<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Jinny Seo on Medium]]></title>
        <description><![CDATA[Stories by Jinny Seo on Medium]]></description>
        <link>https://medium.com/@jinnyse0?source=rss-8f670cd51c8a------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*PJBQNXcLR9p9lHxkDBcEhA.png</url>
            <title>Stories by Jinny Seo on Medium</title>
            <link>https://medium.com/@jinnyse0?source=rss-8f670cd51c8a------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 03:02:17 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jinnyse0/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[11 More Best UX Practices for Building Chatbots]]></title>
            <link>https://medium.com/chat-bots/11-more-best-ux-practices-for-building-chatbots-67362d1104d9?source=rss-8f670cd51c8a------2</link>
            <guid isPermaLink="false">https://medium.com/p/67362d1104d9</guid>
            <category><![CDATA[chatbots]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[ux]]></category>
            <dc:creator><![CDATA[Jinny Seo]]></dc:creator>
            <pubDate>Mon, 30 Oct 2017 14:45:43 GMT</pubDate>
            <atom:updated>2018-05-07T15:53:18.675Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QBk8g8py1nb_Dhd_3vRPtw.jpeg" /></figure><h4>What we learned from building our own chatbot.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*vkeAZdTYbZ2Lal07DZKIqw.png" /></figure><p>Part 1: <a href="https://chatbotsmagazine.com/19-best-practices-for-building-chatbots-3c46274501b2"><em>19 Best UX Practices for Building Chatbots </em></a><br>Part 2: <em>11 More Best UX Practices for Building Chatbots </em>[You are here.]</p><p>Earlier this year, <a href="https://chatbotsmagazine.com/19-best-practices-for-building-chatbots-3c46274501b2">19 Best UX Practices for Building Chatbots</a> was published after guerilla testing a variety of Facebook Messenger chatbots. But you can only learn so much from testing other designs. More recently, in true Sabre Labs fashion, we built our own chatbot, Sabre Labsie. Here are top 11 best UX practices we learned.</p><h4>1. Your chatbot needs a voice.</h4><p>It’s a tired saying for people who have been in the chatbot space for any time, but creating a voice is easier said than done. Putting the advice to practice, we were surprised at how much more engaging Labsie became. Discussing how to develop a voice and a personality would take up another blog post, but we found the simple act of putting a name and a face to a chatbot was a huge help in finding a voice.</p><h4>2. Be conscientious of how the representation of your chatbot affects the user experience.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*seu5U7uhGy7OJG2TBYTnVA.png" /></figure><p>People respond to faces. We anthropomorphize things by nature, and chatbots are no exception. How users perceive your chatbot absolutely affects how they engage and interact with it; the profile image you choose for your chatbot plays a large part in shaping user perception.</p><p>The majority of current chatbots are represented by images of inanimate objects or abstract shapes like logos, microphones, or speech bubbles. While using inanimate objects to represent chatbots isn’t <em>wrong</em>, you should think about how the representation will affect the users.</p><p>We’ve found that the most engaging profile images are of faces. The image doesn’t have to be of a human face, but just something that a user can talk and ascribe a character to and <em>talk</em> to.</p><h4>3. Increase wait time with each additional message to give users time to read.</h4><p>One of most common and noticeable mistakes made in chatbots is loading messages without wait times or loading messages with wait times that are still too short. It’s incredibly distracting when messages are sent too quickly, as the new messages will move the text that the user is still reading.</p><p>The wait times should of course, depend on the length of the messages. What we didn’t know until we actually built Labsie was that the wait times also depend on the order of the messages.</p><p>When messages are sent right after one another, it feels more natural to pause a little longer and longer with each new message, even if they may be around the same length, so that users have time to finish reading older messages before newer messages are sent out.</p><p>To better illustrate wait times, below is a screenshot from <a href="https://www.pullstring.com/">PullString</a>, the conversational authoring tool and platform used to create Labsie. Unfortunately, PullString has moved focus to voice and <a href="https://www.pullstring.com/blog/pullstrings-voice-first-future">does not provide support for chatbots anymore</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2JlDD5MN_AT5tvGhUyf1pA.png" /><figcaption>Notice how the wait time increases for each additional message although some of the text lengths are similar.</figcaption></figure><p>Increasing wait time by a half to a full second more than the “base” wait time with each additional message felt “right” for Labsie. You could probably formally compute this based on message length and order, but Labsie was a small enough project that we manually tweaked the wait times between messages.</p><p>Originally, we wanted to have typing indicators pop up during wait times, but at the time of Labsie’s creation, that was not an easy thing to do. We eventually realized that while typing indicators are a nice touch, they’re basically a fancy wait mechanism to give users time to read.</p><p>No one really missed typing indicators as long as the wait times were “right”; users only complained when they felt rushed to read messages.</p><h4>4. Make sure buttons are functional at all points of the conversation.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*NoR00xIp9EGQQxcUQbNhiQ.png" /><figcaption>Buttons stay even after a user interacts with them.</figcaption></figure><p>If your chatbot employs <a href="https://developers.facebook.com/docs/messenger-platform/send-messages/template/button/">buttons</a>, users <em>will</em> scroll back and try to use the buttons again. This has been proven time and time again in various user tests.</p><p>If you’re not aware, Facebook provides quick replies and buttons as two main methods of user inputs in chatbots. <a href="https://developers.facebook.com/docs/messenger-platform/send-messages/quick-replies">Quick replies</a> disappear as soon as they are clicked; buttons do not.</p><p>While you shouldn’t expect everyone to scroll back to look for buttons, a good chunk of users will, and accordingly, you should keep those buttons functional.</p><p>Hopefully, all your buttons are context-independent and it makes sense to access them any time during conversation. If they’re not, there should still be an appropriate error response telling the user the button is not valid for use anymore.</p><h4>5. Learn when it is appropriate to employ quick replies or buttons.