Over the past few months the world of digital products and UX has been buzzing with talk of bots. Silicon Valley’s biggest and brightest have been busy implementing new bot technology, and the era of the bot seems to be just beginning. So, we set out to get to the bottom of the bots; to explore just how innovative bots really are, and to ask the tough questions. Does every product need its own, personalized bot? And, what exactly do bots mean for UX designers and planners?
Bots, Who Are You?
Bots, short for robots, are simply programs that allow users to interact with an application through dialogue. It’s probably safe to assume that you’ve already encountered bots, as they are integrated into your desktop and mobile operating systems, they’re in automated voicemail services, and in your messaging applications. Bots can be an extension of an existing service, or a product in and of themselves. For example, you might talk to a service bot when calling your favorite shoe store to place an order from your home.
Bots have been popularized as personal assistants that allow you to interact with multiple interfaces simultaneously. Imagine an assistant who in the blink of an eye has simultaneously booked your flight, ordered you an Uber, set the outgoing message on your email, and texted your mom your flight ETA. In a nutshell, the bot dream.
Demo: Paypal payment messenger – Karol Podle
Why Is Everyone Talking About Bots?
There are several key reasons why we’ve seen this surge in bot technology:
1: We’re Sick and Tired of Apps But Addicted to Messenger
There are currently over three million applications available for users to download from the various app stores. However, the rate at which users are downloading new apps to their mobile devices is starting to slow. The average smartphone user has 42 applications on his device, but spends 90% of his time on just nine or ten of them.
Not surprisingly, users spend the majority of their time on messaging and communication apps. These popular apps (like whatsapp, facebook messenger, snapchat) seamlessly integrate bots into our everyday lives without us even noticing, and allow companies to reach users in their day to day lives without the users having to download a specific application.
2: Artificial Intelligence Got More Natural
The technology that drives bots, Artificial Intelligence, is improving rapidly and dramatically. Currently, all the major players from Silicon Valley, with Google leading the charge, are investing effort and resources in developing AI technologies. These new functions, such as smart searches and identification systems, were previously just the figments of an imaginative tech-guru’s daydream. As a result of these new advancements, we can enjoy bots that not only understand natural, innate speech, but are also able to talk to users with a natural ease, and logical integration that wasn’t possible before now.
3: We Want To Break Free From the Keyboard
With the advent of the “Internet of Things,” and the prevalence of mobile devices, computers have entered every sphere of our lives. Some of these situations make using the keyboard, or even tapping on the screen, difficult if not impossible. Running, driving, and even cooking are examples of day to day activities where bots can assist us practically. Thanks to bots, we have the ability to give directions, ask questions, and manage tasks, all hands-free.
Google Home an assistant-style bot helps you manage your home (and life) screen-free
iPhone, Android, Facebook: Bots are Everywhere
Bots aren’t new. Since the dawn of computers we’ve been trying to find a way to start a conversation with them. From the BBS systems of the 80’s, to the MIRC chat technology of the 90’s, through modern operating systems like Linux. These bots are primarily geared towards advanced users, who know the specific lines of code needed to operate said bot. Advanced users can use bots for shortcuts when performing complex operations. Recently, AI technological advances have taken the power of bots away from the computer geniuses, and literally placed the control in the hands of everyday users. By implementing intuitive and simple bot technology into personal devices bots have slipped into the mainstream, and become an integral part of our modern technology.
Since its first implementation, the conversation on digital assistants and bots has been dominated by one name, Siri. Apple’s Siri is a personal, digital assistant, it is voice operated, and has been integrated into iOS from iPhone 5, iPad, appleTV, and is integrated into the newest version of the macbook. Overall, Siri is pretty talented. It allows her users to interface with many different applications through voice commands. Soon Siri will even be able to connect to your computer to locate files and perform a variety of tasks on the web. Google Assistant may be less stylish and lack the glamour of Siri, however in many ways Google’s version of the personal assistant is more ambitious. Google Assistant is a combination of personal assistant, search engine, and app launcher. The main draw of Google Assistant is the bot’s ability to learn the needs of the user, and optimally to anticipate them, and present a response before the user can even ask the question.
