Conversational Design Series: Designing Conversational Interactions & Planning a Flow Diagram

calendar Updated October 25, 2024
John Mowat
Manager, Conversational Solutions
Conversational Design Series: Designing Conversational Interactions & Planning a Flow Diagram

Hello and welcome! Here at Master of Code Global, as a Conversational AI company, we’ve decided to share our approach to conversational design in this series of articles. Please enjoy and let us know your thoughts!

This series shares our approach to designing conversational experiences. The first part delves into the theory behind key best practices, followed by an in-depth look at structuring a conversational design flow chart.

The objective of building a conversational interface is the same as any other UI you might implement: to help your users accomplish their tasks easily and efficiently. Whether a searchable FAQ or pictures of retail goods, a UI lets users interact with your data to reach a successful outcome. Conversational AI is unique in doing so with a more personable, human-like touch – but it still needs to be efficient to be successful.

So how do you build an efficient Conversational AI solution? With a well-considered design, which involves more than charting a few boxes and arrows to arrive at a satisfied user. We’ll look at 4 of MOC’s best practices for designing the flow of your bot’s conversation with the user, then in second part of this article we’ll actually map out a flowchart.

Note that all of this is distinct from actually writing dialogue which is a topic we’ll investigate in subsequent articles. If we were building a house, this is where we consider a floorplan, not where we decide how to decorate the walls.

Conversation design (CxD)

So let’s take a look at how to design a conversation while preventing it from getting side-tracked by unexpected situations!

Bot Conversation Flow: Keep Moving Forward with the Call to Action

One of the most important concepts in conversational design is to always (kindly) push the user on to the next step of their journey. Remember, they are interacting with the system to accomplish a task and reach a successful outcome, but they may not know how to get there on their own.

Consider the two greetings at right:The first image asks the user a question, it’s a “Call to Action” or CTA. Ending each of the bot’s turns with a CTA ensures that the conversation never stalls.

Meanwhile, the prompt on the right ends with one of the worst scenarios in all of Communications: dead air.

What is conversation design

Nothing about the greeting on the right guides the user to their next step. In fact, dead air is so disastrous to a successful interaction that in one of the most demanding forms of real-time conversation of all – the world of improv comedy – there is a whole theory around managing such moments, called Blocking.

Key idea: Don’t block your users, always step forward. End each of your bot’s turns with a question, prompt or other Call to Action.

Focus the User’s Response

If you review the screenshot in the previous section, you may notice that even the greeting on the left opens the bot up to some unexpected responses. While we provided menu buttons, the greeting itself was quite open-ended – some users may miss the buttons on the left and type a response to the question which was asked: “How can I help you today?”

This question invites a wide range of responses, many of which are likely outside the scope of your bot. To reduce the amount of unexpected input that users send your bot, focus the range of responses they can easily give. Consider the difference between these two dialogues:
Principles of Conversational Design
In the example on the right, we’re asking the user such an open-ended question that they are certain to reply in ways the system wasn’t built to handle. In the example on the left, we constrain their options – of course they may still type something unexpected, but at least we’ve increased the chances they reply with one of the options presented (even if they type it instead of pressing a button.)

This technique is actually the positive realization of a well-documented fallacy called the false dilemma. Happily, in conversational design we’re able to apply it for the good of the user, to help them achieve their goal.

Key idea: Focus the range of inputs a user is likely to give to promote responses your bot is designed to handle.

Circular Approach in Conversation Design Process

Suppose we’re designing a bot for a shoe store – consider these scenarios which deviate from the basic flow or happy path:

  1. User realizes they accidentally clicked on “loafers” instead of “sneakers”.
  2. User copy & pastes into the chat an image of the sneaker they want to buy.
  3. User says “actually I think I liked that other one you showed me” but the bot is unable to communicate with its NLU process to determine the intent of this utterance.

Are these all the same type of error? Do they each impact bot design in the same way?

No, the first scenario is different from the others, and recognizing the difference helps design a more effective flow:

  • In scenario #1, the user gave valid inputs and the system is functioning correctly, but they want to make a change. Proactively giving the user options to address this situation is a good practice we’ll discuss below.
  • Scenario #2 is a technical limitation as the bot isn’t able to interpret the image.
  • #3 is a system failure.

