How to score your first AI quick wins: FAQ chatbots

Posted by Stephanie Organ on 21-Nov-2019 10:16:20

In Artificial Intelligence, Chatbots, quick win projects

Taking on Artificial Intelligence (AI) in your organisation, particularly for the first time, can be risky business, which is why it is so important to have quick win projects. Quick win projects will motivate your team and build up confidence for coming projects, but they will also provide immediate value for your company. Building up momentum is essential when introducing AI, which is why we’ve put together this quick win series to make it easier for companies to take those first steps with AI. Let’s jump in with the first in the list, FAQ chatbots.

A bit about the bots

Don’t flinch, they’re really not what they used to be! Chatbots have come a long way since all the hype back in the early days. Just as the omniscient hype cycle would predict, expectations would grow and grow, and when the bots didn’t deliver, chatbot technology was shunned into the trough of disillusionment until slowly but surely, the technology made some real progress and is finally delivering true value. The power of chatbots is finally being harnessed as they are no longer being deployed as standalone solutions, but rather integrated tools that streamline a wider process. And this is being reflected in the market which is currently facing extraordinary growth, with 80% of all companies expected to have chatbot solutions set up by the end of 2020.

As the chatbot industry advances, things like improved natural language processing and sentiment analysis mean that chatbots can provide better services, understanding more complex request, giving more fluid responses, and assisting in a wider range of areas than before. When it comes to building your own bot there will be a play off between cost and complexity, but as long as you keep your companies needs in mind, you’ll get a great return on investment either way.

The value

Chatbots are the ideal project to boost revenue as they save time otherwise wasted answering mundane questions and increase availability of support, which is what your customers want. With someone available 24/7 you’ll have happier customers and happier staff who will be able to focus on more engaging and productive tasks.

In the age of impatience, people want answers straight away. No one wants to wait on hold on for an answer that could have been served up immediately by a bot. Customers are demanding 24 hour support from their favourite brands, and you can give it to them without having to cover the costs of all those extra staff. As with any AI, you should build it because it serves a purpose, not just because you can. This is why we think that building an FAQ chatbot is a great starter project as it’s hard to think of a company that wouldn’t benefit from one to some extent.

If your company has an FAQ page, spends a lot of time answering repetitive queries, or has staff that need to access a knowledgebase for answers, setting up an FAQ chatbot will undoubtably save you time and money. If taking the time to set up the knowledgebase for the bot outweighs the time saved by the bot, then it’s already a clear cut path.

The project

You could invest a lot of money in creating a top quality chatbot that can guide your customers right through your sales process from start to end, making intelligent suggestions judging from the customer’s sentiment, but that’s going a little overboard for your first project when you could simply follow this guide to get your very own FAQ bot up and running for a fraction of the cost. Additionally, no coding skills are required so there’s no need to hire developers to get the job done.

There are several tools available to help you create your bot without coding. Our go to choice is Microsoft QnAmaker.ai which converts information about your organisation into a structured knowledgebase of linked questions and answers that can be used to train the QnA bot on Azure Bot Services. QnAmaker is a really painless method to give you a coherent solution that can be scaled up using different packages.

Spoke is a self-service AI designed to support internal processes which looks up answers to employees questions and directing those it can’t answer to the right departments. It can save up to 50% of support’s time by answering itself and its knowledgebase can be gradually built up over time by first using it as a ticketing system.

Chatfuel is a popular chatbot builder than integrates seamlessly with Facebook messenger making it a great option for those with no coding skills and companies who mostly engage with customers via Facebook. You can get the basic features for free, commit to $15 a month if you want to delve a little deeper, or you can dip into premium packages if you want to spend a little more and receive better support.

These suggestions all cover slightly different needs, so you should pick the bot that best fits your companies requirements, which should be related to the goals you want to set for what you want the FAQ bot to achieve. Setting these goals is an important step in measuring the ROI.

Next you need to consolidate your knowledgebase from FAQs, training materials, handbooks, etc and feed this into the bot for training. Most tools include inbuilt testing so that you can try different ways to phrase your questions and match the new phrases to in-built answers. Ask a few people to get involved in this stage so that you can cover lots of different phrasing scenarios.

Now you're ready to deploy. Make sure the bot is accessible in the relevant places and monitor the progress. It is important to check whether people are getting use out of the bot and to calculate it ROI.

More quick wins

If you enjoyed this project and would like to discover more ways to win fast with AI, or if this project wasn't relevant to you but you're still itching to get to get started on something else, you might be interested in our 7 Quick Wins Projects guide. Download it now and get your primary projects under way.