</h4><p>Continuing our thread of thought about buttons—it’s good to think about how to appropriately use UI elements in a conversation. These are not hard and fast rules, but here’s what I learned about the differences between these two UI elements:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pW3GSbrLrJWnTqfDFzDaOg.png" /><figcaption>Technically, they’re called button templates (left) and generic templates (right), but I call them both button menus.</figcaption></figure><p><strong>Buttons Menus</strong></p><ul><li>Do not disappear after usage</li><li>As such, they should be functional regardless of what state the conversation is in… i.e. these UIs are context-independent</li><li>Good for facilitating the navigation of overall bot architecture</li><li>Usually written as the chatbot speaking</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-N7rH4EN8SDyJ5vdnzZErg.jpeg" /><figcaption>Examples of quick replies</figcaption></figure><p><strong>Quick Replies</strong></p><ul><li>Disappear after usage</li><li>As such, use them when you don’t want users accessing these options again… i.e. these UIs are context-dependent</li><li>Good for facilitating the flow of conversation</li><li><em>Always</em> written as the user responding</li></ul><p>Just as different UI elements such as radio buttons and checkboxes serve distinct purposes on GUI platforms, buttons and quick replies each make sense in different contexts.</p><h4>6. Write quick replies from a first-person perspective—that is, as a user would respond.</h4><p>There are debates on <a href="https://medium.com/@jsaito/is-this-my-interface-or-yours-b09a7a795256">how things should be worded for traditional GUIs</a>, but most users will not notice whether a GUI is using the first-person or second-person perspective. This not the case with chatbots. Conscious thought should be given to how you word every element, especially quick replies.</p><p>Quick replies are user replies. They are inputs that speak for the user. As such, quick replies should be written from the perspective of the user.</p><p>When we first started building Labsie, we made the mistake of speaking from the third-person perspective to write quick replies. We quickly realized our mistake as soon as we actually talked to Labsie.</p><p>It was disorienting and unnatural responding to Labsie with a reply that said, “Learn about labs,” when the question was, “What would you like to talk about?” Instead, we shortened the reply to just, “About Labs,” which is presumably what someone might actually respond with in an actual conversation to the original question.</p><h4>7. Learn to say “<em>no”</em> tactfully, <em>for</em> the users’ sake.</h4><p>There are maybe two or three instances in Labsie which require text inputs from the user. The rest of the user inputs come from buttons or quick replies.</p><p>This was a conscious decision made to reduce branching and keep the users on the happy path. (If you’re wondering why, <a href="https://medium.com/@eunjinnyseo/hey-adam-thanks-so-much-for-your-reply-6a7af853b310">here are my thoughts on NLP and chatbots</a>.) This meant we had to figure out creative ways to word quick replies that meant “no,” without making the users feel rude.</p><p>Originally, quick replies to yes-or-no questions were simply written as, “Sure,” and “No thanks.” Continuing this pattern, we realized we felt bad for Labsie, since it was receiving all these terse replies.</p><p>In other words, users were being forced to feel rude due to our failure to craft polite quick replies, especially when the replies were negating or denying.</p><p>This is when I SOSed <a href="https://www.linkedin.com/in/tim-haynes-691890/">Tim</a>, our storyteller. (Pro tip: Find a wordsmith for your chatbot.) It was amazing how quickly he came up with brilliant ways to politely tell Labsie, “No.” Below is an example of the makeover:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eMcB_RaU6gELS_SpBZIO4w.png" /><figcaption>Another screenshot from PullString</figcaption></figure><h4>8. If you‘re asking for a text input after using quick replies, remind the user to type.</h4><p>An initial worry was that users wouldn’t know how to interact with quick replies, since they were a pretty novel UI element. It was a needless worry. No one struggled, even when one had never interacted with a chatbot before. What we didn’t realize, however, was how quickly users learned to <em>expect</em> quick replies.</p><p>If your chatbot employs quick replies but then suddenly requires a text input, be sure to indicate explicitly that the user now needs to type. The simplest way to do this is to say something like, “Please type your answer,” but I wonder if graphical solutions will emerge. Maybe the text field will be highlighted or the keyboard will be pulled up to indicate a text input.</p><p>If you don’t indicate the need for users to type, they will wait for a few seconds, waiting for quick replies that will never come… most people eventually figure it out (still, why make them anxious?), but some will think your chatbot is broken.</p><h4>9. Craft clever responses to cover all ground.</h4><p>Labsie is scripted from beginning to end. With enough work, it is possible to make scripted answers to quick replies into a seamless experience. Scripting believable and relevant answers to open text fields is a much greater design challenge.</p><p>Opening the floor for text inputs is basically an invitation for users to go off the happy path. At the very least, users will try to “trick” your bot. To deflect tricksters without making the branches balloon, Tim crafted more clever responses, as shown here.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/594/1*i-XGPOjfKAaGl4P6qbkFuA.png" /></figure><p>There is an art to crafting a proper, open-ended question and still crafting a finite set of responses. In the example above, the “user” responses are connected to sets of answers. “Red” gives the first answer; we used a list of dozens of the most common colors that all result in the second answer; any other string of text brings the user to the third answer.</p><p>It’s not perfect… someone could still “break” the seamless flow of the conversation by typing something like, “I hate suitcases.” But in that case — no matter how ridiculous the user input — the third answer still comes across as polite, upbeat, and maybe a little snarky and helps to fulfill the personality we’re trying to create with Labsie.<br> <br>Done properly, responses to open-ended questions can be great opportunities to help cement the personality of your chatbot. But it’s important to make sure that your “catch-all” answer works for a breadth of unexpected inputs.</p><h4>10. Prune your branches: For conversational interfaces, linear navigation is superior to nested navigation.</h4><p>Often times, conversational interfaces are explained through GUI analogies. (For more GUI to LUI (Linguistic User Interface) analogies, check out this <a href="https://medium.com/swlh/a-natural-language-user-interface-is-just-a-user-interface-4a6d898e9721">great article by Matthew Honniba</a>l.)