A Little Less Conversation

Thoughts on Conversational Interfaces and Integrated Services

The past months have seen a lot of talk about the inevitability of conversational interfaces: Chris Messina declared 2016 to be the year they finally happen; Sam Lessin wrote about how this represent a fundamental shift; Quartz released their new conversational mobile experience. This post is an attempt to offer an alternative perspective on that discussion.


What’s Broken Today

There are many things broken about how we interact with phones and computers today. Just think about how much time we spend finding, installing and handling apps for various odd purposes. How often we are interrupted in what we’re doing to enter the same information over and over again: email addresses, personal details, and credit card numbers.

These services are not there when we need them.

What’s stranger still is how disconnected these services are from one another, despite the fact that we tend to use them together. After all, we decide whether to take a cab or the bus when we find out how far we’re going. We choose bars, restaurants and cafés to meet up at while planning dates with strangers and nights out with our friends.

These services do not understand how we use them.


The Proposed Solution

Conversational interfaces are often discussed as a possible solution to both of these problems. There will be no need for apps because you can just send a message to any service you want; no need for interfaces because you’ll talk to them as you do to your friends; and since the messaging platform already knows you, these services will know you too.

The Chinese messaging platform WeChat is an often mentioned example, and it’s easy to understand why. Over half a billion people use the service every month and the average revenue per user is seven times that of WhatsApp, the largest messaging platform in the world. But I wonder if the fact that it’s a messaging platform is really that important?

I’d argue that what’s interesting about WeChat isn’t messaging at all, but how it provides access to services you need, in the context you’re in, and makes it easier for you to interact with them. There is no need to install anything and all your personal information is already there. In essence, it’s about solving the problem of services being there when you need them.

It’s about being able to do what you need in the context you’re in.


An Alternative Perspective

The second problem is more challenging: How do we design services that better understand how we actually use them? Here is where I’d like to talk about integrated services, or contextually relevant services made available from within another service.

It could be being able to compare routes on CityMapper with the price of an Uber without leaving Google Maps. Browsing bars on Yelp and concerts on Songkick while planning a date on Tinder. Comparing schedules on Google Calendar with ticket prices from Skyscanner while planning a conference with your colleagues on Slack.

Experiences like that require us to consider not just the relationship between a service and the people using it, but also between the service and other services. It’s about identifying the moments a service is needed, what other services are used in those moments, and finding mutually beneficiary ways to integrate.

Because services become better when they can be used together.


Moving On

Ultimately, I’m talking about more than just integrating services. It’s about moving beyond the concept of apps and services as isolated entities. About unified user experiences where all the tools and services we need in a particular context are readily available.

In a way, perhaps this is the reason why the idea of conversational interfaces has proven so compelling, and my biggest gripe is that we’re talking too much about a specific how and not enough about the general why. Bringing other services into the context of messaging might be an interesting place to begin, but messaging is still just one context among many.

For now, perhaps the best place to start is to talk more about the problems we’re trying to solve. How we can improve the ways we interact with services, design them for how they are actually used, and make them available where and when they are needed.