8 Ways to Improve Chatbots and Boost Customer Satisfaction


Here are eight ways to boost the performance of contact centre chatbots and improve Customer Satisfaction (CSat).

1. Put an Escalation Path in Place

Where chatbots add most value for businesses is in answering routine questions quickly. If the questioning becomes more complex, the business needs to have an option to escalate the interaction to a human advisor.

That advisor should be able to pick up the thread of what has gone before and, where necessary, switch the call from a chat window to a voice interaction.

This escalation path is critically important because if the customer is left feeling their query has been inadequately handled, they may stop doing business with the brand.

This escalation path is critically important because if the customer is left feeling their query has been inadequately handled, they may stop doing business with the brand.

Jeremy Payne

It’s also important that businesses put sufficient resources in place to manage any escalations, ensure customers’ waiting times are kept to a minimum and that appropriate service level agreements are implemented and always met.

2. Keep Feeding Chatbots With New Information

While chatbots do fulfil an invaluable role within many contact centres, businesses need to realise that implementing and then operating them can be a major commitment.

Jeremy Payne

Contrary to popular belief, chatbots are not cheap to run, they can be time-consuming to maintain and they need to be fed continuously with real-time relevant information to stay up to date.

If an organisation neglects its chatbots, fails to provide them with sufficient data, educate them or give them what they need to learn and grow, those chatbots will become redundant, in the same way that a website that has not been updated will.

So, the lesson for businesses is clear: keep your chatbots fed at all times.

Thanks to Jeremy Payne at Enghouse Interactive

For more information on implementing chatbots within your contact centre, read this article on 7 Things They Won’t Tell You About Installing a Chatbot.

3. Make Chatbots More Empathetic

Chatbots are usually efficient and a tremendous help in a variety of cases. But, as with any other automated systems, customers engaging with a chatbot can sometimes feel a lack of empathy or find it a struggle to express their emotions during conversations.

As Genesys recommend, contact centres should aim to make chatbots much more empathetic. They need the

  • ability to capture customer sentiment throughout the conversations, without explicitly asking (implicit measure)
  • ability to adapt the conversation flow and the responses to sentiments
  • ability to explicitly ask at the end of the conversation how the customer feels (explicit measure)
  • ability to compare implicit and explicit measures to improve sentiment capture over time
Remy Claret

Remy Claret

To do this, training is critical. The conversation set that is used to train the chatbot before production should cover various emotional situations and emotions and be part of the reporting process on real conversations.

By doing this, the contact centre enables the tuning teams to figure out the emotional circumstances where the chatbots are not performing well and keep training them to exhibit satisfactory behaviour!

Thanks to Remy Claret at Genesys

4. Use Them to Collect Information in the Initial Part of Conversations

The secret to improving CSat of chatbots is knowing when and how to implement them.

Mashud Ahmed

Mashud Ahmed

If the contact centre has already successfully implemented self-service, and improved CSat by doing so, then it could be the right time to add bots to the mix.

Start with existing live chat offerings, or social media messaging applications – like Facebook Messenger, and then introduce bots.

Once bots have been introduced, use them to automate the repetitive tasks of collecting information at the start of any dialogue and, ultimately, directing the caller to another self-service option or a live advisor.

Thanks to Mashud Ahmed at Puzzel

5. Use Natural Language Processing (NLP) to Make Chatbots Seem Friendlier

Make your chatbot friendlier without being intrusive. It is important to strike an appropriate tone for your audience.

Through Natural Language Processing (NLP), the system learns to recognise real language. This engagement with language allows the software to utilise data to provide customers with personalised information.

Also, the software is able to analyse the customer’s tone, allowing the chatbot to interact and engage conversationally.

Susannah Richardson

In many cases, customers have felt that they are receiving correspondence from a real person, rather than robotic software. With this more colloquial tone, the software will extend the brand into an interactive experience for consumers.

The double benefit of this for the company is that transcripts of conversations can be easily managed and analysed.

Providing further customer insights, the software allows the company to understand what type of customers are visiting their website, what they are looking for and recurring issues that need resolving.

Thanks to Susannah Richardson at IFS | mplsystems

6. Use Interaction Analytics to Make Chatbots More Conversational

The way that people communicate on the phone is different from the way they communicate using a text message, chat, or email.

It is therefore important to be able to identify these differences and then “coach” chatbots accordingly, to ensure customers feel satisfied with the interaction and its outcome.

Frank Sherlock

The goal is to make these chatbots use the same communications approach as a human when carrying on a conversation.

To ‘coach’ chatbots to adapt their style to reflect the way customers prefer to do things, it would be preferable if they have access to analysis of every chatbot interaction. This will ensure that the differences between the preferred approaches on different channels is understood to as great a degree as possible.

Chatbot adaptations would be based on the insights gathered from interaction analytics, and could include changing the way that conversation is started to avoid triggering a negative reaction.

Thanks to Frank Sherlock at CallMiner

7. Leverage Data

With Gartner predicting that the vast majority of customer interactions will be powered by chatbots in 2020, it is more important than ever for organisations to step up their game in terms of learning from customer interactions.

Modern technologies such as machine learning and neural networks can help make use of the data, learn from it and use it to inform how a chatbot should react to new interaction types.

Prelini Udayan-Chiechi

Modern technologies such as machine learning and neural networks can help make use of the data, learn from it and use it to inform how a chatbot should react to new interaction types.

Using technologies like this will reduce the workload for advisors, leaving them the time to dedicate attention to more complex, nuanced and human issues.

Chatbots like Zendesk’s Answerbot offer an advanced knowledge-base solution that utilises NLP to intelligently serve up relevant issue resolutions to customers’ most common queries.

8. Understand Their Limitations

While bots can be “trained” in empathy and friendliness, customers will always appreciate the human touch when dealing with more complex queries, where empathy and understanding can make all the difference.

Prelini Udayan-Chiechi

Prelini Udayan-Chiechi

While chatbots are well placed to deal with the most common issues, like package tracking and information requests, there are issues that still need the human touch, like bespoke product advice.

We see chatbots working best when they’re working hand in hand with advisors. Dealing with gigantic quantities of low-complexity issues is where chatbots shine, particularly when they’re empowered to seamlessly hand off cases to an advisor when needed.

Take the holiday shopping season, for example – Black Friday, Cyber Monday, Christmas – the time when retailers are at their busiest. Research recently conducted by Zendesk found that ticket volumes can increase by as much as 42% over the holidays. It’s at times like these when chatbots can really be a good addition to a contact centre.

Thanks to Prelini Udayan-Chiechi at Zendesk

Do you have any other suggestions for improving chatbots to boost CSat?

Please leave your thoughts in an email to Call Centre Helper.

Published On: 12th Mar 2018 - Last modified: 9th May 2019
Read more about - Technology, , , , , , , , ,


Get the latest exciting call centre reports, specialist whitepapers, interesting case-studies and industry events straight to your inbox.