Bots-as-a-Service

If bots are the next big thing, then whatever makes bot-building easy and cross-platform will be huge.

Bots, or conversational interfaces, are the future of how we use apps and interact with companies. Instead of opening apps individually, where they compete for home screen real estate and our attention, we’ll talk to them from within whichever messaging app we’re already using.

E.g., instead of opening the Amazon app to search for and buy an e-book, you’ll just text Amazon, or send a message on Facebook, Kik, Whatsapp, Snapchat, Signal, Slack, Hangouts, or any of the various chat options you use.

Bots will increase our engagement with apps and companies and make those interactions more natural and easy. We’ll carry on asynchronous conversations all the time. At least that’s what everyone’s saying, and they’ve done so in far more thoughtful, detailed ways than I will here. Let’s assume they’re right. For recommended reading on this topic, see the bottom of this piece.

If bots are the next big thing, then whatever makes bot-building easy & cross-platform will be huge.

The takeaway:

Someone should build a kit that helps businesses build bots easily. There’s a massive opportunity for whoever creates a universal platform for building bots as a service.

To have the widest impact, a kit like this would have two characteristics:

  1. it’s open and universal; and
  2. it’s super easy and accessible for anyone to use.

Open and universal

This kit wouldn’t only help you build a bot for a single platform, like the Facebook Chat SDK or the Apple platform (although Slack has arguably had the greatest success so far with their API and hundreds of user-contributed bots). Why? Because users don’t have the mindset “I need to get an Uber; which messager app is compatible with the Uber bot?” They never will. Bots need to be where the user is, no matter which messaging app they’re on (as long as those apps are at least somewhat widely used).

It’s likely too difficult for anyone except experienced app developers to create and maintain bots for a handful of platforms. Instead we should be thinking open-source, with one bot deployment spanning myriad messaging apps.

Another piece of this universal bot landscape should be an open directory, sort of like Twitter’s publicly searchable accounts meeting Facebook’s messenger experience. Users should be able to find and chat with any existing company’s bot that could help them, and it should be instantaneous. Ideally this happens without the user installing anything additional or giving new permissions. I.e., once the user initiates a chat with a bot, the conversation is open for responses; once the user leaves, it’s closed. In other words, it’s more like a live customer service chat than a series of one-sided marketing communications.

Bottom line: bots need to be where the users already are. The whole point of conversational interfaces is that they’re convenient. Users won’t change their behavior to accomodate bots; it has to be the other way around.

Easy and accessible

Talent is a company’s most important — and limited — resource, especially developers. Making bot building easy and accessible to non-developers not only conserves developer time, but makes use of non-technical team members’ language-based talents.

Bots at this stage of evolution aren’t really AI; they’re logic. They’re “if this, then that” with a human voice layered on top. They’re streamlining common support cases, explaining frequently asked questions, scheduling appointments, and completing orders. With the right framework, any good writer should be able to “botify” a problem or a use case for a company.

Elizabeth Dwoskin at the Washington Post just wrote about this opportunity in her piece “The next hot job in Silicon Valley is for poets.” Her thesis: “writing for AI in Silicon Valley is becoming a hot commodity” for artistic types: the liberal arts majors, the screenwriters, the history majors, and the theater types. These people are providing the human voices in the bot landscape.

But to be as resource-efficient as possible, these writers need to be able to build their own bots without tech help, or at least very little. Think of the evolution in web design from hand-coding HTML websites to templatized Angelfire/Geocities sites to Wordpress to Squarespace and finally to PageCloud at the bleeding edge of web design accessibility for all.


The bot revolution is almost here, but to speed up its arrival and market penetration, there’s a gap for someone to fill this “bots as a service” need.

If you’re working on it, I want to talk to you. Find me on twitter or send me an email (sarah [at] accomplice [dot] co). Thanks to my Accomplice colleague TJ Mahony for spinning out this idea with me.

Further bot platform reading:

Medium: “The Messaging Landscape in 2016,” by Ben Eidelson

Medium: “Imagining MessageKit: Apple’s path to turning iMessage into a platform,” by Matt Galligan

TechCrunch: “Facebook’s Messenger Bot Store could be the most important launch since the App Store,” by Tom Hadfield

Wired: “Slack Is Overrun With Bots. Friendly, Wonderful Bots,” by Molly McHugh

The Verge: “Kik launches a Bot Shop for its messaging platform,” by Ken Yeung

Sarah A. Downey is a Principal at Accomplice (FKA Atlas Venture). Accomplice invests in tech entrepreneurs and creates companies at the earliest stages. Our partners are Jeff Fagnan, Chris Lynch, and Ryan Moore. Located in Boston.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.