While the system needs to handle the types of technical issues in #2 and #3, it isn’t practical or advisable to provide options to resolve these in advance, so they don’t influence the design of the user experience in the same way.So back to scenario #1. The system is functioning correctly, the user simply made a mistake. Users often type and select options quickly on mobile or web interfaces, which increases the likelihood of errors.

We’ve simply arrived at a place where a user wishes to edit or redo the selection they made in the chat environment.

Conversation Design for Chatbots

Conversation Design for AI bots

This sort of scenario could occur even if the user were interacting with a human agent:“You know what? I actually kind of liked the loafers…let’s go back and look at them instead.”

This type of mismatch can occur at any time, and the further into the flow it is, the more frustrating it will be for the user if they have to start over from the beginning to get back on track. How can we prevent this?

Key idea: As often as possible – ideally at each step of the flow – apply a circular approach to give the user the option to back up a step and change their path or give a different response.

Accommodate Text-Only Inputs

One type of unexpected input can be converted into expected input during the design phase: be ready for users to type the text-label of buttons and other controls, rather than using the control itself.

For example, if you provide a button labelled “Continue”, then you need to ensure that wherever this button is shown, the system is also ready for the user to type the word “Continue”. Otherwise, when this unexpected utterance is sent to your NLU process, it likely won’t be able to determine the user’s intent or separate it from other occurrences of the same word throughout your system.

Similarly, if the bot shows pictures of shoes and asks the user to pick one, then it needs to be designed to handle a descriptive input typed by the user such as “the red one,” rather than them clicking the picture of the red shoe. In a face-to-face conversation, people routinely supplement their words with facial expressions, body language and tone. It’s just as essential in bot design to plan for the user to provide input in multiple ways. Some chatbot platforms support this idea of translating labels to text more readily than others, but it still remains a function which must be planned for in advance so the development team can configure it.

This scenario may seem trivial, but it’s often overlooked by designers. If you don’t plan for it, then your users will end up in frustrating situations where they gave simple, reasonable input, but the bot wasn’t able to accommodate it.

How to design a conversation for chatbot

Key idea: Always be ready for the user to type the name (or other characteristic) of buttons, graphics and controls, rather than using the control itself.

Designing for Success!

This article presented 4 of MOC’s best practices for designing conversation flow. These are lessons drawn from our own experience delivering successful solutions and using them will positively impact the design of other chatbots too:

  1. End each of your bot’s turns with a question, prompt or other Call to Action.
  2. Focus the range of inputs a user is likely to give, to promote responses your bot is designed to handle.
  3. Use a circular approach to give the user the option to back up a step and change their path or give a different response.
  4. Accommodate text-only input even when you’ve provided buttons, graphics and other controls.

The next goal is to present the design in the clearest and most efficient way possible so that stakeholders can focus on what’s important: understanding the the user experience flow. We have found the approach presented here leads to common ground with the client, educating them as to the flow of the interaction without overwhelming in unnecessary detail.

Here’s a simple conversational flow chart or map example – throughout the article, we’ll look at samples from a fictional high school which is going to offer a bot to assist their students and staff with commonly asked questions:

Conversational Flow Diagram

Basic Conversation Flow Chat Diagram Layout

It may seem trivial, but a legend helps viewers get oriented and focus on what matters: understanding the conversational flow.

Without a legend, readers may spend time puzzling over why one box is shaded and another one is clear. Even if it’s just a few extra seconds here and there, it’s a barrier to comprehension which can impact their overall reaction to the design.

The particulars of what shapes you use are of course entirely up to you (although we recommend following industry standards where applicable).

But for clarity and convenience, it is often helpful to embed specific UI choices, sample images and so on directly in the chart. For team members and stakeholders who aren’t as immersed in the project, being able to quickly scan over to see if a shape represents say a Quick Reply button or a “real button” is a big help.

Basic Diagram Layout

The basic organization is to arrange use cases in vertical flow. Why vertical? Because readers typically have widescreen-format displays, so a horizontal orientation provides more context and makes it easier to pan around when navigating and searching for other flows:

Conversation Design for Chatbots: Basic Diagram Layout
It’s often helpful to arrange flows outward from a central main menu or welcome message. From a usability perspective, this helps your reader stay oriented and avoids the suggestion of a left-to-right sequence of operations or a priority which doesn’t necessarily exist.

