Guide to Employing AI Chatbots for Your Customer Service

Lisa Burcl

With the proliferation of technologies and brands on the market, finding the perfect chatbot for customer service can be bewildering but it doesn’t have to be.

The past several years have witnessed an explosion of chatbot technologies designed to (hopefully) make customer engagement more instantaneous and seamless for consumers and less time consuming and onerous for businesses.  And as with any early-stage technology, we are currently experiencing a massive proliferation of standards and brands vying to dominate the market (or, at least steal a piece of it) and, of course, watching bemusedly while the programmers work out all the bugs.

For the savvy business owner or business unit manager, though, this confusing survival-of-the-fittest technology throw-down leads to just one all-important question:

How do I know which chatbot my company should be using?  

This is the right time to think about using AI to your business whether you're running a SaaS, ecommerce or service-based business.

Understand the strategy behind the choice of AI chatbots for customer support

If you are reading this article, you are already aware of the massive potential for cost savings a good chatbot represents--resolving up to 80% of customers’ needs without any human intervention, cutting customer service spend by up to 30%, and free your people up for the real, irreplaceable work of actual empathy. The AI-based chatbot cost nowadays also varies based on your business needs and scale.

Still, if you’ve waited ‘til now to onboard a chatbot of your own, you may have given yourself an advantage-giving the technologists a chance to work all the bugs out on the early adopters’ nickel for example.  Or letting strong contenders with expertise in your own niche to rise above the general milieu of the market and prove their staying power. These factors mean you’ve never been in a stronger position to get the right technology for your business.

Still, implementing chatbot technology just because “everyone else has one” is about as likely to result in a good business decision as anything else influenced by the fear of missing out-namely, nil.  Effective use of chatbots requires a keen understanding of your own sales and service pipeline, and how and where exactly a chatbot contributes to your bottom line, and (even more importantly) to the value your company creates for customers.  So, that is where you need to start, by determining who you are going to help with the chatbot, and what they will need to experience the interaction positively.

Choose the AI technology of chatbots that works best for you

Once clear on what your chatbot needs to bring to the party, you are competent to identify the technology and level of service you’ll need to meet those needs.  

Baseline qualifications

The best chatbots of 2018 share a few standard features that should be present no matter what option you select.  If any of these are missing, it should raise an immediate red flag:

First, your chatbot technology should have the ability to recognize when a question is beyond its capacity and escalate it to a human customer service representative-preferably with a transcript or summary of the conversation the bot had with the customer, so your person can step seamlessly into the conversation.  A system that’s inherently bad at this will probably make a career out of aggravating your clientele, and that’s not net positive for anyone.

Second, it should be easy to actually set up the chatbot for customer service.  That means one of three things: If it’s a high-end system, it should come with strong customer support and a high degree of customizability.  If it is designed for custom in-house programming, it should have an open API and/or easy access to third-party extensions. Finally, if it is meant to be “trained” by non-programmers, it needs to have an exceedingly straightforward user interface that your staff can get comfortable using.

Lastly, you need to be sure your chatbot covers the right social platforms.  If you need something that will provide continuity across several platforms as well as your website, your options will tend toward the higher end, whereas you will have more (and less expensive) options if you only need to focus on one or two social outlets.  Whatever your needs, there is a chatbot out there for you. It makes no sense to end up with one that’s weak at your social platform of choice.

It’s worth noting that the majority of good chatbot technologies have provision for a free level or trial period, so if you’re uncertain about your selection, there is no reason not to “try before you buy.”

With these baseline provisions out of the way, you can get clear about what kind of chatbot you need.  Fortunately, these can be boiled down to just a few groups. Since most chatbot reviews we’ve seen don’t break these down very conveniently, we include a short list of the “big names” within each category here:

The ‘Cadillac’ chatbots

These high-end, enterprise-level chatbots typically feature voice recognition (in addition to text capabilities), the ability to draw upon external knowledge bases (such as customer records, pre-existing FAQ pages, and pretty much any other outside source of information), industry-specific options, and heightened reliance on AI for a conversational experience.  They also tend to feature stronger customer support and rely heavily on APIs--the coding involved in deploying these is not for the faint of heart.




IBM Watson





The “easy setup” chatbots

These systems rely on a hybrid of AI and rules-based learning to enable more “in-house” direction and control at a more reasonable price point in chatbot deployment.  These also support enabling the chatbot for knowledge base use, but typically require companies to employ their own “chatbot whisperer” to analyze conversations and tweak the chatbot’s AI logic.  Don’t be fooled by this more rugged approach to AI, though setting up one of these chatbots for customer service can reduce the burden on your call centers by 50% within a few weeks, and up to 70% or 80% with a bit more fine-tuning.



Flow XO

The “specialized functionalities” chatbots

In addition to the packages mentioned above, there are also a few that stand out in their own niches, particularly in professional services, where attuning chatbots for advanced knowledge base access can actually provide accurate up-to-the-minute information faster than a human being.

Clare.AI (Banking)

NanoRep (Financial management)

The Facebook chatbots

These are so prevalent as to merit their own category.  Some of these have the ability to cover one or two other social media platforms, but for the most part, they are Facebook-specific, and focus on auto-respond functionality and automating marketing campaigns in Messenger.



Its Alive



Octane AI

Wanna grow with us?

Schedule a call with our growth marketers and we will help you analyze your marketing, product, and metrics. We will present a Growth Plan with strategies to help you increase traffic, and convert better.

Let's Chat