</p><p>While these analogies can be helpful in thinking about the conversational medium,<em> conversational interfaces should not be direct translations of graphical user interfaces</em>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/1*L3JAZd3EB2RVXH-JYMaECw.png" /><figcaption>Credit: Matthew Honnibal, <a href="https://medium.com/swlh/a-natural-language-user-interface-is-just-a-user-interface-4a6d898e9721">https://medium.com/swlh/a-natural-language-user-interface-is-just-a-user-interface-4a6d898e9721</a></figcaption></figure><p>The GUI to LUI analogy may work for smaller, individual interactions, but on a more macro level, the analogy falls apart. A big mistake we made in building the first version of Labsie was that we attempted to convert a traditional GUI menu architecture into a conversational one.</p><p>Labsie had three main branches that the users could explore. The first, an “About Sabre Labs” branch, explained our team’s purpose; the second branch informed users on emerging tech trends in travel; and finally, the last branch quizzed users to “Discover Your Travel Spirit Animal!”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/786/1*BTzJif04Uk3tZc36wsp3VQ.png" /><figcaption>GUI equivalent of architecture we envisioned for Labsie version 1</figcaption></figure><p>The first two branches had subtopics under them. For example, in the emerging tech branch, we talked about conversational interfaces (e.g. chatbots), digital realities (e.g. virtual reality), and connected intelligence (e.g. indoor location and automation).</p><p>Instinctively, I envisioned a navigation bar with dropdowns. With buttons and quick replies, I thought I could easily translate this architecture into a conversational one. Unfortunately, this proved to be a suboptimal approach. Below is the user flow of Labsie version 1:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cgqsInt2S_CqeTfezRfg0g.png" /><figcaption>User flow for Labsie version 1</figcaption></figure><p>Notice all the arrows going back to “Select topic [menu],” highlighted yellow. This is happening because at then end of each branch <em>and</em> at the end of all the subtopics, Labsie has to provide a way for the users to return to the top of the conversation. Every few minutes, Labsie was asking users, “Would you like to talk about something else?” Talk about annoying.</p><p>Besides being annoying, structuring Labsie like a traditional nested menu was an unnatural way to use the conversational medium. Asking the users to choose which topic they wanted to talk about <em>twice in a row </em>was strange and disorienting.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZaSJQcFFofwTNlNSwt1SXQ.png" /><figcaption>Expectation vs. Reality: Realizing that we don’t navigate real-life conversations through menus.</figcaption></figure><p>We need to think of conversational interfaces as parallel to human conversations, not to GUIs. A real human doesn’t offer a menu of topics to initiate a conversation. Rather, the speaker (the chatbot) opens with a topic. The respondent (the user) then continues with the subject or changes it by responding appropriately.</p><p>From there, the conversation should branch organically. Since chatbots aren’t very good at organic conversation yet, we can control this branching in conversational interfaces by appropriately employing UI elements like quick replies and buttons.</p><p>These quick replies and buttons do not have to be exhaustive, either. Yes, conversations can branch infinitely, but the users don’t have to know about these infinite possibilities.</p><p>Often times, it’s better to put the users onto a happy path and offer ways veer off it, than to show the users <em>all</em> the functionalities up front.</p><p>After this realization, we eliminated submenus completely. We kept the top-level buttons to launch user onto one of the main three branches, but we stopped constantly asking if they wanted to loop back to the top of the conversation again, and only offered this “back” functionality at the end of a branch. The result was a conversation that flowed much more naturally.</p><p>If we visualize the flow of Labsie version 1 and version 2, the former version looks like a tree while the latter is more of a chain:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ymOdC5c3lYi3VGx2jM3mwQ.png" /><figcaption>Flow for “About Labs” and “Spirit Animal Quiz” sections were ommitted for clarity</figcaption></figure><p>What’s outlined above is not a strict rule. The purpose of your bot will be the deciding factor in what is an “optimal” flow. For example, task-oriented chatbots <em>should</em> have plenty of confirmations and back/undo functionalities. However, even then, having an overly nested architecture is still not recommended.</p><p>Designing conversational flows is an interesting space, and we’re looking forward to seeing how it will mature. The general trend seems to be that the more you prune your conversational branches, the easier it is for both your developers and your users.</p><h4>11. Provide a way to reengage the chatbot when your interaction is over.</h4><p>Finally, what happens when an interaction with a chatbot is over?</p><p>It could be as simple as ending your interactions with, “Say hi again if you need anything.” This last message will then be the jumping off point when the user re-engages with the chatbot.</p><p>As an added measure, we also added a single quick reply that said, “Restart,” at the end of Labsie’s goodbye. Whatever you do, make it easy for the user to reengage your bot when they come back.</p><p>Happy bot-building! And feel free to reach out to share your own best practices! We’re constantly looking for ways to improve.</p><p><em>This article is part of Sabre Labs team’s analysis on emerging technology trends impacting travel. Please reach out to Sabre Labs at </em><a href="/service/mailto:sabrelabs@sabre.com"><em>sabrelabs@sabre.com</em></a><em> for further inquiries or discussions on the future of technology and travel.</em></p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fupscri.be%2F58521e%3Fas_embed%3Dtrue&amp;dntp=1&amp;url=https%3A%2F%2Fupscri.be%2F58521e%2F&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=upscri" width="800" height="400" frameborder="0" scrolling="no"><a href="https://medium.com/media/c90883b0516ddbe5c59960c12a04bc64/href">https://medium.com/media/c90883b0516ddbe5c59960c12a04bc64/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=67362d1104d9" width="1" height="1" alt=""><hr><p><a href="https://medium.com/chat-bots/11-more-best-ux-practices-for-building-chatbots-67362d1104d9">11 More Best UX Practices for Building Chatbots</a> was originally published in <a href="https://medium.com/chat-bots">Chatbots Magazine</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Illustrating the Future of Travel]]></title>
            <link>https://jinnyse0.medium.com/illustrating-the-future-of-travel-2304c653153e?source=rss-8f670cd51c8a------2</link>
            <guid isPermaLink="false">https://medium.com/p/2304c653153e</guid>
            <category><![CDATA[illustration]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[apple-pencil]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[design]]></category>
            <dc:creator><![CDATA[Jinny Seo]]></dc:creator>
            <pubDate>Wed, 19 Jul 2017 17:47:27 GMT</pubDate>
