RICHARD’S NOTE: This was the very first blog post I wrote early on in my career. It was originally published in the UX Collective publication in Medium. It’s been several years since I wrote this (way before the release of ChatGPT on 30 Nov 2022) and I haven’t updated it at all since then. I’m posting this as the first post of my blog for posterity and as a sort of “time-capsule” from my early days as a designer.

This article is based on my talk last October 16, 2019 for the Philippine Web Designers Organization’s MiniFFC#56: Unconventional Design Meetup held at Proudcloud HQ in Quezon City, Philippines.

Hello, Internet! 👋

I’m Richard, a Conversation and User Experience Designer working at Senti AI, the pioneering AI company in the Philippines. I’m an active volunteer in local tech communities such as the Philippine Web Designers Organization and UX Philippines. I’m usually part of the many teams that move behind the scenes in the Philippine Tech Community. Now, it’s my turn to share some insightful content for you! This was my first talk about UX that was tailored for professionals!I’ve been working as UX Designer for almost a year now. Working at an AI company, I was constantly exposed to tech that is augmented by Artificial Intelligence and Machine Learning. This environment lead me to explore areas where I could apply UX in AI such as Conversation Design.

What is Conversation Design? 🤯

According to Google’s Conversation Design Guidelines, Conversation Design is a design language based on human conversation. Being a design language, you can apply it anywhere you want. You can use Conversation Design to make your web or mobile application’s interfaces more conversational! ✨

In this post though, we’ll be focusing on applying Conversation Design on where it really shines right now - designing Chatbots!

I want to get started with Conversation Design. Where do I start?

Just like other specializations in the umbrella of UX, there are some things that you need before you get started.

1. A firm grasp on the UX Design Process.
   • In Conversation Design, it’s very important that you have a firm grasp on the UX Design Process because you will constantly be going through that process as you design conversations. You need to be comfortable with doing User Interviews, Contextual Inquiry, Rapid Prototyping, Usability Testing and constant iteration.
   • If you don’t have a firm grasp on the design process but you want to start in Conversation Design, be prepared to exert double effort. Not taking into account the UX Design Process will result in a chatbot that does not create value for the user, even if conversing with the Chatbot feels natural for the user.
2. Decent Communication Skills
   • You need to have decent communication skills if you want to do Conversation Design. You don’t need to have written an article before or spoken in front of a crowd. All you need is to be able to carry a conversation with another person naturally, regardless of your language.
   • If you are not a good communicator, it will reflect on your Chatbot, and the experience won’t be good for the user. The secret to good Conversation Design is pretty much just sticking to how we normally communicate as humans, with all our quirks and nuances.
3. Familiarization with The Cooperative Principle (or Grice’s Maxims)
   • The Cooperative Principle (also referred to as Grice’s Maxims) are taken from the social sciences, specifically linguistics. The Cooperative Principle is a set of principles observed by humans when engaging in efficient communication. Considering these principles when doing Conversation Design is essential to creating conversational interfaces. This will be discussed later in the article.
   • The Principles, if not applied to your designs, would result in a conversation that feels robotic and forced.

Cool! How is The Cooperative Principle essential to Conversation Design? 👀

With The Cooperative Principle, Paul Grice was able to describe how humans communicate efficiently in normal, everyday situations. The overall concept of The Cooperative Principle is that when humans communicate, the participants will instinctively cooperate with one another. However, due to the quirks of human conversation (such as saying something that isn’t entirely truthful), these maxims should be looked at examples of ideal conversation.

1. Maxim of Quality
   • “We instinctively cooperate in terms of the truth of what we say.”
   • As humans, we naturally want to have a truthful conversation with whoever we’re talking to. The Maxim of Quality states that whenever we converse with another person, we should state things that are genuine and true.

   Correct Example:

   • Person A: “Hey Richard, have you finished the screens for the Account Creation Flow?”
   • Person B: “Not yet, but they should be ready 2 days from now.”

   Wrong Example:

   • Person A: “Hey Richard, have you finished the screens for the Account Creation Flow?”
   • Person B (lying): “Yup! I actually already sent it to you a week ago. It must’ve been sent to your spam inbox.”

In Conversation Design, we can apply the Maxim of Quality by making sure that our bot isn’t saying any wrong information. The most important thing to remember with the Maxim of Quality is that our chatbot should always be truthful, and should not be spreading false information.

2. Maxim of Quantity
   • “We instinctively cooperate in terms of the quantity of information that we provide.”
   • The Maxim of Quantity states that we cooperate based on how much information that we give and receive in a single exchange. In an ideal situation, we answer questions directly and avoid adding irrelevant information.

   Correct Example:

   • Person A: “What are you going to do tonight?”
   • Person B: “I’m just going to stay in and watch Netflix.”

   Wrong Example:

   • Person A: “What are you going to do tonight?”
   • Person B: “I have a dentist’s appointment later, but probably I’m going to watch Netflix tonight. Do you want to go with me?”

We can apply the Maxim of Quantity to our Chatbots by making sure that we respond to what the user asks, and not adding any fluff. Chatbots should communicate to the user directly and answer what the user asked. This is to better guide the user in the conversation and to reduce the chance of error.

3. Maxim of Relevance
   • “We instinctively cooperate in terms of the relevance of what we contribute.”
   • The simplest of all the Maxims, the Maxim of Relevance states that what we say must be relevant to the conversation.

   Correct Example:

   • Person A: “Have you seen my phone? I swear I placed it here earlier.”
   • Person B: “Yeah, you actually left it in the bathroom.”

   Wrong Example:

   • Person A: “Have you seen my phone? I swear I placed it here earlier.”
   • Person B: “I can’t find mine either.”

