Why we’re excited about conversational computing

Bots are taking over the web.

You interact with bots just like you would talk to your friends. Ask them a question, and they’ll reply. Give them a command, and they’ll go off and do it.

All without having to install a new app. Sometimes you don’t even have to create a new account.

Natural to use

This was also the promise of web applications: you could run any application you wanted for any purpose, without installing anything other than a web browser. Web apps took off because workers could use software without having to go through their IT departments. Even if your IT department wasn’t an obstacle, there was no need to have a specific operating system. Software just worked, across all your devices, with no technical obstacles.

Bots have all of these properties, too — and they have a consistent interface. Whether you have four chatbots installed or forty, you never have to learn a new set of menus or get your head around a new way of working. You just chat.

In a seminal piece last month, Chris Messina (inventor of the hashtag and open web pioneer) wrote about what he called the rise of conversational commerce:

No longer do you need to convince users to “download and install” an app — they can just invite a bot to a conversation and interact with it [eventually] like they would a person. Zero barriers to adoption, with minimal risk to the user.

(Mobile apps, by the way, are turning into a losing game: most users drop off in the first three days, and the vast majority of smartphone users never download any new apps.)

Bots at work

When XOXCO released Howdy, Ben Brown wrote:

Messaging and notifications. Software automation. Always-on connected devices. Machine intelligence. These elements can be combined to form a very interesting new type of application: the digital coworker — a piece of software that works along side you at your job and participates in the day to day activities of your company as an active and engaged member of the team.

Howdy allows you to ask questions to your team and turns them into a report to help you to make faster decisions. We use it at Known to power standup meetings, but it could also be used to take lunch orders or ask custom questions.

XOXCO raised $1.5m for Howdy, and hired the novelist Neal Pollack to give their bot a unique voice. For me, this is perhaps the most interesting aspect of bot design, and the bit I’m most excited to play with: character development. If you’re building a virtual colleague, it had better be an interesting one.

Bots in the media

While announcing his resignation, Owen Thomas, formerly the editor of ReadWrite, noted that:

There’s an amazing opportunity ahead of us to redefine how newsrooms work, driven by data, assisted by artificially intelligent bots, and delivering narrative experiences through new mediums such as messaging apps.

Indeed, Ditherati, his content site, is “an experiment in short-form content delivered via messaging bots”. Imagine having a friend who sends you interesting news stories from time to time.

I am interested in the potential of telling a story conversationally. Stories are how we interpret the world, and every relationship we have has a narrative. Again, the ability for writers and artists to craft a human story is suddenly much more important in a bot-led application universe.

More pragmatically, there’s also huge potential for bots to be useful colleagues in working newsrooms.

Newsrooms are busy, hectic places, constantly working under tight deadlines. Fast answers are vital. What if you could ask a bot:

• Where is [journalist’s name]?
• Did [photographer’s name] submit their piece?
• What are we missing for tonight’s newscast?
• What is our lead story?

What if we could ask it to let us know when [journalist] files their story? What if we didn’t have to ask it?

Suddenly every human in the newsroom is freed up to do specialist work, knowing that air traffic control is taken care of.

A universe of bots

It’s obvious that there are a million different use cases for this kind of supportive conversational computing. There are lots of repetitive, important tasks we do every day. In a startup, we’re constantly checking what we call KPIs: Key Performance Indicators like monthly active users and revenue growth. Being able to ask a bot to track those for us saves us time, and holds us accountable.

Lots of different kinds of applications are possible. There’s a universe of different contexts to support, data providers to integrate with, and services to provide.

In all of these situations, you’re always having a conversation with information. For the first time, the bot paradigm makes this explicit.

What does this have to do with Known?

So far, most bots have been fairly closed: one bot per application, with very little ability to customize. Meanwhile, every organization has a different context, and a different set of data that they need to have a conversation with.

One of the key benefits of a chat interface is that no matter which bot you’re talking to, the output will always be in more or less the same format. A stock ticker bot will tell you a number; so will a comparison shopping bot, a business intelligence bot, a Wolfram Alpha bot, and so on. Suddenly the data from every application on the web can be combined as easily as talking to a friend. We think this will be particularly useful for small teams of people who need quick answers.

It turns out there are lots of ways we can help you have a digital colleague that is yours.

We’ve been thinking about bots a lot, and we’re building something new, that is very different to anything else out there. And we can’t wait to share it with you.

You should follow Known on Twitter: @withknown. If you enjoyed this piece, click on the heart below.

Originally published at stream.withknown.com on February 2, 2016. Robot photo taken by Justin Morgan.