Grouping Conversational Flows

If your bot offers different features to different user types or personas, then cluster the flows for each user type separately. Suppose the school in our example plans to offer a bot which both students and teachers use, but only teachers can perform administrative actions. By grouping student flows on the left and teacher flows on the right, we make it easy for the reader to learn how to quickly traverse to an item of interest rather than having to hunt through hundreds of different shapes and various flows looking for it:

Conversation Design for Chatbots - Grouping Flows
Apart from your primary value-driven use cases, also include general-purpose flows such as small-talk and farewell paths that a user can arrive at from other flows, and group these together in another area of the conversation map. Similarly, group error and exception-handling flows in an area of the diagram where they won’t muddle the presentation of the intended user experience:

Chatbot Conversation Design

Labelling Flows and Blocks

The flow charts discussed here are a tool used to show how the requirements of a system will be met.

By the time the conversational map is being laid out, those system requirements should already be documented (as use cases or functional points) and clearly labelled with ID numbers or codes. For example, “Requirement S3: the student can ask the system what is available for lunch in the cafeteria.”

So for accuracy and ease of reference, when labelling each flow, include the ID of the requirement that flow implements. This allows the flow to be referenced from elsewhere in the diagram or during review with the team and stakeholders.

Furthermore, each user-facing or significant block in the diagram should then be given a sub-ID based on the flow it belongs to. For example, rather than having to say “in the 2nd box down from the top of flow 3…” it’s more concise and less error-prone to be able to say “in box 3.2…”.

Conversation Design for Chatbots

Even more importantly – while it’s not incorrect to represent every possible transition with a connector arrow, it begins to obscure the salient features of the user experience when 90% of boxes on the diagram have arrows back to the main menu. Including ID’s throughout affords a much more concise and manageable way to do this, called a “goto”:

Conversation Design Labelling Flows
Borrowed from the world of computer programming, a goto keeps things clean and efficient by “jumping” to a reusable item rather than introducing redundancy and noise by replicating it (or the connections to it) over and over again. This is made possible by including ID’s in the flow and block labels.
Regarding these ID labels in the diagram – if the system requirement IDs they are based on are guaranteed not to change, then simply reuse those IDs. But in practice, it’s usually safer to create new IDs for the diagram. When a business analyst changes “system requirement 4.3” to “4.4”, it’s easy to do a find and replace in a word processor or watch as numbered lists automatically update as elements are inserted and removed. But it’s hard to do that accurately in the flow chart – while some diagramming tools do offer a “find and replace” feature, depending on the numbering scheme used, there can be ripple effects where you are forced to manually update many dependent block IDs throughout the diagram.

At best, it’s unproductive to have to re-label the diagram every time requirements are updated; at worst, the flow breaks when this is missed and block IDs aren’t updated correctly, and this may not be caught until it manifests as illogical system behavior in QA or even deployment!

If possible, it’s convenient to hyperlink the use case or requirement from the flow.

Most modern toolsets allow clickable links and hotspots within the diagram, so take advantage of that and make it easy for everyone to get back to the requirements which drove the design choices in a flow, just by clicking on it.

Conversation Design Process

While NLP dialogue is typically captured in a separate artifact, it’s helpful to show sample dialogue which would bring a user to a given flow. This helps orient the reader to what the user is trying to accomplish and sets a good foundation for understanding why the flow is a good match for the intent. Under the testing stage, make sure you identify all issues and aware of all troubleshooting chatbot conversation flows examples.

Summary

It’s easy to drag shapes around in a diagramming tool, but an important part of Conversation Design is to afford stakeholders every opportunity to easily understand the design. The diagram will often be reviewed by a combination of clients, developers, and leadership. This means it’s likely that a given reader will be lacking familiarity with either Conversational AI solutions (in the case of leadership and subject matter experts) or the business domain (in the case of the development team.) The flow diagram is a bridge between these worlds, so it’s important to structure it in such a way that reduces friction, promotes shared understanding, and ultimately helps ensure the best decisions get made for the eventual system.

To that end, we looked above at Conversation Design best practices for basic diagram layout, the grouping of flows, and labeling flows and blocks for ease of reference. In the next part of this series, we’ll build out some flows for an example bot using the best practices described above and in part 1.

Need assistance with designing Conversational AI strategy for your customers? We can help!


















    By continuing, you're agreeing to the Master of Code
    Terms of Use and
    Privacy Policy and Google’s
    Terms and
    Privacy Policy




    Also Read

    All articles
    chatsimple