In the last twelve months, we’ve witnessed a huge surge in the development and adoption of chatbots, artificial intelligence (A.I.), and machine learning. Many startups, including my own (ReplyYes), are utilizing A.I. and chatbots to help consumers engage with brands through their mobile devices in interesting and creative ways. Examples include the Domino’s chatbot, which enables customers to order a pizza through Facebook Messenger; the Burberry chatbot for London Fashion Week, which helps customers order products they see on the runway; and Lowercase Alpha, which helps founders and friends of Chris Sacca’s venture capital firm Lowercase discover some of the best new apps in the world.
Given the increasing interest in the field and the venture capital dollars being spent to build creative chatbot and A.I. solutions, we’ve developed eight predictions that outline where we think things will evolve in the next 18 to 24 months.
- More chatbots will begin to solve real-world problems. In many instances, early chatbots seemed more like technologies in search of problems than customer-centric solutions. As the chatbot hype subsides, technologies mature, and companies get feedback from customers, the problems that chatbots tackle will become more obvious, and, in turn, more valuable. An example is our own ReplyYes’ The Edit, which endeavors to solve the problem of product discovery for music lovers. Through a use of progressive disclosure, short keyword interactions, and machine-based curation of vinyl albums, we give customers a personalized and serendipitous experience to help them find music they love.
- Interaction Design matters — a lot. At first glance, it might seem that a messaging system like a chatbot doesn’t need interaction design, that it’s simply a two-way dialogue between a user and an intelligent agent — like texting with your sister. That’s way off base. Interaction design in the context of chatbots is as important to get right as interaction design for websites or applications, if not more so. The flexible and fluid nature of conversational interfaces requires flexible design paradigms and workflows with the appropriate affordances that create intuitive experiences for users. We believe that interaction design for conversational interfaces will become a design discipline of its own.
- Artificial intelligence will augment, but not replace, human intelligence. The human language is incredibly nuanced and complex. When companies endeavor to solve problems with A.I. and chatbots, a hybrid approach with a human behind the scenes will be vital for success. Getting stuck in an infinite loop with a chatbot is a terrible and frustrating customer experience, and pretending there isn’t a “man behind the curtain” is folly. Humans will trump machines in terms of true conversation for the foreseeable future — so a hybrid approach that melds input from real humans with algorithms will yield the best experience.
- Vertical solutions represent the best opportunity for startups to create value for their customers. Truth be told, it’s really tough to build smart applications that can understand and interact with humans on a near real-time basis. This problem is infinitely harder for applications like Alexa, Cortana, Google Now, and Siri, as their horizontal nature enables users to ask them about literally anything. By constraining the scope of the conversation, startups leveraging off-the-shelf natural language processing (NLP) and customized machine learning techniques will have a fighting chance of building high-quality and ever-evolving chatbots that deliver a delightful and (mostly) automated experience to their customers.
- Bots will not replace applications. Some pundits have hailed the rise of intelligent agents like chatbots as the death of applications and, in particular, apps on mobile devices. That’s nonsense. There are contexts in which native applications, which can provide rich, interactive, and immersive experiences, make a lot more sense, and there are contexts in which lightweight, bidirectional interactions with chatbots will be easier, faster, and more enjoyable.
- Millennials will lead the shift to messaging-based commerce. For millennials and the generations that follow, messaging is the primary way they interact with the world. Messaging with friends and broader networks is the way this generation discovers new products and content today. By pairing discovery with a frictionless ability to transact, we see the 80 million millennials with $200 billion in annual buying power as the early adopters who make “conversational commerce” a category and not just a buzzword.
- SMS will continue to be a primary messaging interface and realm in which chatbots play. While Facebook Messenger, Snapchat, and WeChat have amassed massive populations of users, simple, plain-old text messaging is, and will remain, a robust channel for chatbots. Text messaging has its share of challenges, but it is an open platform in a way that none of the social platforms can or ever will be. Moreover, it works natively on every single mobile phone (smart and not smart) on every continent, and its usage spans all demographics. As long as people have mobile phones and a need for non-voice communication, text messaging will have a place. This creates a fertile and open playground for chatbots.
- Facebook and Snapchat will work to emulate the capabilities and success of WeChat. The U.S.-based social messaging platforms are still in their infancy in terms of providing the conditions and capabilities for businesses and consumers to engage and transact. WeChat now represents one-third of the ecommerce market in China. Over the next 18 to 24 months, we expect Facebook, Snapchat, and others to invest heavily in the features and user experiences that will enable a material shift to messaging interfaces for digital commerce and for commerce, in general.
We’re sure there are many additional trends that we haven’t predicted that will also come to fruition. Let’s keep moving forward together.
The audio problem: Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here