            <atom:updated>2023-07-17T21:23:23.147Z</atom:updated>
            <content:encoded><![CDATA[<h4>A deep dive into my process</h4><p>For the past couple of months, I have been illustrating and designing Sabre Labs’ 2017 Radar Report. The report, published in June, covers nine tech trends, each of which was illustrated using the 12.9-inch iPad Pro, the Apple Pencil, and Adobe Sketch. Here’s a closer look at the process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DHxvD8PJdAscJq_ZG75WSg.png" /><figcaption>The report is available to download <a href="https://www.sabre.com/insights/labs/">here</a>.</figcaption></figure><h3>The Tools</h3><h4>On Adobe Sketch</h4><p>While not the most robust sketching app available on the iPad, <a href="http://www.adobe.com/products/sketch.html">Adobe Photoshop Sketch</a> and its integration into Creative Cloud was essential in my workflow, and I always pulled illustrations into Photoshop for final edits and touch-ups. The layers translate from Adobe Sketch to Photoshop, although not the other way around, unfortunately.</p><p>All the images created with Adobe Sketch are of course, raster files. (If you want vectors, use <a href="http://www.adobe.com/products/draw.html">Adobe Illustrator Draw</a>.) To preserve the image quality, I’d often draw an object very large in Adobe Sketch, upload the psd onto Creative Cloud, then rescale in Photoshop.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wHcygaENeADvxiFSNrxpcQ.png" /></figure><h4>On the Apple Pencil</h4><p>The Apple Pencil was a pleasure to use. The pencil is sensitive and intuitive, although my hand would cramp sometimes from trying to reach maximum pressure on the pencil. Charging the pencil is still awkward and I’m still terrified that I will snap off the metal every time I plug the pencil in, but overall, it is only a minor inconvenience, and I was impressed by the quickness. My biggest complaint is the inability to know the battery level of the pencil. Even if there is a way to find out, I never learned how, even after months of using the app. These drawbacks are only minor hiccups in the workflow however, and I’ll definitely be using the pencil for my future creative endeavors.</p><h3>The Process</h3><p>Creating materials for Sabre Labs is always interesting—we need to represent Sabre as a company but also retain our identity as <em>Sabre Labs</em>, the team of innovation and emerging techs.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-H_Wiky_rp-ydUCo_3Q7EQ.png" /></figure><p>Effectively leveraging Sabre’s company color, red, is integral to achieving this brand cohesion. Once that’s solid, I can then push other boundaries to represent Sabre Labs’ creative and innovative spirit without straying too far from Sabre’s company identity as a whole. Red can be a challenge though; although attention-grabbing and powerful, when used in excess, it quickly overpowers other elements.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/342/1*W3UWUwAzr0Sn8V8RZUapaQ.png" /></figure><p>Working at Sabre, I’ve quickly learned to use red only as an accent color. So starting out, one of my first challenges was to figure out a solid color palette beyond “Sabre Red”. Looking into our brand guidelines, I landed on a calming mint blue, which meshes well with not only red, but with a variety of other colors that are trendy today. This blue became the backdrop of my first illustration, blockchain, and of all the other illustrations to come.</p><h4>Style, Color, Vision, and Blockchain</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QOF4U3jOWC4Z4KMCcUPcqQ.png" /><figcaption>Initial sketches for blockchain</figcaption></figure><p>As the first illustration, blockchain was the most time-consuming but also the most fruitful, as it set the tone for the rest of the illustrations. It was also a technology that was more conceptual in nature, allowing me to play with different styles of illustrations to best find the one that fit our vision as Labs.</p><p>At Sabre Labs, we envision future that is promising and realistic, yet not dystopian or utopian. It was important that the illustrations embodied our attitude. Fleshing out the illustration for blockchain, three more words emerged that further defined our vision for the Radar Report: whimsical, human, and approachable.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1YWuHu2yrk9-2IyPA3hBxg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rJLk_GvR-pdsze3o7fAsAg.png" /><figcaption>Final sketch for blockchain (left); My notes on Sabre Labs’ brand vision (right)</figcaption></figure><p>Decided on the sketch, I came back to the colors. Based on the previously selected blue and of course, red, I dropped in other colors to the sketch. By the time the illustration was complete, I had a solid color palette and a style to lead me for the next eight illustrations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aiKZ_6RqozjJjRa8qgvK4Q.png" /></figure><h4>A Closer Look at the Tools</h4><p>There is a variety of tools in Adobe Sketch, but I only really used the color pencil tool.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*B7ywoG7iwhnIV1RWL8ARyw.png" /></figure><p>When the Apple Pencil is tilted, the color pencil tool produces a grainy texture that added visual interest and cohesion across the illustrations. I also found the tool to be the most efficient in blending colors. I can blend by simply tilting and varying the pressure on my Apple Pencil—no need to fuss with flow settings, and there are no streaks you tend to get when you use a traditional Photoshop brush.</p><p>Layer blend modes are available on Adobe Sketch, and I use them mostly when I want to increase contrast without going into Photoshop. Uploading psd’s onto Creative Cloud can be tedious and breaks the flow, so I avoid switching platforms unless absolutely needed. My go-to blend modes to heighten shadows on objects are <em>multiply</em> and <em>luminosity.</em></p><p>The ruler was essential in keeping edges sharp and angles consistent. I opted for an isometric perspective whenever possible in the illustrations to add to their whimsy. The hexagon ruler was especially useful in drawing isometrically.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gKtP_Sq6-6Otc_1ESA_Sow.png" /><figcaption>Keeping my angles consistent</figcaption></figure><h3>The Illustrations</h3><p>Finally, below are the nine illustrations in the order as I completed them.</p><h4>Trend 1: Blockchain</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3tOSvqfDV8seLWzXiwHcFQ.png" /><figcaption>Blockchain, 2017.</figcaption></figure><p>My first illustration and my favorite. The concept is quite literal—depicted are technologies, built on chain of blocks… I had a lot of fun drawing out the use cases of blockchain. (To find out what these use cases are, <a href="https://www.sabre.com/labs/radarreport/">read our report</a>!)</p><h4>Trend 2: Space Tourism</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7-RjZwfH9KZb1iz3UB4Piw.png" /><figcaption>Space Tourism, 2017</figcaption></figure><p>I was looking at a lot of retrofuturism as I began this project. Although I wasn’t consistent about it, you can find vintage influences throughout the illustrations. Here, it’s a vintage postcard from Mars.