nline marketing has evened the playing ground for companies. Big or small, every company has access to prospects and one who has mastered the power of words leaves their competitors in the dust.
Naturally, mastering the power of words may not be everyone’s cup of tea. Hence we have writers and UX designers helping companies build engaging content. However, when it comes to designing conversations for chatbots even professional writers fail. You may wonder why?
Conversations do not have hard and fast rules. Thus, designing them takes insight, technical knowledge, strong language skills, and most of all, the intuition to predict human implications during the conversation.
However, no task is impossible. With no-code chatbot builders, you can build conversation flows easily once you have gained some insight. Keep certain tips in mind while designing the conversation flow.
This article can be treated as a guide to help you get started on designing conversation flows for your AI chatbot using a no-code bot builder. The article covers the following aspects:
1. Understanding the nuances of human conversation
2. Implementing the nuances of human conversation in chatbot conversation flows
3. What are the fundamental elements of conversation design?
4. Do’s and Don’ts of conversation designs
5. Things to keep in mind before script writing
Understanding the Nuances of Human Conversation
A good conversation has the power to make us feel energized, positive, and centered. This is one of the reasons why humans are addicted to texting, chatting, or calls. However, all of these apply to human-to-human conversations. Can conversations with AI bots make humans feel the same way?
The answer to the above questions directly depends on how natural the flow of conversation is between the bot and the human. Natural conversation flows between a bot and a human are ones that mimic a human-to-human conversation. To strike perfection in building a natural conversation flow there are 3 things one must keep in mind.
1. Sensitiveness to Underlying Cooperation
Every human conversation works based on some common knowledge, level of politeness, and implications.
Human 1 -” Hi! Do you know who is going to the party? ”
Human 2- “ Yes.”
This conversation feels unnatural and rude because the nature of the conversation is uncooperative. Although, the question asked is a yes and no type. Humans understand that the individual wishes to know the names of the people who are attending the party and is not expecting just a “yes” as a reply.
2. Encouraging Equal Participation
One sign of a good conversation that engages both individuals is turn-taking. Conversations should be back and forth or else it would turn into a monologue. Naturally, if an individual keeps giving out information when asked a single question, the conversation loses its relevance and value. Thus, conversations need to have equal participation from both parties. It guarantees attentiveness and engagement.
3. Context and Relevance
Context plays an important role in the type of conversation humans have. A context consists of the immediate environment ( are they in office, school, club?), the relation shared by the two individuals, and of course, the topic of conversation.
Once you consider all the points, the conversation will be natural and appropriate. Relevance is directly linked to the information being passed on during the conversation.
Human 1 [whispers] - “ Do you have a bookmark?“
Human 2 [loudly] - “ Yes! I also have a pen and a pencil.”
This conversation will lead to awkwardness and simply breaks all the underlying understanding of context and relevance.
Implementing the Nuances of Human Conversation in Chatbot Conversation Flows
To mimic the same level of genuineness and ease in the conversation between a bot and a human, one must design the conversation flow such that it follows the nuances of a human conversation.
Note: Even though the best bot conversations are ones that follow the unsaid rules of human-to-human conversation, it is important to be transparent with users that their assistant is a bot (not a live agent).
What are the fundamental elements of conversation design?
While designing a conversation flow, certain elements are a must-include in any conversation. Listed below are the elements and why each of them must be added to your chatbot conversation.
Salutations or greetings are the best way to introduce your bot, explain its function, and grab the opportunity to engage prospects for lead generation.
Avoid leaving conversations open-ended. Instead, add a closing statement. This adds maturity and a conclusion to the conversation.
Your closing statement may look like this: “ Thank you for placing your order. It will be delivered to you in 2-3 business days. To track your order use link ____.”
You can close multiple conversation paths effectively using Konverse’s jump feature. The jump feature allows you to maintain the conversation loop. You can use this feature to take the user from one conversation path to another while not leaving it open-ended.
Suppose the user-bot conversation comes to an end, the bot can ask - ‘Would you like to go to the main menu?’
If the user opts for yes, another part of the conversation will be triggered with the help of the jump feature. And opting for no will indicate closing the conversation.
Conversation flows must contain questions. It may be in the form of a prompt, or to gain information. Questions are a great way to encourage turn-taking and keep the bot in check. Without user feedback on the question asked, the bot cannot continue the conversation flow.
Acknowledging user input reduces confusion and makes the conversation feel easy.
Your acknowledgment statement may look like this: “ Alright! Got it. Thank you.”
Information statements should be well structured, clear, and to the point. The statement must be followed by a prompt to either keep the conversation going or for it to reach an end.
Bot- “ You can book an appointment with us on the following days. Would you like to know anything else?”
Suggestions are excellent at guiding users to discover new features, find the right content, or make the best decision for themselves which is within your power. Suggestions can also drive your prospects to interact with your content and create awareness.
Confirmation messages are important for users to ensure the bot understands their input. Repeating the vital information in your confirmation message makes the process convenient and provides assurance. You may provide a user with the provision to correct any information and reconfirm it if any changes were made.
Buttons or options allow your user to stay focused on the topic of conversation and not deviate. It also encourages quick replies which save time and effort for the user. With Konverse AI, you can use a variety of conversation nodes (Konverse in-built feature) for this purpose.
Adding audio-visual elements or emoticons in appropriate parts of the conversation flow makes your content-rich and captivating for users. Konverse allows the inclusion of videos, calling cards, images, and audio to enrich your conversation flow.
