Masterclass: Writing chatbot answers your customers want to read

You’ve built your chatbot model and spent countless hours training and testing it to within an inch of its life.  

And you’re confident you’ve done a good job in creating a chatbot that delivers a great customer experience.  

But have you thought carefully about the answers your chatbot is giving?  

Badly written answers can have a significant impact on your customers, leaving them with a bad experience and negative thoughts about your company.  

Obviously, this is not what you want!  It’s crucial you get this part right too, so here’s a few tips to bear in mind when creating the content for your chatbot answers:



Answer length

Try to avoid long answers – be as concise as possible – your customers might be put off having to read a very wordy or long-winded answer, or simply not have the time.  

If the question is simple and short, then the answer should also be simple and short.  

If you cannot avoid a simple answer, try to incorporate a link where your customers can get further information if they want – let your customer decide how much they want to read up on.  

You will also need to consider the channel the chatbot will be used on, and the length of the answer may be dependent on this - whether it’s on mobile, on a website or using voice.  

You might want to consider different answer variants depending on the channel.


Avoid starting answers with 'yes' or 'no'

Don’t start any of your answers with a ‘yes’ or a ‘no’.  

Instead, try to partially repeat the question content within the answer, and try to write the answers that fit as many possibly variants of the question as possible.  

This acts as an indirect confirmation and will give your customers the confidence that your chatbot has understood the question they’ve asked, instead of the chatbot just answering ‘Yes, that’s right, you can do!’.


Independant answers

Create answers that will make sense even without looking at the question.




Answer variations

Consider adding answer variations to a few of your answers, perhaps to the more important or frequently asked questions.  

This will make your chatbot more dynamic, less repetitive and help to give that human touch.


Avoid open questions being asked by your chatbot

These types of questions have a higher risk of your chatbot misunderstanding your customers and getting the response wrong. 

For example, you should avoid questions like ‘What would you like to do?’.  

A closed question would be more appropriate here, like ‘Would you like to see our client portfolio?’ or ‘Would you like to find out more about our customer retention rate?’.  

Try to lead your customers to ask questions that you know your chatbot will answer!  


Type of language

You’ll need to consider if you should create your answers using formal or informal language.  

If the chatbot is for a financial or legal institution for example, you’ll most definitely need to use a more formal tone.  

Perhaps you’re a start-up or a small independent company, so you might want the tone of your answers to be more informal.  

You’ll also need to take into account your customers, if your target audience is for a younger generation, match the tone of language appropriately to a young audience.


Multi-media content

Think about using more than just text-based answers.  

Use links to provide additional information, buttons that can be clicked, or add an appropriate image to accompany the text to give an interesting angle to the answer.


All these tips are worth trying to incorporate into your own chatbot answers, and always bear in mind how you want your bot to be experienced by your customers.  

Don’t let your chatbot ruin your reputation!

To find out more, visit

Read through our library of useful content in the QBox blog.

Ready to give QBox a try? Get 5 free tests here.

Alison Houston

Alison is our Data Model Analyst and builds and trains chatbot models for clients. She also provides advice and troubleshooting support for clients who are struggling with the performance of their own chatbots.

Follow me on LinkedIn