experiencing with chat bots as a customer, and from your company’s point of view.

Let’s first look at a handful of chat bots, and then see how they could help improve customer service and customer experience, finishing with some warnings and suggested steps.

  • Kik’s Weather Channel chat bot that “remembers your zip code” and gives you local forecasts, but it isn’t “too smart: Ask it “How hot is it?” and the bot prompts you to stick to its prescribed inputs””1.
  • Microsoft’s Tay, a Twitter chat bot launched in May 2016 that quickly bombed when it “learned” off-color comments from customers and played them back, and was quickly pulled down.
  • DoNotPay, a chat bot created by a young Stanford undergraduate student, that allows customers to contest parking tickets (in the UK and NYC, today), prompting them to enter problems such as “The offence was before I bought the car” or “I entered the incorrect date on a permit”, with more than 160,000 successful rollbacks.2
  • Tim, “a Virtual Agent answering Airbus Helicopters’s clients FAQ. He guides them through the Customer Extranet portal. He is an expert of the Technical Documentations proposed on the site and is dedicated to make the customer experience easier.”3

How can these and other chat bots help improve customer service and customer experience? Chat bots, and their precursors “virtual agents” (such as Alaska Airlines’ Ask Jenn4) and early AI apps, promise to satisfy two of the seven Drivers of Best Service that my co-author David Jaffe and I presented in our 1st book The Best Service is No Service: “Create engaging self-service” and “Deliver great service experiences 5.

Chat bots can also tick off many of the 39 sub-needs in our latest book Your Customer Rules!6, underneath the seven Customer Needs for Me2B customer experiences, for example:

  • Predict intelligently (part of “You know me, you remember me”)
  • “Give me the advice that I need, when I need it (part of “You give me choices”)
  • I don’t want to have to contact you (part of “You make it easy for me”)

Related, chat bots can help reduce customer service operating costs by deflecting contacts from agents to self-service; take advantage of “crowd sourcing” to improve solutions across many users; and lead to higher levels of revenues, as many chat bots are gearing up to arrange for reservations or upgrades.

Reflecting on these early days, do chat bots “cut the mustard”, and can they truly improve customer service and customer experience? Some are already making a difference, like DoNotPay and Ask Jenn, but others are simply falling all over themselves to be cool, without a clear differentiation from “classic” web or IVR self-service. Some also remind me of a solution that sputtered 20 years called Intellisystems that promised to “help callers diagnose and resolve support issues via telephone, fax, e-mail or the Internet. The system works much like a human expert, asking a series of questions to analyze the situation, then offering specific resolution steps.”7 This promising system sank when it led customers down paths that frustrated them, much like situations where companies force their agents to use scripts and the customer exclaims “I already tried that!”

What are some of the early warnings and lessons learned applying chat bots to customer service and customer experience?

1. Find a real problem, or as some concrete producers splash on the sides of their trucks, “Find a hole and fill it.” Figure out where classic web and IVR self-service aren’t working by calculating your current self-service FCR like you do in your contact centers, and consider replacing those with FCR, or containment rates, that are under 40% with chat bots. Examine where your customers might prefer a speech-enabled Q&A process that chat bots can uniquely meet, perhaps by text or speech mining their comments to surveys or in their conversations with your agents.

2. Guide the customer, but do it flexibly. Don’t confine customers to your chat bot’s “prescribed inputs” but do provide the steps needed to solve the problem, like DoNotPay is doing.

3. Learn from your customers, but install guardrails. The Vine chat bot by Kik responds with “Sorry, we’re keeping this PG-13. Try again!” when asked inappropriate questions.

4. Track usage, then refresh and update or replace your solutions. Nothing new here, but with the rapid adoption and equally rapid failure of chat bots, if the chat bot isn’t getting used, or if it’s being abused, drop it and replace with something else.

5. Follow tried-and-true AI solutions that overcome known weaknesses. These three weaknesses can kill your chat bots: (a) “The program cannot modify any of the components of the architecture; (b) The program cannot modify itself; and (c) The program cannot “learn” a function outside of its hypothesis space.”8 You might want to use expert 3rd-parties to create, host, and upgrade your chat bots, like you might be already doing with apps and contact center software.

6. Don’t over rely on chat bots. Keep the personal element and personal connections, enabling “escape to agent” as you already do with your IVR system and should be doing with your web support. In some cases, and for some customers, the live support is still the best solution – not as a crutch but when the automation is stumbling or frustrates customers or, as we describe in The Best Service is No Service, when both customers and your company find “value” in the interaction.

1. Julia Carrie Wong, “What is a chat bot, and should I be using one?”, the guardian, cited in https://www.theguardian.com/technology/2016/apr/06/what-is-chat-bot-kik-bot-shop-messaging-platform

2. Will Oremus, “”World’s First Robot Lawyer” Is a Chat Bot That’s Actually Useful”, Slate, cited in http://www.slate.com/blogs/future_tense/2016/06/29/a_chatbot_for_parking_tickets_shows_how_bots_can_actually_be_useful.html

3. Chatbots.org, a compendium of 478 chat bots, and counting! https://www.chatbots.org/applications/customer_service/

4. https://www.alaskaair.com/content/about-us/site-info/ask-jenn.aspx

5. The Best Service is No Service: How to Liberate Your Customers From Customer Service, Keep Them Happy, and Control Costs (Wiley, 2008). The full list of the seven Drivers of Best Service = Eliminate dumb contacts, Create engaging self-service, Be proactive, Make it easy to contact your company, Own the actions across the company, Listen and act, and Deliver great service experiences.

6. Your Customer Rules! Delivering the Me2B Experiences That Today’s Customers Demand (Wiley 2015). Here are the 7 Customer Needs that lead to a winning Me2B culture, breaking down into 39 sub-needs:
1. “You know me, you remember me”
2. “You give me choices”
3. “You make it easy for me”
4. “You value me”
5. “You trust me”
6. “You surprise me with stuff that I can’t imagine”
7. “You help me do better, you help me do more”

7. http://www.prnewswire.com/news-releases/intellisystems-announces-email-back-capability-77320427.html

8. Oren Etzioni, “Deep Learning Isn’t a Dangerous Magic Genie. It’s Just Math”, cited in http://www.wired.com/2016/06/deep-learning-isnt-dangerous-magic-genie-just-math/

Bill Price

Bill Price is the President of Driva Solutions (a customer service and customer experience consultancy), an Advisor to Antuit, co-founded the LimeBridge Global Alliance, chairs the Global Operations Council, teaches at the University of Washington and Stanford MBA programs, and is the lead author of The Best Service is No Service and Your Customer Rules! Bill served as Amazon.com's first Global VP of Customer Service and held senior positions at MCI, ACP, and McKinsey. Bill graduated from Dartmouth (BA) and Stanford (MBA).


  1. Great tips. Do you mind if we share with our customers at UBot Studio, who are building their own bots?

    The really revolutionary bots will take all this into account – and we’ll probably stop referring to these earlier renditions of chat bots as bots at all. Honestly, at UBot Studio we’ve been wondering what are bots meant to do for a long time. Customers of ours have been building much more complicated “bots” than this for 6+ years now, and most don’t converse! Hopefully soon we’ll move towards bots serving functions for people instead of functioning based on the limitations of our ‘bot-building’ imaginations.



Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here