At the most recent I/O conference, Google announced an exhaustive revamp of Google Assistant across a number of platforms. These innovative advances range from a Google Home, a speaker that would allow users to interact with Google assistant seamlessly from anywhere in their home, to two brand new messaging apps Allo (think whatsapp) and Duo (videos), all of which are integrated with not only Google Assistant but also with external bots.
Demo: Integration of Google bot capabilities and Allo in a group chat (Google Blog)
Facebook, meanwhile, has cornered the market on messaging apps largely thanks to the resounding popularity of Whatsapp and Messenger. The brand new bots integrated into Messenger are the starting gun for the race to make messaging apps a thriving new business arena, which in turn has strengthened Facebook’s vice-like grip on the world of digital business, marketing, and products. Last May, Messenger was opened up to developers causing a crushing influx of new bots across a variety of arenas. These bots have begun to revolutionize our experience as users and consumers across many products and platforms. From ordering a pizza to movie recommendations, bots are now integrated and integral in our modern user experience.
Demo: Booking a flight via messenger – Isil Uzum
The brand new Messenger bots have been met with a mixed response. Many users complain of slow response times, and an overall negative user experience. Facebook, in an attempt to quell the skeptics and the critics, explained these are simply the birth pains of a first generation of bots that will be sorted out in the versions to come. Facebook has also invested in acquiring start-ups like wit.ai, a key developer of bot technology, in what seems to us to be an active and concerted effort to levy bots as an essential part of business promotion and user experience.
The Challenge: An *Almost* Design Free Interface
As UX professionals, how do all these bot innovations affect us? Well, one of the primary challenges with bot technology is the lack of graphic interaction. These bot-systems are fundamentally text or voice operated, so the integral design elements are not graphic but rather tone or text driven. A bot in a messaging app, with its own style and limitations, can exist without any control over the visual interpretation of its communication. This means that we have to redefine the boundaries of user experience design by returning to the basics of UX/UI, which as always, comes with its own unique set of challenges.
Orientation and Discovery
Anyone who has lost their way trying to navigate through their personal voicemail system, or wasted time dealing with an automated directory service, has had a first-hand experience with lack of screen design and navigation aides. One of the main problems when designing bots is the lack of visual cues to assist the user in orientation. Without navigation menus, set grids, or breadcrumbs, users can find themselves disoriented, and lost in a UX world devoid of visual guides. Bot planners have to be aware of the lack of visual orientation aides, and build bots with more intuitive voice menus that help the users to understand their options, and create a natural sense of control and flow.
The Transition From Text to Visual Interface
Screen based bots allow us to rapidly switch between different modes of interface, for instance from textual conversation to more classic modes of interaction, through complex integration, by clicking or touching. Take a look at the example below. This conversation begins with a simple text conversation between two friends about buying sunglasses. When the bot is added to the conversation, and we can see that the levels of interface are richly layered, and yet surprisingly intuitive. Beginning with choosing the parameters of the desired sunglasses, through simulating how they will look when worn, to a 360 degree model, the bot is seamlessly integrated into the conversation, and shopping experience, leading the user through the various stages of the process.
Demo: Integration of screen-based button and product gallery in messenger conversation – Isil Uzum
This illustrates one of our greatest challenges, the separation between visual interaction and classic written elements of interface. As designers, the challenge lies in knowing when in a conversation to use solely chat-based elements, and which aspects of the interaction are better suited for a visual interface. This is no foreign conundrum – we’ve all been forced to switch between two interfaces on our phones, trying to balance which is more relevant, and all the while struggling with the headache of the physical act of switching from text to voice to a visual interface, or vice versa.
Flexible Input and Double Entendres
It’s impossible to know how any specific user’s experience with a bot will turn out. As designers, the order of operations and tasks might seem innate, but can be complicated and very intimidating to a less than tech-savvy user. When we design screens without normal visual aides (i.e. buttons, menus, information fields, and other familiar UI staples), and instead replace those elements with a conversation, it opens up a pandora’s box of possible input and reply nightmares. As designers our challenge is to understand the line, or lines, of questioning that lead our users to a specific piece of information they might need, and reverse engineer the conversation. A conversation with a bot opens up a wide range of possible outcomes.
