Last Updated on September 13, 2022
Traditional marketing strategies consist of content on websites, blogs, social media, and email. And there have been many transformations in that content over time – the text has given way to visuals and media; businesses focus on relationships and trust rather than on hard sells; savvy consumers what “social proof” of the value business brings to the table. Still, most of the traditional content is one-way communication.
Enter the world of chatbots and two-way conversations – a far more personal way for businesses to communicate with their audiences. You have already experienced this new technology if you have used any personal assistant via your smartphone or experienced in-home devices such as Alexa.
Now it’s time to move this technology into your overall marketing and customer relationship strategies, building a chat user interface (CUI) that will serve your users well, respond to their questions/needs and overall provide a personalized experience that leaves them satisfied.
Chatbots, AI/ ML and Why Chatbot Scripts Matter
Chatbots can be simple, which means they are pre-programmed to provide set answers to set questions. These are useful for such things as FAQ elements of a site. The bot picks up keywords from the user’s question and is then triggered to provide a set response.
Chatbots can also be complex, and this is where the better experience comes into play. Because they are embedded with AI and ML, these bots learn as they go, through natural language processing algorithms that allow them to converse with users at a more humanized level.
One of the most famous early chatbots was Poncho the Weather Cat. In addition to providing a daily weather forecast to his users, he had other dialogues that were hilarious and even provided horoscopes. His developers continually added features and engaging dialogue which served to increase his popularity.
The script was obviously critical to the popularity of Poncho, and it will be critical to any chatbot you create.
So, let’s look at how you create engaging and compelling chatbot scripts.
1. Study Popular Chatbot Scripts
There are lots of them out there – Use them and understand what makes them so engaging. They all have audiences, and their language and style are relatable to that audience. Identify your audience and be certain that you have a handle on that language and style before you begin to write chatbot scripts. A bot for Platinum jewelry, for example, will “sound” far different than one for a fast-food restaurant.
2. Identify Topic Fragments that You Want to Address
Beyond just a bot that is programmed to respond to frequently asked questions, what other topics for conversation can you identify that will promote dialogue between a consumer and the machine? For example, can your bot learn to make suggestions for upgrades or other products based upon what those customers are currently looking at or purchasing? Check out Tacobot, Taco Bell’s chatbot who takes orders through dialogue with the customer, and then suggests options and additional products. List your topics and the details of what you want to be included in your script on those topics.
Topic fragments also include things such as a conversation starter or humorous/witty elements. How will you greet your users? Again, this will depend upon the audience and what will “connect” with them. There’s a big difference between, “Hey there, what’s up?” to “Good day. How may I help you?” to “Hello human. I’m a smart chatbot who learns. Let’s see if I have learned the answers to your questions or issues. But first, how is your day going?”
3. String the Fragments Together
These fragments will eventually be strung together, and different paths will be needed based upon what the user says. For example, if the user says he is having a bad day, there should be a string of jokes, gifs, etc. with the message, “Maybe these will cheer you up.” If the user is having a great day, the bot may respond, “Good for you! Hope the rest of it goes well too. I have a headache but I can still help you.”
These multiple paths are called dialogue trees. And you should develop different paths (trees) for each potential response you may get from a user. Yes, this sounds horribly complex, and it will not be perfect the first time out. But there are tools to help you develop the trees and branches. You can try Twine, Xmind, Chat Mapper or Inklewriter for starters. Or you can use expert speech and scriptwriters from writing agencies, such as Trust My Paper or Studicus. Your other option is to draw your own treemaps on large sheets of paper if that sounds more appealing – it can actually be fun, if this is a team effort.
A Conversation Requires Two
You have two talkers involved in every conversation – the user and your bot. Obviously, you cannot control what your user will say or type, but you can try to predict it based upon past user behavior – questions they have asked, the terminology they have used, comments they have made, etc. You can, therefore, teach your bot lots of different possible phrases, put into your tree, or you can put capable AI behind your bot so that it teaches itself over the course of time.
You should begin with typical sentences they have used in the past. From there move onto words within those sentences and come up with a list of synonyms that they might use too. And don’t forget slang or dialects that they might use.
There is a final detail that you don’t want to leave out – user misspellings. For this, a great resource in Normalizer – it will help to decipher abbreviations, different spellings of the same words (e.g., American and British English), and correct misspellings automatically.
Your bot should have a personality that matches your brand. Remember the comparison between Platinum jewelry and Taco Bell? Remember Poncho the weather cat? When you write dialogue, you have to keep this in mind.
One of the issues you will face is users getting off topic. So, you will have to have responses that will not be critical or judgmental but that will bring the user back to the task at hand.
Some users like to “play” with chatbots, so expect this. A few of the most “famous” questions they may ask are such things as, “Where is the best place to bury a body?” or “Do you like Donald Trump?” Anticipate the playful questions they may ask, and add responses that will be humorous and yet bring the focus back to the real task. A great response might be, “You might want to check with a funeral home on that one, but I can give you options for sides on your food order. Would you like to take a look at them?”
Your Chatbot Scripts Must Sound Naturally Consistent
You definitely want your users to know that they are conversing with a bot, but that doesn’t mean your bot has to sound stilted and “all business.” It should communicate as a real person would in the “world” of the user persona, programmed to use their name if it is given – just a small piece that helps to personalize.
You can also give your bot a name, an age, a gender, outside interests, etc. This makes the entire UX and UI more enjoyable. The more personality you can give your bot, the better. And once you have that personality established, it must be consistent in language, style, and tone.
At Kommunicate, we have spent hours and hours over the persona of our chatbots such as Liz and Eve.
As you craft dialogues, watch for unnecessary words. These are needless distractions, so take out any words you can. It is generally believed that a chatbot’s message or response should be no longer than 60-90 characters. Aim for that.
If you have to provide a lot of information through your bot, at least put it in bubbles, so there is plenty of white space in between.
Lose the Need for Perfection (at First)
When you first launch a bot, expect the unexpected. Users will want to play; questions will come in that your bot cannot respond to because you have not thought of them. Remember this: a chatbot is an evolutionary tool – armed with AI, it will learn; armed with your constant monitoring, it will be improved.
Develop the best chatbot scripts you can, use the best available tools, find the best developer you can, and go from there.
Suggested Read: Chatbot framework, Tools and Type
This post is written by Diana Adjadj.
Diana Adjadj is a seasoned writer, both as a freelancer and as a copywriting expert for Best Essay Education and Grab My Essay. She has special expertise in speech and script writing. When not writing, she can be found showing her prize-winning Maltese “children.
Subscribe here to get the good stuff — we solemnly swear to deliver top of the line, out of the box and super beneficial content to you once a week.
At Kommunicate, we are envisioning a world-beating customer support solution to empower the new era of customer support. We would love to have you on board to have a first-hand experience of Kommunicate. You can signup here and start delighting your customers right away.