Do’s and Don’ts of Conversation Designs
There are certain things which you should include and some which may seem obvious but it’s best when avoided or not included. Listed below are some do’s and don’ts which individuals tend to miss out on. Let’s avoid making the same mistake!
Even though this is obvious, many times individuals forget to continue being conversational throughout the conversation thread. To make your bot seem conversational, ensure the conversation flow does not contain general or impersonal statements. The best way to go about this is to consider whether or not you would say the written statement in a conversation.
Be to the Point
Being to the point may feel more difficult than you think. While designing your conversation flow you may feel tempted to include extra information. The trick is to carefully design the conversation thread in a way that it feels short, sweet and guides the user to the right content. Cramping information into a single message can break the natural flow of the conversation.
Do Design Multiple Channels at Once
While writing the script for your chatbot, keep in mind that you would like to deploy the same bot in multiple channels such as WhatsApp business API, Instagram, Facebook, website, and so on. With Konverse AI, you can easily deploy and design no-code chatbots for multiple channels. Konverse also provides data and analytics on your most used channels and conversation flows.
Avoid Being Vague
Trying to use clear-cut statements leaves no space for doubts or confusion. Ambiguity or vagueness in statements or questions may lead to miscommunication. If your conversation flow leaves an Individual feeling more confused or left with more questions than before then the purpose of the bot will have been defeated.
Don’t Give Commands
Avoid giving commands through your AI chatbot to users. The beauty of AI technology is that people find it user-friendly. Thus, if you’re having to write down commands then you aren’t doing it right.
AI chatbots have entered the playing ground in the past decade. While they are equipped to solve many issues in a jiffy, like every technology, bots have shortcomings too. The best way to handle situations that the bot cannot handle is to provide alternative options. Apologies tend to highlight bot shortcomings. No user wishes for an apology in return for a request. Instead, adding prompts or alternative options for the user proves to be more effective than an apology.
Wrong response: Bot- “ Sorry, your account cannot be linked to your email address.”
Preferred response: Bot- “ Unfortunately, your account cannot be linked to your email address. Please try again or link to a different email address.”
This might be one situation where the bot fails to perform a task it is trained for. However, a second situation may arise when the bot encounters a situation it is not trained to solve.
For example: If a user asks an eCommerce bot about the weather conditions of a city.
Wrong response: Bot-” Sorry, I don’t know.”
Preferred response: Bot-'' I'm not sure. Would you like to talk to our agent ______[ agent name] to solve your query?”
With Konverse AI, you can choose to present the option of routing bot conversations to a live agent in case the user demands something that the bot cannot handle. This will allow you to resolve complex client issues quickly and seamlessly.
Don’t Forget Bot Persona
While forming bot conversation flows, one needs to keep in mind the personality or persona of the bot. You must follow through with the traits in the entire conversation. A break in character may leave a negative impact on the user experience.
Don’t Leave out Prompts
Prompts are a great way to present alternative or different options to the user. This allows a user to proceed with their preferred option quickly. Making decisions easy for users reinstates positive interaction and aid through the AI bot.
Bot: “ Your order has been placed. Would you like to continue shopping? “
Things to Keep in Mind Before Script-writing
To ace your script for the conversation flow, there is some prerequisite knowledge you must gain. Listed below are some of the topics and a step-by-step guide on how to build a bot. This will allow you to better understand the bot-building process and smoothen your script-writing process.
Study your Target Audience
Before you begin your script-writing process, consider the type of users you will be appealing to. Your language, use of words, information, and so on should be tailored to assist your target audience.
Type of Chatbots
There are mainly two types of AI chatbots. First, being a rule-based chatbot and the other being NLP-powered chatbots.
Rule-based AI Chatbots
These chatbots are easy to build, maintain, and a powerful tool for communication. These bots are recommended for small businesses or businesses with low budgets. Rule-based chatbots make use of conversation nodes/blocks to follow a strict path of conversation flow. Sudden user inputs may lead to the rerouting of bot conversation to live agents instead.
NLP-powered AI Chatbots
These chatbots are equipped to conduct mature conversations depending on the user input. NLP-powered chatbots are trained using FAQ training and entity creation. Although NLP-powered chatbots are considered more difficult to build and expensive to maintain, with Konverse the process is much easier.
NLP-powered chatbots are recommended for big businesses or businesses with ample budgets.
Step-by-step Bot-building Process
Using the steps provided below, you can easily build a bot on the Konverse platform.
- Step 1- Sign in to your Konverse account using your verified Email and Password.
- Step 2- Select the box that has your company name written on it.
- Step 3- Click on the Dashboard and opt for ‘Bot-builder’.
- Step 4- Select the ‘Story builder’ option and start designing the conversation flow. You can choose from a variety of conversation nodes (text, images, cards, quick replies, etc) as per desire.
- Step 5- After you have designed the conversation flow. Choose the ‘test bot’ option from the ‘Bot builder’ drop-down list. This option allows you to try out the AI chatbot before its implementation!
Note: Ensure you save or update your conversation flow each time before testing the bot. To save your conversation flow, click on the ‘save’ icon placed on the right top corner of your screen on the ‘story builder' page.
Want to kickstart your bot-building process?
With Konverse AI as your bot-building platform, you can start designing your no-code chatbot in a jiffy. You may opt for a demo, or use our resources to stay up to date on new and upcoming trends in AI for marketing.