Companies like Amazon are dumping AI-powered chatbots in favor of “dumb” bots that make customer service smarter. That’s because rules-based chatbots are cheaper, easy to implement, 99-percent effective, and solve issues fast. Learn how they can transform your customer service and boost customer satisfaction when you join this VB Live event!

Register here for free.

“We have pretty strong opinions about the right and wrong ways to use conversational interfaces,” says Mitch Lee, manager at Credit Karma and co-founder of the instant messenger bot Penny. “People would say, well, with natural language processing we’ll just be able to mimic a human and it’s going to be amazing. That went wrong.”

He points to the Poncho weather bot, which launched with fanfare on Messenger, and then crashed and burned.

“People would spend just as much time trying to figure out how to ask about the weather as actually getting information about the weather, versus just typing into a search bar, “weather San Francisco” and getting the answer you want,” he says. “It’s such a great example of where these bots were going wrong.”

Rules-based bots that offer pre-populated responses are far more accessible and user friendly, he says, and improve the customer experience significantly, especially in complicated situations where the customer requires guidance — but might not even know how to ask for what they need.

“You give someone an empty chat box and say, what do you want to know about your finances, and the number one thing they ask is, ‘How do I save money or how do I invest,'” he says. “But those aren’t very productive questions versus, ‘Hey, what order should I pay off my debt in?’ That’s a question you might not even know that you should ask.”

Pre-populated responses can help provide structure to a conversation and make it more productive, Lee explains, and it’s an especially good use case for support. Email isn’t a great experience because there’s high latency — fire off an email to support, wait a day for the automated response saying somebody will review your question, and then wait longer for a response. Calling and getting trapped in IVR is most customers’ nightmare, and a quick way to help customer satisfaction plummet.

And while live chat can be a great experience, it’s still asynchronous communication, there are often long wait times, and the customer service humans behind the keyboard are often already using canned responses to the customer’s questions — a surefire sign that you can automate that function programmatically, and be looking at major cost savings.

Bots are immediately available, 24 hours a day, and with structured questions and responses, bots can also move a lot faster through problem resolution, because they’re getting exactly the information they require in order to solve the issue. Then, when you hit the odd edge cases that don’t have an automated way to be resolved or require subjective judgment, you can kick it up to a human.

“At least at that point you’ve already dispositioned probably 80 or 90 percent of customer issues,” Lee explains. “Then you’re just left with the last difficult-to-manage 10 percent of questions that humans can take a look at.”

In the end, Lee says, implementing a chatbot comes down to this bottom line: Are you serving people? Are you able to serve more people as a result? Are you providing the most value?

“I think through conversation, the answer is yes,” Lee says.

To learn more about how rules-based bots can fit into your customer service strategy, how to implement and leverage the power of these bots, and the results big companies like Amazon are seeing with simpler bots, don’t miss this VB Live event!

Don’t miss out!

Register here for free.

You’ll learn:

  • The difference between NLP and rules-based bots and why it matters
  • Why companies like Amazon are turning away from natural language processing-driven bots to rules-based bots
  • How to deliver mobile and web-based customer service that works, using the right bots.
  • How rules-based bots make the customer journey more effective


  • Abinash Tripathy, Co-Founder and Chief Strategy Officer, Helpshift
  • Mitch Lee, Manager, Credit Karma and Co-Founder, Penny
  • Leslie Joseph, Principal Analyst, Forrester Research
  • Rachael Brownell, Moderator, VentureBeat

Sponsored by Helpshift