As the interest in chatbots and virtual assistants continues to grow, natural language conversation will surely continue to overtake browser and apps as the medium of choice in coming years.
Barriers to entry for new bot builders continue to decrease. The pace of bots released grows as bot building platforms and connectors multiply, offering increasingly powerful and easy to use tools. In April 2017 Facebook messenger achieved 100,000 bots on its platform , whilst Microsoft claims 20,000 bot developers on its’ Bot Framework . That’s a lot of bots!
Continued increases in Monthly Active User (MAU) numbers on conversational channels further attest to the importance of conversation as a preferred medium for users, 1.9Bn MAU on Facebook messenger, 320M MAU on twitter, WeChat fast approaching 1Bn MAU, and 1.2Bn MAU on WhatsApp. (Statista.com)
Although A.I. powered bots continue to improve rapidly, they are still in their early days and limitations remain evident. Bots, although quickly becoming commonplace in e-commerce sites, supplementing the browser interface and guiding potential customers, overall are yet to gain significant traction in broader community applications. Why is that? Other than with the major proprietary ‘assistant’ bots most users are still not making contact either primarily, or directly with a bot. Mostly remaining as secondary interface only. Is it because they are not very useful, or perhaps too hard to find?
In the case of the proprietary ‘assistants’, they may appear more powerful than most other bots, however their ‘expertise’ or knowledge remains broad but shallow. How often does Siri end up serving you a Google search? These proprietary assistants will continue to improve, but will always be limited by the resources that their creators either want, or are able, to devote to them. The information they serve by their very nature will be biased towards whatever their corporate owner wants them to provide (or manipulate) you with. Their usefulness is limited by the resources and impartiality of the corporation.
Outside the major corporates, most bot builders domain knowledge is specialised and limited, and their resources scarce. As such it will make sense for them to specialise and develop deep but narrow domain expertise bots. So does the user have to talk to many bots for many answers? Even then, if a developer manages to develop a useful bot, monetization of this service or knowledge remains a problem.
So as the number of bots increase new challenges arise for users, builders and service providers. For users, issues around whether they should, for convenience, just accept the broad, shallow, and maybe biased knowledge provided by the corporate ‘assistant’ bots, or should they spend precious time trying to find and curate multiple domain specific bots to resolve any deep domain queries or requests they may have? As it stands options to easily find and assess bots are limited, and who wants to do this? Why should users be required to maintain multiple conversations and interfaces? That is not a lot of fun (or utility) and it represents significant retention issues for bot developers.
Should the increasing number of bot builders and service providers focus on building broad bots to capture the most users, or remain within their specific domain of expertise? How do they encourage a user to access and remain engaged with their specific bot? Just as importantly, how do they monetise the service they are providing?
What if all bots (IoT devices and any digital services) could talk and exchange their unique information or knowledge directly with each other? What if a user could easily connect to this entire ecosystem through any of the bots (or services) already connected to that network? That is, talk to one bot to reach the intelligence of all bots. What if each bot (or service) creator could choose which bots it (and their users) spoke to, and how all its interactions were routed, ranked, resolved and monetized? What if the information exchanged had a marketplace for all connected bots and services to transact in? What if that marketplace allowed information to be fairly priced and transacted through frictionless, trustless, micro-payments?
We asked these same questions in defining our vision for The Hut34 Project and we invite you to participate. Together we can find the answers to those questions and more.
Join us, and perhaps together we can be one step closer to creating the world’s first superintelligence?
Source: Chatbots Magazine