Not all bots are created equal. To find one that can help you meet business goals, you’re going to need one that can do much more than obey a few simple commands or respond humorously to questions. If your criteria for success goes beyond entertainment and engagement, and you’re looking to achieve tangible business benefits — such as reducing call volumes to the contact center or offsetting online shopping cart abandonment over the holiday season — then understanding the difference between chatbots is critical.

If you’ve ever engaged a virtual assistant like Alexa (or Google Assistant, Siri, or Cortana), then you probably don’t have a lot of confidence in their ability to help improve your business in a serious way. And you’re correct to be skeptical. These types of chatbots won’t be able to help you decrease operational costs or improve your Net Promoter Score. They lack awareness of context, and as such, they don’t empower consumers to conduct and complete transactions.

These types of chatbots operate without a specific context and are classified as “unbound” — that is, their territory includes any realm of knowledge and consumer need. Commonly known as virtual assistants, they’re great for turning on your lights or playing your favorite song. And if they make a mistake, it’s really no big deal. You’ll eventually get the right song to play without too many tries.

In business, the stakes are much higher, and losing one customer because of a preventable error is one too many. It can be devastating to a brand if a chatbot gives a customer the wrong information, for example about coverage on an insurance policy, only to discover the error when it’s needed the most. This is where intelligent chatbots come in.

Business bots have to be smarter

Intelligent chatbots rely on deeper and more robust artificial intelligence (AI) and are integrated with essential enterprise systems that hold customer data. They not only parse big data for key pieces of information, but they can even learn from it. Intelligent chatbots that are built for the enterprise and serve in customer-facing roles are known as virtual agents or virtual customer assistants. In some cases they are powerful enough to conduct transactions and resolve customers’ issues as well as humans can.

One important reason some chatbots can do this is that they are “bounded.” They operate within a relatively narrow context framed by the industry or business in which they are deployed. Whether deployed in a wireless carrier, utility, bank, retail or ecommerce business, or insurance company or mortgage broker, the context of the industry ensures these intelligent chatbots can support a conversation in a way unbounded ones cannot.

Chatbots aren’t just useful for online chats. They can be deployed on many other interfaces: mobile, social media, messaging apps, voice response, and SMS texts. Most importantly for improving the overall customer experience, intelligent enterprise chatbots can anticipate customer intent.

All about intent

To start, an intelligent chatbot analyzes data such as the customer’s identity, feedback, the channel used for contact, date, time, location, and other information in real time. It also taps into a company’s internal data stored in customer relationship management (CRM) systems and elsewhere to gain more insight into the customer. By analyzing all that data using artificial intelligence and machine learning, the chatbot can anticipate customer needs. This includes using natural language processing (NLP) to analyze text or conversational speech (even slang) to more accurately gauge intent.

One of the hallmarks of excellent customer service is “next issue avoidance.” That means the bot provides not only what the customer needs to get out of the present interaction, but it anticipates and prevents the next likeliest issue. For instance, if a customer has contacted their wireless company to ask about data roaming rates, and data shows that more than half of customers also ask about international voice and data plans, the chatbot can proactively serve up that information.

Should an enterprise chatbot encounter an issue it can’t resolve, it can easily pass the interaction off to a human being. This is another way that bots built for business differ from unbound chatbots. Among other things, that means the chatbot remembers where the customer is in the conversation so that when they’re transferred to a representative, the information can already be queued up with no need to repeat information (a well-known pain point for consumers).

Chatbots can be amusing distractions on social media or in automating your home, but you don’t want to entrust your brand’s reputation to a simple conversational interface with a limited scope, when the technology — if used right — can in fact provide an engaging and useful frontline for customer service. Make sure you provide your customers with the best service possible by investing in the right kind of chatbot.

Scott Horn is the chief marketing officer at [24]7, an AI-driven software and services company.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.