</p><h4>Trend 3: Quantum Computing</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*immGe0s88w3GBQVg1gGLpg.png" /><figcaption>Quantum Computing, 2017</figcaption></figure><p>Another one I had a lot of fun with — featured here is the Apple Classic computer with a qubit diagram on the screen, a “vintage” plant behind the computer, except in true millennial fashion, it’s a succulent.</p><p>And of course, if we’re going to talk about quantum physics, we can’t forget Schrodinger’s cat.</p><h4>Trend 4: Autonomous Delivery</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Gw_svTM1DHTQPPbCTiA-QA.png" /><figcaption>Autonomous Delivery, 2017</figcaption></figure><p>Admittedly, the first thought was to draw drones, but I didn’t want to be so literal. The reasons are twofold: 1) Drones are ugly, and 2) the image would become outdated too quickly.</p><p>Initially, the idea was a walking pizza delivering itself, but we decided autonomously delivered medicine was a more important thing to focus on.</p><h4>Trend 5: Neural Interfaces</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jpPE3exhBD7336l35iWlMg.png" /><figcaption>Neural Interfaces, 2017</figcaption></figure><p>Brains are fun to draw! We may implant chips in our brains one day!</p><h4>Trend 6: VR Gets Physical</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8ftt8BuMAgEXi9QMDqo5TA.png" /><figcaption>VR Gets Physical, 2017</figcaption></figure><p>This one was painful to get going. I really didn’t want to illustrate a guy looking awestruck in a headset for the hundredth time, but who knew coming up with a different idea would be so difficult?</p><p>To make things trickier, we weren’t just talking about VR but about how VR could map to the physical world. Still unsure on the clarity of this image, but visually, I do like it.</p><h4>Trend 7: Artificial Intelligence</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*g5UTwv9fQ8nFSjTpicyHug.png" /><figcaption>Artificial Intelligence, 2017</figcaption></figure><p>Starting this illustration, there were a lot of opinions on what <em>not</em> to draw. We didn’t want any scary-looking android heads or implications that AI was going to take over the world. The final idea came about after brainstorming with the team — we believe in the future of human/AI collaboration, so instead of <em>just</em> the machine, depicted are the human and the machine working <em>together</em>.</p><p>Drawing the little details in the background as a nod to the importance of machine learning in AI was fun :)</p><h4>Trend 8: Augmented Reality</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7l0vsiX0dMuroeMEZpGEoQ.png" /><figcaption>Augmented Reality, 2017</figcaption></figure><p>Another tough one conceptually. We wanted to emphasize scanning and surface detection used in AR.</p><h4>Trend 9: Trusted Presence</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IRDs3QQem9FPNjzRqdxcsw.png" /><figcaption>Trusted Presence, 2017</figcaption></figure><p>This was the last illustration, and it shows. But I do love drawing eyeballs.</p><h3>The Final Report</h3><p>I hope this walkthrough has piqued your curiosity in our <a href="https://www.sabre.com/labs/radarreport/">2017 Radar Report</a>. It’s available to download here at <a href="https://www.sabre.com/insights/labs/">sabrelabs.com</a>.</p><p>For further inquiries or discussions on the future of technology and travel, please reach out to Sabre Labs at <a href="/service/mailto:sabrelabs@sabre.com">sabrelabs@sabre.com</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2304c653153e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[19 Best UX Practices for Building Chatbots]]></title>
            <link>https://medium.com/chat-bots/19-best-practices-for-building-chatbots-3c46274501b2?source=rss-8f670cd51c8a------2</link>
            <guid isPermaLink="false">https://medium.com/p/3c46274501b2</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[chatbots]]></category>
            <category><![CDATA[guide]]></category>
            <dc:creator><![CDATA[Jinny Seo]]></dc:creator>
            <pubDate>Fri, 20 Jan 2017 21:59:47 GMT</pubDate>
            <atom:updated>2020-06-02T19:57:27.746Z</atom:updated>
            <content:encoded><![CDATA[<h4>How do we make chatbots usable? What we learned from testing various chatbots.</h4><p>Part 1: <em>19 Best UX Practices for Building Chatbots</em> [You are here.]<br>Part 2: <a href="https://medium.com/@eunjinnyseo/11-more-best-ux-practices-for-building-chatbots-67362d1104d9"><em>11 More Best UX Practices for Building Chatbot</em></a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mIysC1UCu0e-AfrOkBs74Q.png" /><figcaption>Image from <a href="https://unsplash.com/photos/xYjWeyS235w">Rawpixel</a></figcaption></figure><p>At Sabre Labs, we explore how exciting, up-and-coming technologies impact travel. Lately, one of these technologies has been <a href="https://chatbotsmagazine.com/conversational-interfaces-beyond-the-hype-457c18290644">conversational interfaces</a>. In 2016, we jumped on the chatbot bandwagon with enthusiasm.</p><p>Through our research and prototyping, it quickly became clear: Best practices for good UX in conversational interfaces are still yet to established.</p><p><a href="https://chatbotsmagazine.com/the-bot-playbook-7bb6d181a6a9">Usability practices</a> for graphical interfaces have been long standardized; anyone looking to improve their websites are pointed to trusted guides like Steve Krug’s <a href="https://www.amazon.com/Dont-Make-Me-Think-Usability/dp/0321344758"><em>Don’t Make Me Think</em></a><em> </em>containing well-established usability practices. Unfortunately, for chatbots, these guidelines are only sometimes applicable, and often irrelevant.</p><p>In hopes of establishing some usability guidelines for chatbots, I conducted guerilla user tests at a local university in Dallas, asking a total of 15 users to interact with one of three travel-related Facebook Messenger chatbots: Skyscanner, Hipmunk, and Kayak.</p><p>For each test, the users were asked to to 1) book a flight, 2) change the destination, and 3) change the departure and/or return dates.</p><p>Below are the key findings, observations, and themes that came out of the guerilla testing.</p><h4>1. Strive for consistency: don’t create expectations that the chatbot can’t meet.</h4><p>The number one rule of <a href="https://faculty.washington.edu/jtenenbg/courses/360/f04/sessions/schneidermanGoldenRules.html">Schneiderman’s “Eight Golden Rules of Interface Design”</a> still applies to chatbots: Strive for consistency.</p><p>Seeing that the Skyscanner chatbot accepted weekdays as inputs, a user typed in “Wednesday” as his departure date, which was successfully parsed to the correct date.</p><p>Emboldened by his previous success, he then typed in “Monday” when prompted for the return date, only to be met with an error response. He was bewildered by this inconsistency — why was his return date invalid, when it followed the same format as his departure date?