How to Deploy a Knowledge Base Chatbot
9 Jul 2024
5.5 min read
How many questions do your customer service team answer every day where the answer is on your website, or in a public knowledge base? If it’s a lot, you might want to look into AI chatbots, specifically a knowledge base chatbot.
What is a knowledge base chatbot?
A knowledge base chatbot is an AI-powered chatbot that answers customer questions based on information in your company's knowledge base. It can sometimes be called an FAQ bot because it is summarising information that might be contained in FAQ pages.
Are knowledge base chatbots new?
In a sense, not really. There have been chatbots that use suggested articles to help solve customer issues. In that sense they act a little like search engines, matching keywords to articles.
This experience has its place, but it’s not great for the customer. Using a chatbot like a search engine that just directs customers to read a whole page to find the one bit of information is not the most helpful thing.
More modern knowledge base chatbots go a step further. Using the same sort of tech that underpins ChatGPT and other generative AI engines, the chatbot essentially finds the relevant information from the knowledge base or FAQ and tells the customer the answer.
Implementing a robust knowledge management strategy is crucial to ensure that the chatbot can access and provide accurate information efficiently.
This means that the exchange is much more like an actual question and answer than a glorified search engine. The chatbot can even rephrase the answer, rather than simply repeating it.
For example, take the question: “Do you ship to the UK?”. An old school chatbot will identify “ship” as a keyword and will offer a link to a number of articles related to shipping. Slightly smarter ones will understand the question is about shipping locations and prioritise articles about that. So the response looks something like this:
A new knowledge base chatbot will understand the context, will then search the knowledge base for information about shipping, and to the UK, and can relay that answer. This looks something like this:
The difference is clear. This approach:
Understands the full context of the question
Gives a natural sounding response
Requires no extra effort for the customer
Adds extra, useful context in a short message
The benefits of a knowledge base chatbot
Relieve pressure on your agents
As we said at the start, if your agents are having to provide answers that customers could find for themselves, then a knowledge base chatbot can take that pressure off by effectively answering customer questions.
This frees up your agents to focus on questions they are better suited to answer, and means that your customers will get better answers, faster.
Look at the questions you get regularly, and if they fall into this category then you can start to build a business case for a knowledge base chatbot.
Improved customer experience
Knowledge base chatbots are a great method for helping customers to self-serve by providing self-service resources. Many customers would prefer to find the answer out for themselves, rather than speak to an agent.
This dual benefit enhances the experience for both your customers and your support team.
Knowledge base chatbots can answer these sorts of quick questions, and save customers time that would otherwise be spent wading through page after page of product info.
Used well, they also help route customers who just have a quick question away from agents who can help customers who need a bit more help.
Removing barriers to purchase
Answering key pre-purchase questions through a knowledge base chatbot removes any friction from the buying process. Rather than getting bored or distracted waiting for an agent to answer the question, the knowledge base bot can provide an answer in seconds.
This removes any obstacles on the path to purchase, and should increase conversion rates ultimately.
Constantly updated answers
If you change something within your knowledge base, such as adding a new destination you ship to, or adjusting your returns policy, then it can be a pain to update the information everywhere that is relevant.
With a knowledge base chatbot, that process is much simpler through, because the chatbot can easily be retrained. This can even be made automatic, so that it’s always checking for the latest information available.
How to get started with a customer support strategy
Getting started is easy.
The first step is to ensure that your knowledge base is in a format that is easy to access by a bot. If it lives on your website under a public URL, you are already there.
The next step is to decide how you want your bot to respond, i.e. what tone you want it to take. This can be playful, or serious or anything in between. Your support team should be involved in deciding the tone and voice of the bot to ensure it aligns with your brand.
Finally, speak to a knowledge base chatbot provider to get set up, and go through their process. Some are entirely self-serve, while others will help walk you through it, or even do it for you.
After that, you should think about what comes next.
Setting Up a Custom Knowledge Base
Setting up a custom knowledge base is an essential step in building a knowledge base chatbot. A custom knowledge base allows you to create a centralized repository of information that is specific to your company’s products or services. You can create knowledge base articles that provide detailed information about your products or services, including troubleshooting guides, FAQs, and how-to tutorials.
To set up a custom knowledge base, you need to identify the information that you want to include in the knowledge base. This can include product information, customer support articles, and FAQs. You can also use analytics tools to identify the most common customer queries and create knowledge base articles that address those queries.
What's beyond a knowledge base chatbot?
These bots are a great way to dip your toe into AI for customer service. But they don’t tell the whole story of what is possible with AI.
Think again about the questions your customer service team get asked. How many customers could be served by a knowledge base chatbot, compared to those requiring specific order information?
The chances are the ones that could be answered by this bot will amount to around 10-20% of questions at best. Whereas questions about late or missing orders, the status of a specific return or refund, and so on, will be a much higher share. Being able to answer these questions with a chatbot will be much more valuable to you and your customers.
Enter DigitalGenius. Through deep integrations with your carriers, your order management system, payments system and the rest of your tech stack, our chatbot can pull information from orders and relay that to customers. But more than that, it can then take actions depending on the information available.
If the order appears to be lost, a replacement can be shipped. If a return has been received, a refund can be issued early. If a package is damaged, a refund can be offered.
This allows retailers to fully resolve pre- and post-purchase questions from customers, helping them to manage peaks more successfully and help relieve pressure on their team.
To find out more about how DigitalGenius can help you create a knowledge base chatbot, and go beyond, read our customer stories here. Then make sure you speak to our team.