Follow JRNY on LinkedInFollow JRNY on MediumFollow JRNY on Twitter

Blog updated 20 February 2019

All chat solutions are not created equal.

If you’re considering adopting chat technology to generate and convert more leads for your business, then it’s important to consider what type of solution will best fit your organisation.

There are regular chatbots, and then there are emotionally intelligent digital assistants. If you’re wondering what the difference is, then you’ve come to the right place.

Regular chatbots

Just like most chat solutions, regular chatbots deliver instant responses to customers through your website. Some might also integrate with messaging apps so that your customers can interact with your company through chat channels they regularly use (e.g. Facebook Messenger).

Regular chatbots offer predetermined responses to a recognised message from the user. If the bot doesn’t recognise an exact input, it will display an error message. They deliver the same response to every user, regardless of their mood, tone of voice, or stage in the customer journey.

These chatbots rely heavily on the use of buttons and quick replies. That’s not a bad thing at all and buttons can be very effective; it just means that if the user types a response instead of clicking a button, it’s less likely the chatbot will deliver an appropriate response for the individual user.

Now let’s talk about the next generation of bots which I’ve called ‘emotionally intelligent digital assistants’.

Emotionally intelligent digital assistants

These more sophisticated solutions not only deliver more accurate responses, but can offer different replies based on the user’s customer persona, mood and stage of the customer journey.

Here’s some more detail about some of the technology used to achieve this…

Natural language processing - understanding

A good chatbot needs to understand the words, syntax and deliver an answer.

When the chat engine has then classified the words into categories, the platform then determines semantics - that is, the meaning and context of the user input. Natural language processing is the technology that assists with this part of an emotionally intelligent chat solution. It may be used by some regular chatbot as well.

Machine learning - improving over time

Machine learning is the technology that assists your chat solution in delivering more accurate responses over time. That is, if an input is not recognised the first time, the machine will add this input to its ‘bag of tricks’ and can be taught to deliver the correct response next time. A sophisticated chat solution will deliver better customer experiences over time as it is used by different types of people.

However, simply understanding what customers are saying is not enough to provide the best possible customer service. You also need to learn about their feelings during the interaction and deliver a tailored response.

Sentiment analysis - the emotional element

What is sentiment analysis?

A chatbot, ideally, should not deliver the same experience to everyone it interacts with. Particularly for large enterprises with many different customer personas - like insurance or utilities companies for instance.

Sentiment analysis is a layer on top of the natural language understanding technology that allows the engine to 'understand' the user’s mood by analysing sentence structure.

Here’s an example

User 1 - negative mood:

"My car was broken into!"

User 2 - neutral/positive mood: 

"Hi there, can I please make a claim on my car insurance?"

A human agent can easily detect the difference in moods between the two users. User 1 seems more blunt, maybe in a bad mood and likely wants a solution quickly. User 2 seems more neutral about the situation.

Sentiment analysis is trying to mimic this ability of humans to deliver a response tailored to the user’s mood, but automatically and at scale. Not only can the chat solution determine that both users want to make a claim on their car insurance, but it can offer a response that is tailored to their mood.

A tailored response might offer an angry user quick reply buttons to get their solution quickly, or even offer the opportunity to speak with a human straight away. A neutral or happy person might be more patient and welcome a friendly tone back from the chatbot. They might be more willing to offer feedback about the chat solution at the end of the interaction as well. A chat solution with this ability allows business to deliver a better customer experience.

Going beyond just detecting mood, a chatbot that can categorise a user as a particular customer persona, at a particular stage of the customer journey, will offer even more personalisation.

More personalised interactions with leads can increase lead conversion rates and improve customer satisfaction, and this of course is good news for your bottom line.

At JRNY we build digital assistants for our enterprise clients, and we’ve seen the value in offering a platform that goes beyond just delivering the same answer to every user. Feel free to get in touch if you’re wondering how this could look for your company: letschat@jrny.ai.

Sharing is caring
FacebookLinkedInTwitter

What's next?

Partnering with Insurtechs

Barriers to Partnership with Rita Yates, CEO @ InsurTech Australia

Bindy Egden
October 8, 2019
Artificial Intelligence

Talking Data with Dr Ratneesh Suri, Head of Analytics & AI, IAG

Bindy Egden
September 24, 2019
Ready to
start the JRNY?