Everything you know about chatbots is wrong

So, the buzz about chatbots has swamped your workplace. If you’re on the verge of commissioning a chatbot development project for your enterprise, it’s the perfect time to understand some misconceptions and make sure you don’t fall for them. This guide brings the most common of these misconceptions to the surface.

We need to build a comprehensive chatbot

Enterprises Think About Chatbots

Yes, a comprehensive chatbot will be great for your enterprise, but if you’re still building your AI-tech abilities, that will take you a very long time. Enterprise IT teams can get stuck on the idea of building a chatbot that does many inter-related tasks (for instance, answering customer queries, helping them select a product, and tell them how to use it). This is possible, but not very easily achievable for enterprises beginning to find their feet in AI technologies. Instead, focus on these points:

  • Identify the most basic, often repeated questions that your customer service team answers.
  • Build very basic chatbots that promise answers to these questions.
  • Enhance them with a mechanism to seamlessly connect the user with service personnel when needed.
  • Focus on enhancing user adoption for your chatbots.
  • The conversations that these basic bots handle will go a long way in helping you build smarter chatbots subsequently.

Chatbots will replace customer service personnel

Just like everything about artificial intelligence, this is more conspiracy theory than anything else. Chatbots must be thought of as, well, programs that understand natural human language, and can quickly search through large databases of massive information, and create outputs in natural language. These bots are meant to assist humans by taking care of their repetitive and mundane tasks, leaving them with more time for value-adding tasks.

For instance, in Deakin University in Victoria, Australia, an IBM Watson-based chatbot called Deakin Genie has been deployed to take the load off teachers and coordinators. The chatbot answers all basic questions that beginner students have about courses, administration, and general campus life. The chatbot has been in action for two years and is getting better at answering questions more accurately, as well as answering a larger number of questions because it’s constantly being exposed to larger datasets.

Aren’t chatbots only good for customer service?

Build a Comprehensive Chatbot

That’s a good question. Though initial use cases and applications have been focused in customer service, it is naive to believe that chatbots are limited in their applications only to this function. In any business function where information exchange is necessary (which basically means every business), chatbots can serve some great purposes.

For starters, chatbots can take the role of virtual assistants for each employee in your enterprise, helping them with inputs where they need them, and enabling them to track and improve their productivity. In human resources, chatbots can go a long way in shortlisting candidates that meet the minimum qualification criteria for a role.

The applications are relevant for a wide range of industries, right from education to retail, from medical to financial. Enterprises need to start with baby steps, and once their chatbots are able to make use of the daily data exposure to improve, they can be augmented to perform more sophisticated functions than merely servicing consumer queries.

Chatbots need to be linked to expansive product info databases

This assumption is true to some extent; after all, your chatbot needs to be “aware” of all the products your enterprise offers, your brands, and your service lines. However, it’s not what enterprises should really focus on, at least for their first chatbot projects. Essential to a chatbot’s success is its ability to sustain conversations. This means that the first priority must be given to using real communication logs from your organization’s databases so that the chatbot can sustain a conversation across a wide range of situations. These information sources include your enterprise customer service call logs, chats, and transcripts.

Chatbots are potential replacement of apps

Chatbots Are Good for Customer Service

The opinions in the tech cybersphere are highly divided. We believe that chatbots and apps are not exactly meant to serve the same purposes. Whereas apps are a mechanism of delivering the complete digital experience that an enterprise has to offer to its customers, chatbots are focused on delivering conversational experiences. This makes text a very important (though not the only one) component of chatbot based experiences.

Apps, on the other hand, are more focused on delivering visually engaging experiences and are also able to deliver sophisticated capabilities (such as sale, profile updates, and account creation) within the app interface. Chatbots, although they can integrate with applications, are not likely to replace apps. When you commission a chatbot development project, don’t attempt to make it a replacement of any of your enterprise apps (unless the app is purely about conversations).

More misconceptions

Apart from the misconceptions covered above, here are some more, which are too nuanced to deserve a separate section, yet common enough to impede your tech team’s vision for leveraging chatbots for great business outcomes.

Chatbots are only text-based: Not true. In fact, voice-activated chatbots are the next in the value chain. Users can even initiate interactions by uploading an image or a video clip. Natural language processing algorithms can work across content formats, and enable chatbots to do the same.

Chatbots always have to work using AI: Not true. Simplistic chatbots can be created purely using conditional logic. Of course, the capabilities of these bots will be meager as compared to, for instance, chatbots that use machine learning.

Chatbots could frustrate users: Agreed. If a chatbot can’t seem to understand a customer’s queries, or can’t get further than a “sorry I can’t help you with this; maybe something else?” kind of messages, users will be frustrated. However, programmers can make chatbots smart enough to “sense” this and route the interaction to a human agent.

See the light

Chatbots can deliver tremendous benefits to enterprises. There are some blind spots though; be smart enough to separate the wheat from the chaff and don’t let misconceptions spoil your chatbot journey.

Photo credit: Shutterstock

Leave a Comment

Your email address will not be published.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Scroll to Top