Talking Up Chatbots at the Berlin Chatbot Summit
Talking Up Chatbots at the Berlin Chatbot Summit
The Berlin Chatbot Summit covered the chatbot landscape, conversational user experience, and more. Read on to get the low-down.
Join the DZone community and get the full member experience.Join For Free
Start coding something amazing with the IBM library of open source AI code patterns. Content provided by IBM.
Now in its second year, the Chatbot Summit has grown into a series of international events that span cities. We attended the Berlin summit to find out more on a tech trend that is growing at an unprecedented pace.
The attendees (and speakers to a lesser extent) were refreshingly diverse, possibly due to the Berlin location, or due to chatbots being a relatively new technology (in their current form) and thus lacking historical cultural baggage and inertia.
Are Bots Taking Over?
While people talk of robots taking over our jobs the reality is far more likely to be AI-driven chatbots that take over particular tasks rather than whole occupations. Research by McKinsey suggests that as many as 45% of the activities individuals are paid to perform can be automated by adapting currently demonstrated technologies. In particular, jobs that involve a level of analytics, research, or processing that a computer can do better than a human are the most impacted —speakers spoke of the tasks of accountants and legal and tax advisers replaced by chatbots, typically leading to retraining and reclassification for employees. The conference largely focused on customer service roles where chatbots are utilized (often answering questions that arguably could be replaced by a good Frequently Asked Questions page and search engine.
You could probably find a use case for chatbots in most aspects of life. The question is whether you want to. For those who live a highly connected life, researchers at Stanford University have created Woebot, a bot designed to create an alternative to in-person therapy. Drawing from a therapeutic framework known as cognitive behavior therapy, Woebot asks people how they’re feeling and what is going on in their lives in the format of brief daily conversations. Woebot also talks to you about mental health and wellness and sends you videos and other useful tools depending on your mood and needs at that moment. You can think of Woebot as a choose-your-own-adventure self-help book that is capable of storing all your entries and gets more specific to your needs over time. At a time when telemedicine is slowly becoming a reality, services like this seem like a natural, inevitable extension.
Another growing where bots are deployed is in entertainment. Hello Barbie, companion chatbot built into a Wi-Fi-enabled talking Barbie doll. Barbie has a conversational repertoire of questions and answers that exceed that of some adults including “So do you find it easy to keep New Years’ resolutions or do you find them difficult to stick to? And hats have so much attitude! How did you feel wearing it?” It’s fair to assume that such developments will grow into other areas of entertainment, with companies such as Disney already heavily invested in chatbots.
A number of conference participants questioned our reliance on the keyboard and pointing device as interactional interfaces, wondering if voice activated interfaces would replace them in function and user adoption.
Depending on the complexity of the tasks for your chatbot, the coding behind it may not always be that complex, but designing the user experience around the code is the more interesting challenge, just adding an interface to an API is not enough. Adrian Zumbrunnen spoke about trying to turn his website into a chatbot, think about how you would represent each section? For example, a Contact Page would be a perfect use case for a replacement chatbot. One of the UX mantras you hear frequently at chatbot events is “don’t just replicate a form,” think about how the form would be represented as a conversation with another person and how you would flow from one field or question to another.
Users generally expect conversational UIs to build upon steps they have taken so far. For example, if you have already mentioned the recipient of an email, then a user will assume that any further steps in the current conversation are also about that person, so your bot should possibly assume (or check) the same.
Another major difference with conversation UIs is that they are expected to maintain history or state from previous conversations. Whilst you don’t always expect a website to remember what you did on a previous visit, people generally expect (even the most fractured) conversations to have a semblance of history within them. Looking at this from a positive angle, this also gives designers the ability to do more with personalization, preemptive design and the knowledge you have about a user from an authenticated account.
As a pair of writers, it was good to hear reiterated over and over again the need for good copy in chatbots. Getting the balance of humor and smarts correct is a delicate balance that is dependent on the context. For example, humor in a financial planner bot would be inappropriate, but fine in a holiday booking bot. The same applies for adding ‘smarts’, if a user is made to feel stupid when using a bot, they are unlikely to use it again. Understanding the context of language is important, for example, consider the difference between “Great!” and "Oh Great. It’s a subtle change but can lead to misunderstandings.
One of the more interesting talks was by Beerud Sheth, CEO of Gupshup who have created Interbot, a platform for bots to talk to each other. Developers can upload their own bot to Interbot or use bots that have been uploaded and shared on the platform. Interbot can piece together bot functions to combine functionality and AI smarts.
Through intercommunication, bots will be able to:
“… collaborate, delegate, transact, negotiate, repair, and self-upgrade.”
An example would be a customer service bot that has an array of bots behind it from each department of a workplace (for example, accounting, legal, HR) that could communicate information to the front line bot, expanding its capabilities.
Sheth believes that in the future town and neighborhood bots will communicate. For example, a bot in home security alerting local bots in the case of an intruder. We can expect bots to be a part of Vehicle to Vehicle communication in autonomous cars, with scenarios like people would pay to get someone to move out of the fast lane.
Too Much Talk?
Like most conferences, there was the usual glut of stalls of B2B services. Questioning the companies, most couldn’t offer a compelling point of difference or competitive advantage to other companies exhibiting, although many had been operating in the space for over 10 years. With emerging tech especially, one always gets the impression that attendees expect to hear magic answers to solve all their problems and questions, but it’s such a new field, no one is really an expert with all the answers yet and often attendees looks non-plussed or unimpressed as they slowly realize this. The chatbot landscape is fresh, new and open to experimentation, if you have an idea that might suit the medium, you can become the expert.
Opinions expressed by DZone contributors are their own.