Messaging bots won’t replace apps

One app (or bot) to rule them all?
One app (or bot) to rule them all?
Image: Reuters/Beck Diefenbach
We may earn a commission from links on this page.

Facebook’s long-expected addition of chatbots to Messenger recently confirms early signs already seen with SlackTelegram, and Kik: Chatbots are all the rage, and along with the shift to conversational commerce, represent one of the major trends in 2016. And while bots themselves are hardly a new idea (we saw them in IRC and SMS decades ago), there is a broad industry consensus that they’ll become broadly transformative. Some have even suggested they’ll eventually supplant our app-based ecosystem. While I agree bots represent a major opportunity, I also believe they won’t replace apps, but will rather live in symbiosis with them.

The walled garden challenge

As with WeChat and other early incumbents, Facebook Messenger remains a walled garden, controlling the users’ online experience while also owning and controlling their data. This remains a friction point for companies which want more control over both. Beyond that, most popular app categories (services, retail, games, social sharing apps) have an internal messaging feature which chatbots within walled garden messaging services can’t address. For those reasons, I’m concerned these chatbots miss an opportunity to evolve messaging beyond the current paradigm.

By contrast, I see an opportunity to turn each message into a bot and micro-app, with services intelligently surfaced and integrated into the overall messaging experience. Instead of having to shift from conversation to conversation based on the brand you’re interacting with, I believe the future lies in conversational AI, in which all our applications are seamlessly threaded into messaging.

Bots should unify multiple just-in-time services

The first iterations of messaging services that enable bots mainly focus on having a series of conversations with different agents. The user has a conversation with service A and another with service B. These conversations are completely separated from each other and usually single-task oriented.

Some might argue that this addresses the problem of having to download apps to access services, however it does not change much in terms of switching from service to service to complete a task. The only difference is the user stays in the context of the same app, benefitting the publisher of the messaging app, but not improving switching costs for users.

At the same time, companies using messaging bots also face the same discovery challenges they’re already dealing with in app stores. To counter this shift in power, I could see mobile OS providers opening up their notification center APIs a bit more  and enabling web applications to trigger push notifications with actions on mobile and desktop.

The big shift that bots in conversational UI bring is that many different services can sit behind the same UI, ready to be invoked when needed. In a world where the message is just a container, and the content can be a micro application, the conversation view serves as a timeline. The user can always go back and forward in time and access the services they’ve used in that context before, just by scrolling.

Apps define context

Messaging apps are an important step in the evolution of messaging tools for general communication. The work Facebook, WeChat, Apple, and others have done on messaging is fantastic and we all benefit from better tools for communication, even though not every bit of our communication will fit in them.

As messaging technology and platforms became more accessible, more and more vertical applications started to include messaging as part of their experience. Marketplaces connecting buyers and sellers, communities connecting people with similar interests, and businesses connecting with customers — all of these are examples of applications that can provide a much better, tailored experience for their target audiences than a horizontal messaging app can.

Most importantly, applications define the context of the experience and separate conversations into contextual branded spaces. I know that all the information I need when looking for places to stay can be found in AirBnb or HotelTonight, and that I can go to Trunk Club for everything regarding my fashion needs. When the context is set applications become the container of the context.

App notifications and in-app messages as service delivery mechanisms

Just as applications define context, the user expects content to match. The brand elements, such as notification icons, visuals in the product and details that differentiate the experience from every other similar provider out there.

Looking at bots in the context of in-app messaging, the app notifications and rich messages themselves become service delivery mechanisms that deliver the right service in the right context in the time needed to provide maximum value.

The beauty of this is that an application can be powered by many services that are surfaced through the same interface. For example in a conversational commerce scenario, a service can serve the list of sneakers (inventory items), another service collects the payment information, a third service collects the shipping info, and a fourth sends delivery notifications. Instead of doing this through multiple, separate chat bot message threads, all of this happens in the same interface while still enabling the providers of these services to really specialize at a single task across hundreds or thousands of applications.

Evolving AI through in-app messages

The ability to gather in-app conversational data is really important, especially as AI and messaging bots are trained to become smarter. But as noted, Facebook and other walled gardens will never allow you access to that data. At the same time, in-app messaging often generates conversational data these services typically miss.

In a world where there are specialized services for very specific tasks, surfaced through notifications and messages the “intelligence” sits in the cloud and can constantly improve, without major changes to the application layer. For the intelligence piece to evolve and improve the access to the right, contextualized data is key. This brings us back to the point of access to identity and conversational data.

May a billion bot channels bloom

Every messaging interface in any application, in either a messaging app addressing billions of people or an in-app messaging experience as part of an application, is a channel for humans and bots to interact. The big opportunity for service providers building bots and bot platforms lies in servicing users through all these interfaces, wherever the users are, at the right time in the right context.

While messaging bots may not replace our app ecosystem, it’s wonderful that Facebook has affirmed their importance. I’m excited to see how more and more channels emerge every day—a rich ecosystem of bots and apps, thriving together.