</p><p>He would’ve had a better experience if he had never been given this expectation. Rather than attempting to give an illusion of high-level NLP (natural language processing) and then failing, you should set correct expectations that match your chatbot’s abilities.</p><p><em>Functionality trumps all. If the chatbot can only partially parse certain formats, forgo the format altogether, and set expectations and provide examples of formats that your chatbot can process.</em></p><h4>2. Strive for consistency: provide a way back to the “main menu” somehow.</h4><p>Many users would complete a task effortlessly, and then ask, “So… where do I go from here?”</p><p>In traditional GUIs, it has become second nature for us to look for the home button whenever we want a fresh start. For conversational interfaces, the solution is still forming.</p><p>Is it simple as just training the user to say “help” or “start over?” Will this prompt be part of the onboarding process, or at the end of an interaction? Or will the solution not be conversational at all, but simply another button on the interface? Kia Niro took an interesting approach to the onboarding process by using the carousel to explain how to use a Facebook Messenger chatbot.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1ts-Ex1h-TGZrGbnzUAY0g.gif" /></figure><p><em>Regardless of the approach, the user should know how to get to the main menu and how to reset the conversation at any time during the interaction.</em></p><h4>3. (Outdated) Until it becomes an established practice, don’t rely on Facebook’s persistent menu.*</h4><p>Facebook preemptively formulated a solution to the above problem: the <a href="https://developers.facebook.com/docs/messenger-platform/thread-settings/persistent-menu">persistent menu</a>.</p><p>The persistent menu, in theory, is everything a lost user needs: it’s always there, it communicates the basic capabilities of your bot, but it’s also behind a “3-caret icon on the left of the composer,” which is just a fancy way to describe a hamburger menu.</p><p>While following <a href="https://developers.facebook.com/docs/messenger-platform/guidelines">platform guidelines</a> is critical to create consistent experiences, I can’t advocate for the use of Facebook’s persistent menus… because no one clicks on them.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DsmCsRnEffwa2Y59FIt5Vw.png" /><figcaption>Can <strong><em>you</em></strong> find the persistent menu?</figcaption></figure><p>Out of 20 users we tested, only one even thought to click on this hamburger menu. The information icon on the upper-right corner got more attention than the hamburger menu ever did. Three people hovered over the helpful looking “i” on the upper right-hand corner when they got lost, but the hamburger menu never even registered to these observant users.</p><p>Anti-hamburger menu-ers, here’s example #1001 of hamburger menus failing.</p><p>*Facebook has since updated this functionality, but this is still a good example of why you shouldn’t hide critical functionalities.</p><h4>4. Don’t make the users go tl;dr.</h4><p>In its chatty excitement, the Hipmunk chatbot hits the users with four opening messages, totaling nearly 350 characters, which most users scrolled right past.</p><p>In general, anything more than three lines of text seemed to activate the <a href="http://www.urbandictionary.com/define.php?term=tl%3Bdr">tl;dr (too long; didn’t read)</a> response in users.</p><p>How long your message will seem will depend on the screen size of the device, but a good rule of thumb is to have no more than 90 characters per message (around three lines on mobile).</p><p><em>Having to scroll up to see content is an instant trigger for tl;dr (too long; didn’t read). Try to keep your message to less than 63 characters (around two lines). Multiple messages shouldn’t add up to more than ~140 characters at a time.</em></p><h4>5. Test your chatbot on different devices to make sure users are not scrolling to read.</h4><p>Hipmunk sent an image in addition to its lengthy intro, making users scroll up to read its messages and all the buttons it offered.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_ZpdNBhGbl7z7FmrKy-Y9w.png" /><figcaption>Messages can look vastly different depending on device sizes and mode of access.</figcaption></figure><p>Test your chatbots on different devices. One of the widely-touted “pros” of chatbots is that the frontend does not need to be designed — but they are still being displayed on a screen. Testing the functionalities across various screen sizes is still very necessary, and a quick way to gain usability points.</p><p>NOTE: Keep in mind as well that users will have their keyboards pulled up on mobile. So realistically, you only have half the phone screen as your playground.</p><p><em>Don’t send multiple messages that will not fit in the display of a mobile phone. This makes the users have to scroll up to see content and is a surefire way to make them glaze over.</em></p><h4>6. Use UI components as “guard rails” in a conversation.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kNre7kWHfDg5lmTiF03vjw.png" /></figure><p>The beauty of conversational interfaces is its flexibility. The paths that the user could theoretically take are virtually infinite. Unfortunately, natural language processing is currently nowhere near being able to handle this.</p><p>Buttons and quick replies are almost integral for keeping users on the happy path and away from too many apologetic error messages.</p><p>They will also help to prevent users from encountering the dreaded “blinking cursor” scenario, where users are faced with an empty text field and are unsure of how to proceed.</p><p>Watch out though, these guard rails can limit our users if we are not careful…</p><h4>7. WARNING: Buttons and quick replies give an illusion of limited choices.</h4><p>For better or worse, when users are presented with UI controls, they will not think to do anything other than to interact with the given controls.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*I7l5AZDIWsDOCt6_pL5_lw.png" /><figcaption>Expedia–good; Skyscanner–bad.</figcaption></figure><p>This was a big failure point for Skyscanner. When users were asked to change their departure date to one that fell outside the two button options, they were simply… stuck.</p><p>When asked why they had not typed in the date instead, most users replied that they had not thought to do so. Others replied that they weren’t sure if the chatbot would be able to process their text input.</p><p>Most users will no longer perceive sending a message as a viable action when a chatbot offers buttons or quick replies as a way to respond.</p><p>This is a natural response to being given a set of choices that is probably exacerbated by the fact that chatbots are a new technology. Users are mistrustful of new technologies, and this mistrust translates into safe choices.</p><blockquote>Users are mistrustful of new technologies, and this mistrust translates into safe choices.