Let’s get social Nearly 70% of today’s consumers use social media to engage with businesses and the trend shows no sign of stopping. + Read More
The idea of automating and scaling one-to-one conversations using AI chatbots has created a lot of hype, but a lot of bots have failed to live up to these lofty expectations. Part of the problem is that many chatbots lack the ability to provide a worthwhile experience for customers. Most people understand that a chatbot can provide an answer to a simple question, but so can an FAQ page. For a chatbot to impress, it needs to be able to understand what a visitor is trying to do, help them do it, or put them in touch with a person who can. This goal inspired the latest version of our AI powered chatbot.
In June we added a machine learning and NLP engine to power our bot, and with this update we are putting that engine to work by giving the bot the ability to handle more complex customer requests, route chats intelligently, and converse more naturally. Let’s explore some of the new and exciting features of our chatbot:
If a chatbot is to transcend a strictly informational role, it must interact with the business ecosystem around it. With our last chatbot update, we gave it transactional capability which allowed it to do things like place orders, book meetings, and check on order statuses. With this update, our chatbot has more tools at its disposal to help visitors accomplish their goals:
Webviews: These are custom HTML pages which can be displayed inside the chat window. These pages work like web browsers and can be shown to visitors when a certain intent is detected. This is handy for complex, multi-step forms which require data from APIs to populate their fields. An example of this might be a flight booking process where a visitor first needs to specify when and where they want to go, then select from a list of itineraries, and finally provide identity information.
Dynamic Forms: Forms can now change dynamically based on prior field selections. For example, if a visitor selects BC as their province, only cities in BC will be available as options under city.
Location Based Responses: A very common question for businesses with multiple locations is: “which of your locations is nearest to me”. Without knowing where a visitor is chatting from makes answering this question impossible. Our latest version of our chatbot takes care of this problem. Once a visitor shares their location with the bot, it can contextualize any follow-up questions based on their location.
One of the biggest problems with chatbots is interacting with them can often feel… too robotic. Having a bot that doesn’t repeat itself or hit you with a wall of unmanageable text, can make it easier to interact with and more approachable. That’s why we’ve introduced a series of new features aimed at creating a better, more conversational experience for your visitors:
Broken Up Responses: Instead of serving up large blocks of text, our Chatbot can break responses into bite size chunks, just like humans do when they text. This is useful for providing complicated step-by-step instructions to support questions or for asking follow-up questions after an answer has been provided.
Conversational Forms: Our bot can now gather information using a series of questions rather than presenting visitors with a form. A visitor that was ready to convert might be discouraged when presented with a lengthy form. Use this new functionality to make providing necessary information feel like a conversation rather than a chore.
Answer Variety: A bot that says the same things over and over can seem like an annoying younger sibling (“I know what you are, but what am I?”). Now you can program our chatbot to have multiple responses to the same intent. This way, users that ask the same question twice won’t get a carbon-copy response each time. This will make your chatbot seem a little more personable and a lot less repetitive.
Quick Replies and Buttons: Suggested responses can now be displayed as buttons above the typing area. Selecting one of these buttons can trigger a follow up response, direct users to a URL, or prompt them to fill in a form. This is especially useful for mobile visitors for whom typing can be a challenge. It can also help direct users down a path that a bot is equipped to handle, leading to more satisfying visitor experiences.
Everyone knows that there are some conversations that bots just aren’t meant to handle. When visitors are getting frustrated or a VIP logs on, businesses want to be able to take the reins of a chat or at least monitor what’s going on. Still, very few chatbots offer the visibility and escalation tools needed for companies to have confidence that a chatbot will raise its hand when it needs help. With this update, we’ve added much more flexibility for administrators to define what the chatbot to agent workflow looks like:
Smart Triggers: A new system for defining how the bot-agent relationship will work has been added to the Comm100 control panel called “Smart Triggers”. “Smart Triggers” can use visitor information, message content, intents, or repetitive bot responses as conditions to trigger follow up actions like immediately routing a chat to an agent or sending a notification to an administrator. This makes it easy to ensure that the chatbot knows what to do if it gets stuck.
Bot Specific Routing: The same routing rules that applied to live chat can also be applied to the bot. Before this release bots could only take a specified percentage of total chats, now you can specify what percentage of chats you want the bot handle when a routing rule is triggered. For example, if you had a routing rule which said that anyone with a shopping cart value over $1,000 needs to be routed to your best agents, you could make sure none of those chats are ever handled by our chatbot.
If you’d like to find out more about our next-level chatbot, get in touch with our team! We’d be happy to answer any questions you have and show you how our latest chatbot can be a trustworthy, capable, and approachable companion to your customer service team.