chatbot | 7 min read
Guide to Designing an Amazing Chatbot
Published by Bhumish Sheth on 01-Jul-2021
chatbot | 7 min read
Published by Bhumish Sheth on 01-Jul-2021
There are chatbots, and then there are good chatbots that ensure successful conversational outcomes and benefit business processes. Deploying a bot combines both technical and creative talents. Your job isn’t over by just learning how to build a bot on the bot builder. You also need to understand the intricacies of the conversational flow and the strategic reasoning behind it.
Consumers are still not very open to using bots, but with the right design that leads to a sound user experience, 86% of consumers are open to using it. This proves that the formula of chatbot success is majorly attributed to its design. That makes me want to ask you, how well are you designing the conversational interface that provides value to the users and ensures a wholesome conversational experience?
There are some basic protocols to chatbot design that one needs to follow that enhance the chances of bot success.
The term chatbot combines both rule-based and AI chatbots. Rule-based chatbots are built by charting a decision tree. You can develop complex conversational flow charts to fulfil your bot’s purpose without the use of AI. On the other hand, AI bots use Natural Processing Engines (NLP) to answer complex questions. You can either adopt one or build a hybrid bot combining both. To understand their differences in detail, you can check out the following article:
Both have their set of benefits and limitations, but you need to conclude what’s best for you. While deciding, consider the following parameters:
Your business functions largely define what kind of bot you should deploy. Does your business require intelligent conversations with your users? Some sectors like travel, hospitality, eCommerce, and restaurants require AI bots to answer users’ specific questions. But not every conversation needs that level of personalization or intelligence. Define the pain points you’re trying to address and determine if your bot would need personalised conversations or basic conversations. Accordingly, decide if you need an AI or a rule-based bot.
AI bots take a higher amount of investment and technical expertise than rule-based bots. As a result, most SMEs prefer opting for a rule-based approach capable enough to address repetitive and straightforward queries.
Your customer queries can either be simple and can be solved within minutes or can be complex and takes time and effort from the agent to solve. Determining what type of query you receive on an everyday basis can help you choose the right bot. Use a rule-base for simple questions and an AI bot for complex queries. You can also deploy a hybrid bot to cater to both types of queries at once.
Customers are still hesitant to speak to a chatbot, primarily because they cannot offer a humanised conversational experience. Your conversational script is the key to break this barrier.
Bots are conversational marketing tools, which means the conversation is at the crux of it. We cannot speak of designing a chatbot and not talk of the conversational flow. To develop an engaging bot script, make sure you define the following points:
The best part about a conversational platform is you have the liberty to retain the familiarity of replicating a verbal conversation. People are used to having a casual tone on messaging platforms. Use that to your advantage. However, this also depends on the product, the audience and the platform you’re using. A person buying a luxury watch needs to be approached with a different tone and language style than a person looking to buy casual wear. Strike a balance between the right amount of playfulness, formality and personalization.
Imagine a situation where you get a very lengthy message on your phone or even have someone speak for too long at once without allowing you to engage in the conversation. The experience is simply unpleasant. Businesses make a common mistake while developing a bot script. They want to communicate everything at once. If your visitor wanted to read the same thing that they read on the site, what value is your bot providing? The best way to make your bot engaging is to have small message threads instead of text-heavy messages on your bot. It makes the interface look cluttered and boring.
For long businesses have looked for a method to have real-time conversations with their consumers, and now they have it. No one is willing to just read on a bot. They’re looking to talk- let them. Your flow needs to allow input from your customers. Give them options- what are they looking for, what do they want to know, how can a bot help- make it about them and not you.
Many consumers find it creepy and uncomfortable when they try to replicate human-like conversations on a bot. You don’t need to design your bot to be a human to provide an emotional touch to the conversation. It can be what it is and still have an empathetic tone. This can also make your bot stand out and stay longer in the mind of the consumer.
To establish a friendly conversation from the start, let your bot introduce itself. This message holds importance because it will dictate the tone and personality of the rest of the conversation.
Unless you’re deploying an AI bot that can answer open-ended questions, ensure that you provide suitable options for your visitors to choose from. This will also require you to analyze the common customer queries that they’d need quick answers to.
Once you provide a solution to the visitor query, ensure that the visitor is redirected to the main menu. Keep the flow less complicated and easy for the visitor to navigate.
Think of a conversation between two friends on an instant messaging app. It’s filled with emoticons, images, GIFs, and videos. I’m not suggesting overdoing it, but you can use some to add colour to an otherwise mundane chat and make the conversation appealing and engaging for the visitor.
Always check every word, sentence, and phrase in the bot script. I cannot emphasize this enough. No matter how much of a friendly rapport your build with the visitor, it still expects professional decorum from a brand. Hence, even the slightest grammatical error can result in an unpleasant experience for the visitor.
Your bot isn’t just a random interface. It is going to be a representation of your brand. So what are you going to do to differentiate your bot from your competitors? How are you going to use the bot to forge an emotional connection with your consumers? The biggest hurdle that bots need to cross is the ability to..well..not sound like bots. There’s a difference between making it sound human and human-like. Your goal is to achieve the latter. How would your brand talk, represent you, and market you if it was a person. Design your bot in a manner that just one look at the bot and your visitor just knows which brand it’s speaking to.
Bot interfaces actually allow you to replicate your brand image on the bot. Customize the bot accordingly. Give it your name, your brand colour, logo, font, and your preferred language, just like Dominos did it with its bot “Dom” and Whole Foods with its bot. Retain your colour scheme and brand logo in the bot header to provide a branded conversational experience. A renowned hospital, Zydus Hospital named its bot “Zye” that assists website visitor in getting their answers.
Keep two words in mind here: Human Handover. Etch it in your brain, write it on your hand but remember it. You aren’t building your bot to show off your latest tech. You’re doing so to address the gaps in customer communication and streamline business processes through conversation. Hence, conversations are at the core here, not chatbots. At the end of the day, it’s us, humans, that design and builds the bot. There may arise multiple situations where your bot isn’t programmed enough to solve a particular problem. Let’s just say your customer is old school and just doesn’t want any bot interaction. What are you going to do then? The customer service rule book doesn’t allow you to leave your customer hanging just because it refuses to adapt to your method of communication. Never compromise their experience. Know the limitations of your bot and design it to allow a live agent transfer in your bot conversation. Because you’ll always encounter customers that value human interaction above anything else, and there’s nothing wrong with that.
If you’re decided to deploy a chatbot, you’ve already gone through the preliminary research that goes into determining why you need it in the first place. You must’ve identified a business problem. Whatever that may be but realise that the only reason your customer will use it is that they want an answer to their problem. You need to align your goal to the customer’s to utilise the bot’s truest potential. Some of the most common reasons businesses deploy bots are:
Although your motive is to generate leads from a bot, be subtle about it. Think about how you can differentiate your bot from a form. Consider the following example of a lead generation bot:
If you’re willing to automate your support to an extent, identify the primary questions, your customers would like answered based on your services. Add options for the customers to choose from to solve their issues.
Probably an essential part of designing a bot for booking appointments is a calendar integration into your chatbot that allows your visitor to choose a date of their preference.
Just like how you have FAQs on your webpage, similarly, you can have FAQs on a bot that provides rudimentary information to the consumers.
When you’re deploying a bot, you’re not just deploying a speaking machine on your site. You need to add flavour to bot conversations, and designing plays a prominent role in giving that edge to chatbots. Your bot design can ultimately help you position in the conversational marketing arena, allowing you to pave your way through the new dimensions of customer communication and improving customer engagement.