A bot programmed to set up reminders for upcoming appointments, is required to be able to decipher the two following sentences and deduce that they actually refer to the same task:
The bot has to not only understand the ambiguous elements of the two tasks (does the user mean 8am or 8pm), but it also has to be able to consolidate the textual name “Shani,” with the contact “[email protected],” and understand these two very differently worded inputs are one and the same. The challenge in designing this bot, and bots like it, is a problem of understanding. We are required to find a way for the bots to understand a variety of messages, and decode them in a way that consolidates our scattered ideas and natural way of speaking.
Designing a Conversation
Even the most simple of human conversations is built on layers of language, social clues, natural rhythm, and unspoken nuances of emotion. Therefore, our bot must be designed to handle the obscurities of human interaction. One of these challenges comes to a head when designing a conversational bot that must understand not only contextual cues, but also that two separate sentences, texts, or statements are related, and more importantly how they are related. Bots must understand when the information in one of our phrases is related to a previous statement or if it is a remark that stands alone.
Creating a Natural Flow: The bot connects elements of both conversation and context
The above interaction is a great example of a bot that understands context and conversation elements. The bot is able to understand that the first and the second questions are related, and provide the user with the necessary information. As humans, we instinctively know that the user is asking for clarification about the bus he had previously mentioned. The challenge is getting the bot to understand that the use of “there” is building on the previous question.
Tone, Personality, and Most Importantly Style
When we design a screen interface for a user we rely on many different design elements (color, shape, movement, etc.) to convey a desired feeling, identity, and atmosphere. When planning and designing bots, our ability to create a visual interface is very limited, so we have to rely on writing style, tone, and context to convey desired creative elements.
The design of bots with specific personas, names and style of speech, allows us to manufacture a distinct feeling, identity, and, at the end of the day, user experience. How a bot speaks and interacts verbally, or textually, with the user will affect the outcome of the bots service value and the user experience as a whole. For example, a bot that is embedded in an app geared towards teens should speak with more slang and colloquialisms, whereas a bot that will be implemented as a tool for designers should understand and speak the technical lingo. Bots should understand how to use language to influence human emotion and psychology, for example phrases such as “let’s get started,” “you’re almost there,” “thanks,” and “good luck,” and understand when and how to use them to engineer a desired mood or feeling.
Dealing with Errors
As the wonderful world of bots continues to develop, and bots become more widespread and integrated into our technology, we face a whole new rash of errors, misunderstandings, and problems with user interface. These first experiences with bots are pivotal for everyday users because they set the stage for how, and if, the future user experience will continue. If first-time users find themselves mulled down in semantic errors and usability issues then they will be apprehensive to use anything that looks like bot technology in the future.
As in normal conversation the bot must help the user to clarify the request
The conversation above should be a simple back and forth, however the bot is first confused by unclear audio, then it gets bogged down in semantic understanding. We as designers need to be aware of these possibilities in order to prevent unsatisfactory user experiences and interactions before they can happen.
Bots, Bots, and Only Bots?
As with any new form of technology, we have to take a step back and ask ourselves, are bots here to stay? Is this new platform worth investing time, design, and money in, or will bots be another passing trend? On one hand, bots hold the promise of a bright future; a future where programmers and app designers can reach a large audience, simplify and avoid installation issues, and increase audience engagement. On the other hand, the challenges are just as big, and include dependence on external platforms, limitations of design experience, and an ever evolving interface.
We recommend you think of bots as a part of a larger world of interface and AI options. A world which is increasingly more available for designers to use and integrate into our products. Depending on the goal of the app, and the general user experience for a specific product, bots can help to strengthen certain aspects of your app. The challenge is that a bot alone cannot carry a product. In spite of all this, if your product provides a structured and clear system of information, an organized catalogue, performs specific simple tasks, or provides basic services, then bots just might be right for you. Adding bots to a product that fits these parameters can help add value to your users, and open a new and promising world of possibility.
Want to learn more about The Uprising? Here are a few bot-related links we recommend:
- For Designers > Matty Mariansky from Meekan, published two pieces in AListApart focusing on reviewing interface design principles. (Part 1 and Part 2)
- For Developers > There are a plethora of startups offering tools and services that can help build bots, including Wit.ai (acquired by Facebook one year ago), as well as Watson by IBM, and Chatfuel (creates bots without using code)
- For Mock-ups > botsociety.io, tools for creating mock-ups of messenger bots (coming soon in other messaging platforms as well)