We’ve reached an inflection point: the conversational interface is now viable for a much wider set of tasks and applications. The story of how we arrived here is for another day. But put simply, with advances in AI, machines are increasingly able to converse intelligently with users. This is a huge boon since we spend a lot of our time using chat based apps (Facebook Messenger, Slack, SMS). The dominance of the conversational interface is the starting point for any predictions for the future of user interface and user experience (UI/UX). Here are three of mine:
1. Businesses will build fewer lousy apps
I am no sports fan. On the rare occasion when I take my boys to a Yankees game (against strong oppositions from their Boston grandparents), the last thing I want to do is download the Yankee Stadium app to order a hot dog. There are two million apps in the Apple app store and 28 slots on the first page of my iPhone. It is beyond optimistic that any businesses’ app will make it into slot #86 of my phone. The interface challenges, though, go well beyond managing to secure real estate on my home screen. As a user, what I want is to buy a hot dog without leaving my seat. The current app workflow requires me to:
- download the app over cell tower
- open the app
- create a login
- create a strong password
- enter my credit card information
- navigate to the menu
- select the hot dog
Halfway through the app download, I’ve already bought the hot dog from the stand, the old-fashioned way.
A much better solution is to make that same purchase via a pre-authenticated messaging medium such as email, SMS, or Facebook Messenger, with far fewer steps. The AI agents on the other end already know my identity, my credit card info, and my preferences. A few simple exchanges later (made in plain English), a hot dog will be delivered to my seat. When I leave the stadium, I don't have to waste time deleting the Yankee app.
AI agents will play a larger and larger role as AI gets better. At x.ai, we believe that artificial general intelligence is many years away from reality. In the meantime, we are focused on building a vertical AI assistant (also known as narrow AI). By tackling a single domain—meeting scheduling—we can build something delightful with today's technology.
2. Messaging will transform the enterprise space
Messaging platforms are already massive in the consumer space (e.g., Facebook, WhatsApp, SMS). The enterprise space is heating up quickly. Launched in August 2013, Slack went from zero to over 800,000 daily users. It’s now worth around four billion. In addition to the core chat capability, there are hundreds of integrations and bots on Slack. These bots enable users to conduct polls, manage projects, and receive analytics reports from the likes of Google Analytics, New Relics, and Mixpanel. At x.ai, we connected services like Github, Amazon Web Services, and Twitter to Slack. The team can access the latest social mentions, receive notifications for machines that finished training their machine learning models, and get reminders on best lunch spots.
Slack's tremendous traction is attracting serious competition and the race to penetrate the enterprise space with messaging has started. It will be interesting to see which companies achieve dominance in the end.
3. A new class of designers
I envision a new class of interface designers who specialize in dialog design. For example, we created a new role called “AI Interaction Designer” at x.ai. Their job is to embody the persona of Amy and Andrew, our AI agents. Staying true to these personas, they design and build the responses and actions that our users see and experience. At times, they utilize familiar tools like user research to further understand our users. Other times, we build custom tools and invent methodologies for them to optimize and create dialog elements. We are in the nascent phase of this field of design. AI Interaction Designers may very well be a required role for any company building an AI Agent in the future.