</blockquote><p>Thus, UI components in chat is a double-edged sword. They are great for keeping users on a happy path, easing them into a new technology, and easing the mental and physical load during interactions. However, they can also easily trap your user by creating an illusion of limited choices.</p><p><em>If your chatbot utilizes buttons or quick replies, include an option to nudge users into typing a response if needed, as Expedia did.</em></p><h4>8. Make interactions predictable.</h4><p><a href="https://developers.facebook.com/blog/post/2016/09/12/new-messenger-features-payments-ads-enhanced-mobile-websites/">Payments in Facebook Messenger</a> is still in beta, so most chatbots involving shopping eventually redirect users to their respective websites.</p><p>One of the biggest problems with these redirects was that it was impossible to distinguish which buttons would result in a redirect and which buttons would continue the conversation in Messenger.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EkklGefN61zwtsQFTMIAIg.png" /><figcaption>Out of these five buttons (“Show Me,” “Clear preferences,” “See flight details,” “Create a fare alert,” “Search hotels”) can you guess which will redirect to Hipmunk’s website?</figcaption></figure><p>One could argue that the blame is on Facebook for using the same button UI for two distinct kinds of interactions. However, it’s our job as designers to work with the platform and make sure that we give users the best experience possible. Unpredictable interactions will result in distrust and users will be discouraged from exploring and discovering your chatbot.</p><p><em>Make sure all interactions are predictable. If clicking a button opens a new page, the users should be made aware, most likely through the button text.</em></p><h4>9. If you must redirect users out of Messenger, use webviews.</h4><p>Facebook offers <a href="https://developers.facebook.com/docs/messenger-platform/messenger-extension">webviews and extensions</a>, allowing “custom experiences that still appear to be part of the Messenger thread.”</p><p>Webviews pull up over your conversations, mitigating the disrupted flow by giving the impression that the conversation is being continued inside the Messenger app.</p><p><em>If you must redirect, use webviews to make for a smoother conversational experience.</em></p><h4>10. Make your redirects relevant.</h4><p>It was astonishing how many Facebook chatbots redirected to a booking page without maintaining any of the user entries from the chatbot.</p><p>If the user just spent five minutes filling out information on your chatbot, that information needs to be absolutely be persisted in the redirect. Do not inconvenience the users without being able to offer something in return.</p><p>The number of chatbots that did this was so high that it makes me think that it may have been a technological limitation at the time, which makes me wonder, should the option have been presented at all in the first place?</p><p><em>Any information the user provided the chatbot with should be persisted in the redirect. If this is not possible — reconsider offering the option at all.</em></p><h4>11. Know when to keep context, and when to reset context.</h4><p>Context-keeping is great — it’s touted as one of the winning features of conversational interfaces. However, inappropriate context-keeping made changing dates completely unintuitive in Skyscanner.</p><p>Skyscanner had a bug/feature that made changing dates virtually impossible: it retained the previously inputted dates when users tried to modify their search.</p><p>In the below example, the user had first searched for a flight departing on December 16 and returning on December 18.</p><p>I asked him to change the departure date to December 20 — a date that was later than the return date of December 18. Skyscanner would throw an error, as the “return date needs to be after the departure date.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GNROm0N5B5u-LFMi9YBxcw.png" /></figure><p>Everyone was bewildered by this error message. In the context of the conversation, users thought that Skyscanner had become confused and was trying to change the return dates rather than the outbound dates.</p><p>What was really happening was that Skyscanner had not reset the previously inputted dates, but this oversight resulted in much confusion for the users.</p><p><em>If you’re saving previously inputted information, makes sure that information is being retained and reset at the right moments.</em></p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fupscri.be%2F58521e%3Fas_embed%3Dtrue&amp;dntp=1&amp;url=https%3A%2F%2Fupscri.be%2F58521e%2F&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=upscri" width="800" height="400" frameborder="0" scrolling="no"><a href="https://medium.com/media/c90883b0516ddbe5c59960c12a04bc64/href">https://medium.com/media/c90883b0516ddbe5c59960c12a04bc64/href</a></iframe><h4>12. Wording matters: Clarity over elegance.</h4><p>As you may have noticed, instead of using the more common terms, “departure” and “return,” Skyscanner opted to use the terms “outbound” and “inbound” to describe dates.</p><p>The words “outbound” and “inbound” may be more aesthetically pleasing than “departure” and “return.” Unfortunately, the similarity between words being so similar led to a lot of confusion for the users. It seemed to lead to <a href="https://en.wikipedia.org/wiki/Semantic_satiation">Semantic Satiation</a>, where words lose meaning with increasing repetition.</p><p>This became especially problematic when the users attempted to change flight dates. Users would also mistake the meaning of one word for the other.</p><p>Chatbots that used common, distinctive phrases like “departure date” and “return date” did not face this issue, confirming my hypothesis that the issue was due to the the diction.</p><p><em>Use distinct words that you would use in conversations. Clarity wins over elegance, every time.</em></p><h4>13. Wording matters: Confirm by asking, not stating.</h4><p>Say that a user is trying to book a flight <em>next</em> Monday, but the chatbot interprets it as <em>this</em> Monday, and states, “You’re booking a flight this Monday.”</p><p>This statement is less than ideal for a couple of reasons:</p><ol><li>Users may think that there is no option to correct the mistake</li><li>Users may feel that <em>they </em>made a mistake by saying the wrong thing</li></ol><p>Instead, a better way to confirm would be to ask, “Is it correct that you want to book a flight this Monday?”</p><p><em>Formulate confirmations as questions, not statements. Questions are a natural segue into error corrections and removes the blame from user.</em></p><h4>14. Wording matters: The question will form the user response.</h4><p>To determine the departing airport, Hipmunk will ask, “What airport do you usually fly from?” Most users did not have trouble with this step, but one user replied with “american airlines.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oz1nESsHLm_vB6cEyOWJ0Q.png" /></figure><p>This prompted Hipmunk to ask again, “Where are you flying from? Say something like “from SFO.”</p><p>This simple change in wording resulted in the correct response, “DFW.”