Customer support chatbot
Here you can understand what is a Customer support chatbot and how you can build one (or at least to give you the right direction).
Types of customer support chatbots, based on their functionality:
Button based
They are simple for use and implementation. Easy to use if you are the developer or the client. In most cases, they are at a low cost.
ML (Machine Learning)/Natural language-based (you can see them as an AI chatbots at some places, but those are different terms)
They are the most expensive at the time we are writing this article. They can communicate at a higher level with your clients. So they are the most powerful chatbot that you can use. They are harder for implementation and setup, so you will need an experienced developer.
Hybrid chatbots
They are the best of the “both” worlds. You can use the button-based chatbot for the most common questions and enable the power of the ML when you need (for example) some fields for address, where the user could miss some important information, but the bot will guide him and understand him.
From where to start?
Answer these five questions and based on your answer you can choose three different types of customer support chatbot development:
- Do I have online requests and/or questions from customers?
- What is the number of those requests/questions?
- Are the majority of the requests/questions similar or different?
- Where is the conversation (platform)?
- What is my budget?
Types of customer support chatbots, based on dev approach:
Drag & drop solutions
Questions:
- Do I have online requests and/or questions from customers? – Yes, I have
- What is the number of those requests/questions? – Small or medium
- Are the majority of the requests/questions similar or different? – They are similar
- Where is the conversation (platform)?
- What is my budget? – Small or medium
Here we have platforms that give you the opportunity for building (in most cases) a generic chatbot.
Their pros:
- Cheap – Generic platforms, you have to make your bot alone, because they have big client base.
- Easy-to-use – Even people that are not developers could create a chatbot that works.
- Fast for building – If you have the idea, you can create a chatbot for less than a week.
Their cons:
- Function limits – It’s a broad platform for the masses, that must be used with ease, so do not expect complicated functions.
- Pay-per-user base – If your users raise, probably you will pay a lot more.
- Less user support – You are paying low, so if this platform is without a big community, do not expect personal care. Of course, there are some companies that are exceptions.
- Expensive if you want extra features – As we said, they profit if they have many users, who pay a small price. If you want custom work and solutions, probably you will pay a lot.
Example companies:
- Manychat
- Activechat
Chatbot development agencies
Questions:
- Do I have online requests and/or questions from customers? – Yes, I have
- What is the number of those requests/questions? – Small, medium or large
- Are the majority of the requests/questions similar or different? – Most of them are similar
- Where is the conversation (platform)?
- What is my budget? – Small, medium or large
Their pros:
- Agile and working by requirements – If the chatbot company is good enough you will get what you want, without any limitations. Here we have your idea like it is in your mind with your goals.
- Dedicated support – You are paying more, but getting the personal support and help from the DEV company.
- Own framework – Those companies work with own software, not depending on external services for the core code.
- Machine learning features – Yes, many of them give you the chance to have a really smart chatbot.
Their cons:
- More expensive – You are getting more, so you will pay more.
- Users can’t edit the chatbot (only the company) – If they use their own framework, you can’t edit your chatbot alone, but this way you won’t need to dedicate time for deep testing.
- Some use drag and drop platforms – For us this is a big con. Why? You are paying for the setup of a chatbot in 3rd party platform, not for real development.
Example company:
Linbots.com
Corporate solutions
Questions:
- Do I have online requests and/or questions from customers? – Yes, I have
- What is the number of those requests/questions? – Medium or large
- Are the majority of the requests/questions similar or different? – All different kinds
- Where is the conversation (platform)?
- What is my budget? – Medium or large
Pros:
- Agile and working on requirements – Again we have custom chatbots
- Dedicated support – Many time whole support team
- Own framework – Built from scratch
- Advanced machine learning features and big data – Those companies can provide more ML related best practices
Cons:
- Most of them are expensive
- Users can’t edit the chatbot (only the company)
Example companies
- NativeChat
- Outsystems
Types of customer support chatbots, based on platforms:
Self-hosted and/or system integrated
- JS – Here you have self-hosted (for example JavaScript) application, that is not-dependable of social media or another kind of logins.
- Zendesk – Powerfull solution with good integration to your current system.
Third-party (with options for integration)
- Facebook Messenger
- Viber
- Telegram
ML functions
- Dialogflow
Some best practices
- Create answers for the most common questions
- Use images
- Send to URLs
- Give option for a live agent (if you can)
- Monitor the bot
- Errors
- Analytics
