AI, Blog

What You Need To Know About Chatbots



According to BI Intelligence, about 80% of businesses will use chatbots — which are applications that engage in interactive conversation using natural voice or text – by 2020. Juniper Research also forecasts that this technology will save businesses about $8 billion annually by 2022.

“Due to advancements in AI technology, natural language processing and speech recognition, the cost of developing chatbots has come down drastically, which is fueling the explosive growth of this market,” said Jeri John Deva George, who is the Head of Zoho’s SalesIQ and Cliq.

Yet it’s customer service that chatbots will likely see much of the traction, at least in the next few years. “With chatbots able to engage customers seamlessly around the clock, this is poised to completely change the online customer experience game, while saving time and money,” said Patrick Welch, who is the president and CMO of Bigtincan. “The main goal of these chatbots is to assist customers with getting to their end goal as quickly as possible, whether it is finding out more information, or making a sale. In the end, it’s ideally to replicate the success of top performing customer agents.”

So then, what are the ways to implement chatbots in your organization? What are the best practices and, well, the gotchas? Here are some things to keep in mind:

Set Expectations: Chatbots are not silver bullets. In fact, there are many ways they can go wrong!

“For a great customer experience it is crucial to not try and ‘trick’ customers into thinking the AI chatbot is a real person – and make it clear when they have switched from a bot to a person,” said Chris Bauserman, who is the VP of Product and Segment Marketing at NICE inContact. “Start with a focused pilot covering topics that you already successfully provide self-service options for, test and learn quickly, then iterate and expand from there.”

Understand The Customer Process: Look at ways to better personalize the experience. Otherwise, a chatbot may ultimately be worse than using a traditional approach.

“Customers may get annoyed if they have to repeat all their details to a human agent after having painstakingly typed it into a chatbot interface,” said Michael D. Mills, who is the Senior Vice President of Global Sales at the Contact Center division of CGS. “Failure to centralize customer service information can lead to negative experiences.”

Focus On Data: In other words, there should be ongoing data analytics to understand trends. “This will help a brand build profiles on its customers which will then personalize the experience even more,” said Jonathan Taylor, who is the CTO of zoovu. “Collecting this insight will also help brands understand how the navigation of their site works.”

Think Different With Design: Your experience with designing websites or mobile apps may lead you down the wrong path. Consider that chatbots have their own unique requirements.

“How should your company sound?” said Gillian McCann, who is the Head of Cloud Engineering & Artificial Intelligence at Workgrid Software. “Think carefully about brand and personality and what it says about your company. Also be prepared for users to say the most unexpected things and build in conversation flows that can handle going off topic. There should also be an awareness of local or cultural differences in language.”

The Future of Chatbots

Chatbots certainly have some powerful underlying technology.  According to David Karandish, who is the founder and CEO of  “Natural language processing is used to understand what the user meant. Within NLP, companies use neural networks to make inferences around matching, as well as create the right ensemble of algorithms that use a combination of classical techniques and modern neural networks to ‘vote’ on the best matching response.”

Yes, it’s enough to make your head spin.  But for the most part, chatbots allow for effective interactions at scale.

More importantly, the technology is still in the early phases.  “I must emphasize that chatbots augment humans, not replace them entirely,” said Antonio Cangiano, who is the AI Evangelist at IBM and teaches a course on chatbots. “As a result, it would be a mistake to expect so-called strong AI a la Sci-Fi movies, at this stage.”

This is critical to keep in mind.  But the future does look very promising as chatbots are likely to be impactful for your business. “There will be a shift away from chatbots being simply reactive,” said Stefan Ritter, who is the co-founder and Head of Product at Ruum by SAP. “As AI becomes more advanced and chatbots collect more data, bots will start to develop the ability to predict what a user’s next move might be, or what problem they may be experiencing, and act on it in real time.”

This article was originally posted on –


Mark Halstead

You might also be interested in