AI Chatbot Buyers Guide:
How to Pick the Right Chatbot for Your Organization
We've all, no doubt, browsed a webpage and received a message asking if we need help finding anything. Or we've opened a chat session and been greeted by a humanlike stand-in who asks a few questions and collects basic information before introducing us to a human agent. Maybe you've even had the opportunity to complete transactions in chat conversations, like transferring money between bank accounts, without needing to talk to a human.
These experiences are all made possible by chatbots, a growing technology that is quickly becoming the new way of doing business while putting customer experience (CX) front and center. The era of conversational commerce is here.
A chatbot is a computer program powered by artificial intelligence (AI) that users can communicate with in the digital world. A chatbot can be deployed on websites, mobile apps, social media and messenger apps, SMS/text, and other communication channels to serve as front-line virtual agents to website visitors and customers. With a conversational interface, chatbots are helping customers do everything from understanding shipping and return policies to booking entire international vacations.
There are several different types of chatbots, from basic scripted chatbots to cognitive AI chatbots. And there's a world of difference from one end of the spectrum to the other.
The simplest type of chatbot can serve up scripted answers (i.e. answers written in advance by humans) in response to an exact keyword the user enters. These scripted bots are quick and easy to get up-and-running, and they are relatively inexpensive. But scripted, keyword chatbots can only handle very narrow use cases and often leave the user frustrated and seeking human assistance as quickly as possible.
That's because real humans don't adhere to scripts. Customers don't want – and frankly don't know how – to conform to a bot's constrained workflow. Communication quickly breaks down when the customer goes beyond those scripted answers.
While some companies start with a basic bot for handling very simple inquiries, many quickly find that they want to do more with chatbots than just deliver canned answers in response to an exact keyword match. If you want to use chatbots to mirror human conversation, you need look to artificial intelligence.
AI-powered chatbots allow companies to provide more value to their customers because they have the ability to answer queries in a more human-like way, preserving meaning and context from one response to another. To do this, AI chatbots need Natural Language Processing (NLP) capabilities. This AI chatbot technology enables them to understand the intent of a phrase or sentence and respond to that intent, rather than to a keyword. Check out the difference between chatbots that do and do not understand intent and retain context:
AI-powered chatbots do more than react to keywords; they understand intent.
AI-powered chatbots retain context throughout the entire conversation.
When you have a truly conversational chatbot, the result is a much more human-like interaction. And this capability to have a back-and-forth dialogue is built into the software– you just give it the data (more on that below).
The highest level of chatbot sophistication is Cognitive AI. Cognitive AI is capable of doing just about anything including driving cars, predicting cancer, and beating humans at Jeopardy. These chatbots use the AI technologies we just talked about – Natural Language Processing to understand intents, and conversation skills to retain context – and takes them one step further.
Cognitive chatbots algorithms that are created by data scientists to organize vast amounts of unstructured data, recognize patterns and relationships in the data, and make predictions about the future. Cognitive AI uses a technology called Machine Learning, which ranges in complexity. Simple machine learning clusters data into similar groups and maps those groups to outcomes that it was given. For example, if you teach a computer that everything shaped like this is called a circle, then every time it encounters a shape like this: O it will treat it like a circle.
In complex machine learning, or deep learning, the outcomes are not given to the computer ahead of time. Instead, the computer reaches conclusions on its own based on experience. Humans do not know what the output will be or how the computer reached that outcome. Deep learning is the technology behind these eyebrow-raising accomplishments. And while this technology will certainly change what the future looks like for humans, it is still in infancy. When outcomes are out of human control, the results are unpredictable and can be undesirable.
For example, in a 2016, in a failed experiment that will forever live in infamy, Microsoft launched its chatbot "Tay" on Twitter, and it began learning from all the people it interacted with. Within 24 hours, the company had to pull the plug because Tay started spitting out racist, misogynistic, and politically polarizing tweets. Cognitive AI can and will learn anything and everything, and humans do not know what the output will be.
Of course, businesses need to have more control over the conclusions that artificial intelligence draws, or your chatbot may end up giving away all of your products or services for free after "learning" that it makes customers happy! This is why we recommend human-supervised machine learning. With a human-supervised approach, businesses not only avoid undesirable outcomes, but also cut the significant cost and time required to set up a chatbot with unlimited capabilities. As a business owner in today's world, you aren't trying to cure diseases or solve other complex human problems; You just need to serve your customers as quickly and efficiently as possible. So, choose a chatbot designed just for that purpose. Nothing more. Nothing less.
Companies can stand up a chatbot on a number of applications or websites for customers to access from a computer, smartphone, tablet, or other digital device. Some examples include:
In order to streamline the customer experience, companies often choose to have one chatbot that remains ubiquitous across all of their messaging platforms. Customers become accustomed to engaging with this bot, personality and all, on whichever platform they prefer.
Businesses with a large customer base and/or limited resources can benefit from automating customer service in the form of virtual assistants or chatbots. Chatbots can answer simple, repetitive inquires that often eat up the majority of an agent's time. This allows customers to be served immediately, at any time of the day or night, and frees up customer service agents for more complex issues and higher-value work that require a human touch.
Consumers can ask questions, get advice, and understand options before they make a purchase. 77% of users believe that immediate online help would increase the likelihood of completing online transactions more often, and 85% of businesses believe that immediate online help would improve online sales conversion rates.1
AI-powered messaging makes this possible without forcing consumers to make an unwanted phone call or visit a store. Consumers can easily get the information they need to make a purchase within the online channels they're already using every day. It keeps consumers digital while delivering a personalized experience at scale.
Customers hate reaching out to live agents when they can have their questions answered quickly by a chatbot.
32% of Americans would rather give up sex for a week than wait in a customer service queue.
3 out of 4 consumers prefer to interact with a chatbot than fill out an email form and wait until the next day for a response.
85% of customers prefer shopping online to in-store, as long as they can get help when they need it.
Over 70% of customers believe brands should leverage technology (such as AI) to reduce the time it takes to resolve an inquiry.
With answers at their fingertips, customer satisfaction improves. If the question cannot be answered by the chatbot, intelligent routing transfers customers to a live agent, who can provide the information that they're looking for.
Not all chatbots are customer-facing. There are many business benefits to providing your customer service agents with their own chatbots. These behind-the-scenes bots follow along as an agent chats with a customer, and searches all integrated systems and tools for relevant information, and delivers that information to the agent right in their workspace. Bold360 calls these agents "Smart Advisors."
On average, less than half (49%) of customer enquiries are resolved during the first interaction. A possible reason for this is that 68% of business respondents agree that their agents struggle with the volume of customer enquiries.2 Like a true work BFF, chatbots can help take work off a busy agent's plate – taking on those repetitive, mundane requests that bog down their day.
If a customer interaction needs to transition to a live agent, the chatbot can quickly and efficiently gather necessary customer information — including the customer's history with the brand and the context of the interaction so far — and deliver that information to the agent so they can dive into solving the issue right away. Having their own chatbot working for them behind the scenes greatly improves customer service agent productivity.
1 Fifth Quadrant, Conversational Commerce and ChatBots: Business & Consumer Usage and Attitudes, Nov 2016
2 LogMeIn 2018 AI Customer Experience Report: Impact of Chatbots and AI in the Customer Journey, 2018
Discover smarter, more personalized engagement.