The Maxim of Relevance can be applied to our chatbots by making sure that what our chatbots say are relevant to the conversation and actually answer what the user asked.

4. Maxim of Manner
   • “We instinctively cooperate in terms of the way we strive to communicate clearly, without obscurity or ambiguity.”
   • Lastly, the Maxim of Manner states that we should communicate in an orderly fashion, where every detail is clear and nothing is confusing to the user.

   Correct Example:

   • Person A: “What were you up to before we met up today?”
   • Person B: “I decided to have a chill morning and drink some coffee first before going here.”

   Wrong Example:

   • Person A: “What were you up to before we met up today?”
   • Person B: “I got a cup of coffee at the coffee shop but I forgot to get my keys at the apartment so I had to go back there first and finish my coffee.”

Applying this to Chatbots can be as simple as making sure that what our Chatbots says makes sense logically. This is crucial in conversations where we have to outline steps to our users, such as how they can avail a promotion for a product or how they can open an account.

How is The Cooperative Principle applied in Chatbots? 🤔

Let’s dive into some sample conversational flows and see how we can apply the Cooperative Principle.

Example 1 - Phone Store Chatbot

Let’s say we have a chatbot for a physical store that sells phones. In this scenario, Richard is looking for a flagship android phone and decides to message their Facebook page.

If the Phone Store got an off-the-shelf chatbot without first considering the design of the conversations, this is one of the scenarios that can happen. Instead of picking up Richard’s request of flagship android phones on sale, the chatbot responded generically and repeated what Richard just said. Keeping in mind the Cooperative Principle, making users repeat what they just said doesn’t make the conversation flow naturally.

Improving upon the previous example, we can see here that the Chatbot responds in such a way that it acknowledges what Richard said. This conversational flow, compared to the previous example, sounds like what an actual salesperson in a phone store would say. The change in how the Chatbot treats the user as a participant in a conversation (compared to using canned responses) can make all the difference in engaging your users.

Example 2 - Chatty Customers

Let’s say that we have the same Phone Store Chatbot. This time, with a particularly chatty customer.

In this example, Richard added an unrelated follow up statement (or intents as they are called in Conversation Design) to what he previously told the bot. Normally, the default behavior of Chatbots are to state that they don’t understand the statement and would ask the user to repeat what they said in an effort to reset the conversation. With this conversational flow, it makes it seem like the Chatbot is either broken or not really equipped to converse with actual users.

Observing The Cooperative Principle and applying Conversation Design to the previous example, we can see here that the Chatbot is able to detect the change in statement or intent of the user. This small change improves the overall user experience greatly, with the Chatbot seemingly able to listen to an actual user. For the user, this would improve engagement and make them keep using the chatbot since they know that it provides value to them as a customer.

Always keep the Cooperative Principle in mind when you’re designing conversations. The four principles will help you in making your Chatbots more conversational, more useful, and a little more human.

Why Chatbots? 🤖

I chose to study on Conversation Design because Chatbots are starting to pop up everywhere in our daily lives. If you step back and assess all of the technology that you’re using now, you would be surprised how much we interact with Chatbots on an everyday basis. From our favorite digital wallets and payment apps, to Facebook Pages that instantly reply to your questions.

However, as technology becomes more ubiquitous, there is a tendency to maximize it by offloading repetitious tasks to machines, robots, or computers in general.

Conversation Design plays a huge role in developing chat and voice assistants because we as humans have a tendency to create responses that sometimes sound like robots. It’s a common pitfall for us to think sequentially and forget to consider the nuances of human conversation in our products. Little things that we take for granted such as suggesting to the user what they could do with the product gets lost as we build AI-powered Chat Assistants and, in general, pursue technology that is enabled by Artificial Intelligence.

As a Conversation Designer, your job would be to step back, assess the entire conversational flow, and ask yourself, “Is this how I would talk to another person?”.

Whew! That was a lot to take in! 😬 There definitely were a lot of terms, jargon, and overall a different style of designing products listed in the article. It may not be much, but what you’ve just read is enough for you to understand the role of Conversation Design in creating Chatbots and see how theory can be put into practice.

With that, you’re now ready to level up your Chatbots and create valuable AI-powered chat assistants that you can talk to like a human.

Before I finish the article though, there are a few things that I want you to remember as a Conversation Designer:

• Intelligent Assistants are everywhere.
  • Voice and Chat Assistants are everywhere now, and we can expect more assistants to pop up as ML technology improves. The more Chatbots get pushed to the market, the more important it is to consider Conversation Design. It is our job as Conversation Designers to help people accept Chatbots (and AI in general) and help them see how their lives can be improved with this technology.
• Too much tech is being introduced that creates more problems than it solves.
  • There was a Chatbot I used before that the company used as their primary customer support. It’s a great idea, for sure, but the minute I asked a question that (I assume) wasn’t part of the canned responses, the Chatbot just straight up didn’t respond to me anymore. This just frustrated me because I was really having problems with their product and I couldn’t use it normally. As Conversation Designers working on Voice and Chat Assistants, the responsibility is on us to make sure that the people who use our products have a delightful and valuable experience.
• As we innovate, we sometimes forget the human element.
  • This is especially true when we further push the envelope of technology through Artificial Intelligence and Machine Learning. We become so absorbed in creating the next big product or become the next big unicorn that we forget that ultimately, us humans are going to be using the very same products that we build. We should strive to make delightful experiences not only for us, but also for the people building the product, for your parents, for your friends, and for humanity as a whole. Because its not your computer that’s going to be affected when a product isn’t designed properly. It’s going to be you.

This was my first blog post ever! I hope you learned something as I did when I wrote this. Did you love it or hate it? Comments on the content and how to improve my writing are very much appreciated!