There are over 500 million small to medium businesses on the planet, and without them life as we know it simply wouldn’t exist. But more than 50 million of those businesses fail each and every year, causing devastating effects for the business owners and their families. Can chatbots fix this problem?Dale Beaumont is the Founder and CEO of BRiN, a business advisor powered by artificial intelligence. They've found that there are two main reasons why business owners fail: the first is a lack of education, and the second is lack of support. So they set out on a mission to make education and support instantly accessible to millions of business owners around the world, through their chatbot app BRiN.At the Chatbots Track of Deep Learning Summit in London, Dale will be speaking on a panel that will explore humanising AI assistants. I spoke to him ahead of the event to learn more about the future of chatbots and how we can drive progress in the field.What do you feel are the leading factors enabling recent advancements and uptake of virtual assistants and chatbots? It was Justin Osmer that said, “Give them tools they need and they’ll turn around and build unique and creative things”, so the leading factor to recent advancements is that the tools needed to build chatbots and digital agents are now falling into the hands of developers and even enthusiasts. Plus, the proliferation of smart-phones means we’ve become an ‘always-on’ society, and the demand to get answers and service 24 hours a day, seven days a week is now upon us - so chatbots are here and ready to service that demand. What present or potential future applications of chatbots excite you most? The current applications for chatbots extend to customer service, e-commerce, news, weather and other forms of on-demand alerts. While mostly helpful and efficient, chatbots need to get better if they are going to reach mass market appeal. In the near future I’d like to see bots working hand-in hand with websites. This will help with discoverability and get more people interacting with bots, and once a conversation is started, it can then continue within other services, such as Messenger, Skype, Kik or Slack. Outside of this I’m most excited about the voice capabilities that must one-day come into bots. I believe this will usher in a new era and the focus will shift to digital agents and virtual assistants. With improvements to natural language processing (NLP), chatbots will be able to go ‘off piste’ and give users a unique and highly personalised experience. When this occurs, use cases for chatbots will grow exponentially as they begin to surpass the knowledge and capability of human agents. What do you feel is essential to future progress in this field? After Facebook released it’s bot platform in April 2016, I, like many people, tried several bots only to quickly find numerous limitations and faults. Many of these were due to poor UX design and developers building with a one track mind. With many of these faulty bots it has been this way for months, so clearly they’re not spending anytime improving their bot. This brings to our attention the problem and the solution for chatbots. I think that leaders like Facebook and Microsoft should stop giving tools, and instead give training. It is not about the tool but it is about the experience. Companies that want bots to succeed need to spend the same attention and care showing technical developers and non-technical managers and marketing teams how to design their bots. Further to this point, the large corporations pushing the chatbots field need to recognise that most companies could benefit from a bot, but most do not and never will have the internal capabilities to build them. So they need to invest in the eco-system by training builders. My suggestion is that they develop programs that enable individuals to become a certified bot builder, which will set a better standard for the industry. Next the responsibility has to fall on the companies building bots (internal or external), to monitor every interaction so their bots get better. If people can’t see progress they won’t persist with inefficient technology. For this companies need to assign people or teams to investigate customer requests in order to promptly uncover the break points and fix them. What developments can we expect to see in virtual assistants in the next 5 years? While chatbots are extremely popular right now, I believe that 2017-2018 will be the year the spotlight turns to digital assistants. It started with the release of Amazon’s Alexa, which was followed by Mark Zuckerberg’s announcement that he was building a voice-powered assistant that he described at a "Jarvis for the home". Then Google entered the race with Google Home. Once these devices hit a tipping point, interest in this area will skyrocket and people will begin to demand these type of experiences in their personal life and working environment.First will be the home. In the future these devices will get smaller and faster, and families will have several in their home - and when the hardware becomes small enough they will even be embedded into common appliances as well. Next is the car. Some would say this already exists with Siri, but everyone knows Siri is limited. I believe that unless Apple dramatically improves her capabilities and fast, the door will be open for Microsoft Cortana, Google Now, and even Tesla to develop a voice-based assistant, made for cars. After the home and the car, the next place I see chatbots and virtual assistants making developments in is the office. For years Microsoft’s lab has been experimenting with voice-based assistants and with Cortana’s recent improvements, I believe they are best placed to dominate within this space. Outside of these environments, I believe there is a huge opportunity for thousands of niche digital assistants, which is why we are working in this space. To start with, these applications will be built within native smart-phone apps. However, when the clear winners emerge in each category, and their are some better policies about who owns the IP, niche assistants will become omnipresent in hardware. What do you feel are the biggest challenges for adopting chatbots to everyday life and industry? In the early days of the internet the biggest challenge was educating people. Most people struggled with the question, “So what can the internet do?” and “Why can’t I just use the Yellow Pages?”, and it took years to get people over that hump.
Chatbots face a similar challenge, only this time the questions are, “So what can a chatbot do?” and “Why can’t I just use the internet?” To solve this challenge developers need to first ask this question, “In what ways does my chatbot offer a better experience than the internet?”
If answers to this questions are “Because it is cool” or “Because people can access my product or service within Facebook Messenger”, then I don’t believe your chatbot has what it takes to succeed. Only if you have an experience that is better in several ways than the internet should you continue. Once you have something that offers a better experience, it is a matter of being vocal about your chatbot so people know what it is, where to find it, and most importantly how to use it.
Dale Beaumont will be speaking at the Chatbots Track on 23 September, as part of the Day 2 of the Deep Learning Summit in London.
The next Virtual Assistant Summit will take place in San Francisco on 26-27 January! Speakers include Roberto Pieraccini, Director of Advanced Conversational Technologies at Jibo; Jordi Torras, CEO & Founder of Inbenta; Rachael Tatman, PhD Candidate at the University of Washington; and Alonso Martinez, Technical Director, at Pixar.Discounted tickets are available until 7 October, book your pass at an Early Bird rate and view more speakers on the event website here.