“Artificial intelligence” might come across to many as a buzzword used by enthusiastic tech theorists and marketers, but in this series I’ll show you how AI chatbots are being used right now by real companies with real business objectives. I’m not a theorist or a marketer, I’m someone who’s gotten their hands dirty implementing AI chatbot solutions that provide value to businesses, as well as their customers.
So what could an AI Chatbot do for you? If you’ve got a vision of what a perfect bot could contribute, it might look something like this:
- It delights my customers
- It automates tasks, saving my customers and employees time
- It “learns” about my business and improves itself automatically over time
In part one of this series, I discussed how chatbots aren’t generally vehicles of delight (and why that might not necessarily matter). In this article, we’ll go over how chatbots do when it comes to automating tasks and saving businesses and customers time.
Q: Will an AI Chatbot save time for me and my customers?
A: Emphatically, yes!
There are so many opportunities for automation and time savings with a chatbot! The technology is already here, and the only real limitation is your creativity. Many organizations have already automated away routine customer interactions, whether through chatbot-powered Alexa skills or social messaging chatbots.
Impactful chatbot workflows don’t have to be completely autonomous either. One governmental organization I’ve worked with has been using our live chat platform for years and is now moving into conversational AI. They are located in Europe and have strict regulations around personal information. Every time someone wants to chat with an agent about their case, they are required to provide security information like their mother’s maiden name, their case number, etc. Then the agent has to manually verify this information in their customer database.
The organization was complaining that this is a significant inefficiency for them, as agents are left twiddling their thumbs waiting while visitors look for their case number or type in all of the required information. It was also frustrating for visitors, as they may have waited in a queue, only to find that they couldn’t be served before they provided a piece of information they may not have handy.
We set up some bot flows where, if the visitor wanted to discuss their case, the bot would prompt them for the required information, patiently wait for the visitor to provide it, and then confirm the information against their customer database. The chat would then transfer to an agent who could confidently discuss the case, knowing that the visitor was authorized.
Another company I’ve worked with, an Australian internet service provider, receives dozens of chats every day from customers with service availability issues. Many of these issues could be resolved simply by having the customer power their router off and on, leaving agents sitting around and waiting for customers to complete this simple task (and probably bored out of their minds).
We implemented a bot that, when dealing with situations like this, would recommend some basic troubleshooting steps and only connect the customer to the agent after the customer had given them a shot. This simple and easy-to-implement chatbot flow has saved hours of time for their agents, freeing them up for more complicated and stimulating work.
This leads me to a key question everyone thinking about implementing a customer-facing chatbot should ask themselves:
What should be automated by AI, and what should be handled by humans?
The answer to this question really depends on your customers and your business, but in the end there will be a mix of processes that should be handled by bots, processes that should be handled by humans, and processes that are best handled by smooth bot-agent collaboration.
Forrester predicts that advancements in AI won’t make human agents obsolete but will instead lead to the customer support agent becoming a higher-skilled (and better-paying) occupation leveraging AI and requiring advanced technical and interpersonal skills. Just like in chess, the best performances will be a result of successful human-AI collaboration.
I hope you enjoyed this article and learned a thing or two about how chatbots can lead to efficiencies through automation! In the next article in this series, we’ll learn about how chatbots will “learn” and get more effective over time and see if our dreams of a self-learning super-agent match up with the realities on the ground in 2019.
If you’re interested in chatting about AI chatbots and how they can impact your business, let’s talk!