</p><p>How you formulate your questions will affect the user’s responses. Although both questions had the same meaning, the use of “what” versus “where” resulted in two completely different answers.</p><p>If Hipmunk had used “Where are you flying from as its initial question, I believe the reprompt would have never been activated in the first place. This goes back to the point of users not reading — most are constantly scanning for words that pop out at them.</p><p>Unsurprisingly, words matter a lot for chatbots. The importance of thoughtfully selected diction cannot be overemphasized, especially keywords and phrases that can greatly influence the user’s understanding of the conversation.</p><p><em>Carefully consider the words you choose, especially for points of interactions. Particularly, be cognizant of the leading words such as “when,” “where,” and “what” for question.</em></p><h4>15. Don’t leave the user hanging.</h4><p>In traditional apps and websites, users have patience for only a few milliseconds before they think something is wrong. Designers have dealt with this challenge using loading icons and messages for applications that take more time.</p><p>We currently have a lot more leeway with chatbots since users currently do not have expectations for proper chatbot response times. Instead, user experience in messaging interfaces appears to be subconsciously informed by texting with other humans, which always has some amount of lag time.</p><p>In my user studies, I found that 8 seconds is about how long a user will wait until he or she starts trying to reboot the “unresponsive” chatbot.</p><p>Regardless of this generous leeway, users should always get hard feedback that 1) the chatbot is processing things and 2) a response is on the way.</p><p>Facebook has <a href="https://developers.facebook.com/docs/messenger-platform/send-api-reference/sender-actions">typing indicators</a>, which are useful for smoothing out experiences for short wait times, but they are automatically turned off after 20 seconds.</p><p>If you anticipate that it will take longer than a few seconds for a chatbot to reply to something, let the user know upfront, and monitor the process to let the user know that things are taking longer than normal.</p><p>We already do this with loading messages. For instance, a lot of apps say something like, “Things are taking longer than usual... Try refreshing?” Refreshing isn’t an option in chatbots, but you should still keep the user informed of the process going on in other ways.</p><p><em>The user only knows of what you tell them. Be proactive about loading messages. Don’t make users wonder if your chatbot is broken.</em></p><h4>16. Give opportunities to correct errors throughout the conversation.</h4><p>Neither chatbots nor humans are perfect. Undo and cancel are essential functionalities for a smooth error-correcting experience.</p><p>Until undo and cancel as a functionalities are implicitly understood across multiple conversational platforms, users will not know to call upon functionalities exist unless they are introduced to them. However, once they know this option is available to them, the users will feel that much more secure exploring the chatbot.</p><p><em>Allow users to go back, cancel, and undo. Be smart about informing users about these options at critical points of the conversations, so that the users know how to access those options.</em></p><h4>17. Don’t display important texts as captions.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*M_6yDJIq1fnD-4HZKIZkpw.png" /><figcaption>Expedia’s opening message.</figcaption></figure><p>Facebook offers the ability to place captions on images. Images are eye-catching; captions are not.</p><p>We’ve been trained through years of reading to regard captions as unimportant. The likelihood of captions getting read is even lower, and as we’ve discussed, people already don’t read.</p><p>Don’t do what Expedia did — display critical onboarding information as a caption, which also happens to be set don’t-read-me-gray.</p><p><em>Important details should never be displayed as captions.</em></p><h4>18. Keep the humor safe and professional.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*g3QwsVUFyfRd11kNe6-0NA.png" /></figure><p>There is a fine line between cute and annoying, or even dangerous… like this message Hipmunk sent.</p><p><em>Please don’t joke about charging my credit card. It’s one thing to be funny, but find ways to do it without risking user trust.</em></p><h4>More Reads</h4><ul><li><em>Part 2 of this series:</em> <a href="https://medium.com/@eunjinnyseo/11-more-best-ux-practices-for-building-chatbots-67362d1104d9"><em>11 More Best UX Practices for Building Chatbot</em></a></li><li><a href="https://chatbotsmagazine.com/usability-heuristics-for-bots-7075132d2c92"><em>Usability Heuristics for Bots</em></a></li><li><a href="https://chatbotsmagazine.com/testing-chatbots-how-to-ensure-a-bot-says-the-right-thing-at-the-right-time-58857ceb97f1"><em>Testing Chatbots: How To Ensure a Bot Says the Right Thing at the Right Time</em></a></li></ul><p><em>This article is part of Sabre Labs team’s analysis on emerging technology trends impacting travel. For more insights, download the </em><a href="https://www.sabre.com/labs/emergingtech/"><em>Emerging Tech in Travel 2017 Report</em></a><em>.</em></p><p><em>Please reach out to Sabre Labs at </em><a href="/service/mailto:sabrelabs@sabre.com"><em>sabrelabs@sabre.com</em></a><em> for further inquiries or discussions on the future of technology and travel.</em></p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fupscri.be%2F58521e%3Fas_embed%3Dtrue&amp;dntp=1&amp;url=https%3A%2F%2Fupscri.be%2F58521e%2F&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=upscri" width="800" height="400" frameborder="0" scrolling="no"><a href="https://medium.com/media/c90883b0516ddbe5c59960c12a04bc64/href">https://medium.com/media/c90883b0516ddbe5c59960c12a04bc64/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/705/1*mToBIVQSczojw0mSn2h5Sw.gif" /><figcaption>👏👏Clap below to recommend this article to others👏👏</figcaption></figure><figure><a href="https://docs.google.com/forms/d/e/1FAIpQLSfJDqeR7zbRBfQvx6mouQAhjK4lrCy3z9Z6Ppxln7f1xi12dg/viewform?c=0&amp;w=1"><img alt="" src="https://cdn-images-1.medium.com/max/255/1*rRjUc6JW61AP1UWsa-yItQ.png" /></a></figure><figure><a href="https://chatbotsmagazine.com/be-featured-in-front-of-thousands-of-people-interested-in-bots-e7040c4080df#.hswzhzpp9"><img alt="" src="https://cdn-images-1.medium.com/max/255/1*OGAA3ICxv3ds-RK2Sq4oxA.png" /></a></figure><figure><a href="https://chatbotsmagazine.com/want-to-receive-the-best-chatbot-related-content-in-your-medium-feed-20a9c11f11b6#.1t1yhwcyz"><img alt="" src="https://cdn-images-1.medium.com/max/255/1*1XecGp_cGZ1wGLAldRHLWQ.png" /></a></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3c46274501b2" width="1" height="1" alt=""><hr><p><a href="https://medium.com/chat-bots/19-best-practices-for-building-chatbots-3c46274501b2">19 Best UX Practices for Building Chatbots</a> was originally published in <a href="https://medium.com/chat-bots">Chatbots Magazine</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>