Tell us a little about your needs and one of our Customer Success Specialists will get in touch.
Chat bots – the new craze unlikely to live up to the hype
The start-up world is abuzz with the latest “it” technology – chat bots. From Slack, a start-up megastar itself, to Microsoft and Facebook, everyone seems to be betting the farm on chat bots taking over the world. Chat bots – programs that are embedded in commonly used messaging systems to trigger actions in response to user messages – are on fire, as companies are sinking millions into their development. However, despite the craze, they are likely to occupy only a niche position without gaining a leading role in tech, unlike the predictions made by their proponents.
First thing first, why are chat bots so hot today? There are two main reasons: new sources of revenue and integrations, and the promise of keeping 100% (or close to 100%) of the revenue generated by app transactions.
- Integrations. Many apps nowadays have a very specific purpose, and it’s often difficult to diversify away from it. Chat bots offer an easy solution. Let’s take Slack . In addition to its core messaging purpose, Slack acts as a radio receiver (if you’re old enough to remember what radio is). It receives messages generated by other systems and consolidates them in message streams. It is capable of transmitting messages to other systems through “bots”, but that functionality is not widely used. There is only so much value in being a message board for small and medium size businesses, so the logical change for Slack was to expand its capabilities. Becoming a “hub of everything” vs. being just the notification hub it is today. Chat bots allow Slack to keep everything that made it a winning product (a messaging app), while adding new capabilities using the same model and user experience. Other vendors sense similar opportunities – diversify your system with chat bot functionality to generate ecosystems and revenue above and beyond what they would otherwise.
- Keeping revenue. In a mobile-centric world, only 2.5 names matter – Android, iOS, and Windows. The three companies behind them (Google, Apple, and Microsoft) are the gatekeepers that wield all the power when letting apps tap into their ecosystems. This ecosystem access comes at a cost – 30% of revenue generated by these apps (whether direct sales or ads) goes to the gatekeepers. In-app chat bots would allow bypassing this revenue splitting. Chat bot-enabled WhatsApp, for instance, can be used to order pizza with a single message. Facebook can collect payment from the pizzeria for facilitating the transaction without paying a 30% cut to Apple or Google. The combined revenue that Google and Apple generate from their app stores exceeds $10B. Give that back to the app vendors and the savings to service providers are substantial. With the potential of ecosystem expansion and revenue growth, chat bots look golden.
With these great promises in mind, is the hype around chat bots justified?
Let’s digress to another technology that was supposed to change the world. I’ve seen lots of people using iPhones, Android phones and even a handful, including myself, using Windows phones. But I’ve never seen a single person using voice-based Cortana, Siri or the Google equivalent (I bet, you’re hard-pressed to remember its name). I mean, I’ve never seen them in real life. I’ve seen plenty of examples of product placement in TV shows, advertisements, and movies, but I’ve never seen an actual person speaking into their phone trying to schedule an appointment.
The marketing value of voice-activated assistants is undeniable – when Siri came out in late 2011, not a single day would pass by without news articles about Siri. From a marketing perspective, whatever the development cost was, the marketing benefit was worth the expense. From a usability perspective – not so much. These systems have inherent flaws that won’t be overcome anytime soon, so Siri, Cortana and Google Now will remain nothing more than marketing gimmicks in the foreseeable future.
The same problem is likely to plague chat bots. Not to the same degree as for voice assistants, but close enough. A basic flaw will limit their use. This flaw can be described using the following three acronyms: UI, AI, and UX (user interface, artificial intelligence and user experience).
- UI. Chat bots, as their name suggests, are smart programs that analyze message flow and react to action triggers. To initiate an action, you have to type “keywords” that trigger a bot’s response.
- The appeal of messaging systems is in their simplicity – typing text and emoji requires no training or specialized knowledge. Chat bots would change that. Initially, communication with sophisticated chat bots (those that go beyond basic reactions, such as displaying automated greeting messages) would require knowledge of keywords or formatting to invoke the desired action. Learning to talk to chat bots, especially in the early years of their development, before varying flavours of communication syntaxes naturally converge to a standardized format through evolution, is a step that not too many users are willing to take.
- A second UI challenge is a lack of sense stimulants, visuals being the biggest one. Often, buying products or services involves going through myriads of sense-stimulating cues – from product images (visuals) to the smell of fresh coffee in the morning to opinion sharing with other people. This multifaceted experience contributes to the buying process and satisfaction with the purchase. Chat bots offer a less engaging experience. Typing “I want a large pepperoni pizza at 3 pm today” doesn’t offer the same experience as picking a pizza in person or clicking on images of toppings while assembling it online. Of course, texting would be sufficient for many niche tasks, but it will be a poor replacement for many other tasks that rely on visuals as part of the process. Eventually, chat bot systems will incorporate necessary aids, but this will be a relatively long (in IT terms, of course) evolution process through trial and error.
- AI. To address the issue of users needing to learn how to interact with each bot (the issue mentioned above), AI (artificial intelligence) will have to step in to process natural language. However, after 100s of millions in R&D, Cortana, Google Now, and Siri are still quite inadequate at understanding natural speech. It will take years before machines are capable of understanding human speech to the degree necessary to correctly process the nuances of conversations. The current state of AI and machine learning will inevitably create a bottleneck for chat bot advancement.
- UX. User experience from chat bots will put brakes on their adoption and proliferation also. Having to learn special syntaxes to communicate with each bot will deter people from using them. Advanced AI is not a panacea either. Even with the best of AI, certain inputs from users won’t be processed correctly. What does “oh, man, I’d kill for a pizza now” mean? Does it mean that the person is hungry and is simply stating it to a colleague or is it a signal to order a pizza? At the early stages of their introduction, the chat bots would generate too many false positives and false negatives. Every time a chat bot interrupts at the wrong time or doesn’t react to a command, the user gets frustrated with the bot. Product usage demands consistency and predictability – if a car responded to steering inputs only 99% of the time (a tall order to meet for bot accuracy), you’d probably stop using it after the fist missed turn. If a few keys on a keyboard failed to work occasionally, it wouldn’t be long before you’d stop using the keyboard. Bot adoption will rest on their dependability, and, especially at the beginning, it will be low or will involve a learning curve, impacting user experience.
The second issue with bots (the first being the UI, AI and UX technical shortcomings) is the recent trend of single-purpose applications. With the emergence of the iPhone, users grew to expect simple single-step actions. Most recent successful apps and platforms do only one thing well – Wunderlist only does to-do’s, Slack is for messaging, Buffer and Hootsuite is for social media management, etc. Users associate a specific need with a specific application. Did you know that your MS Outlook has a weather forecast built into the calendar? Even if you do, I’ll bet you still open your weather app or go to a weather site to check the forecast even if it sits right before you in the perpetually opened Outlook. Chat bots attempt to add functionality to other apps. In theory, being able to adjust the thermostat and then order a new movie directly from WhatsApp sounds appealing. In reality, it confuses users that open WhatsApp only to message their friends. At best, chat bots will be underused. At worst, they will ruin the experience with the core product.
Much like voice assistants that have become commonplace in specific applications, chat bots will eventually find their niche spots and serve their intended purpose. However, until then chat bots are likely to follow the typical hype curve (a sharp rise in expectations, followed by a collapse and mistrust in technology, with a modest recovery once their true purpose is established). Right now, we are still in “the hype” stage and everything about chat bots – capital investments, promised capabilities, and expectations – must be taken with caution. The hype is there, but the technology is unlikely to live up to it.
Tags: #chat #chatbot #chatbots #slack #microsoft #facebook #whatsapp #meemim #integrations #wunderlist #hootsuite #buffer