In my new job as a bot-maker and product manager at Quartz, I've been asked lots to share my thots about bots.
For a deep dive about conversational interfaces and what they mean to journalism (according to me), you can check out this Nieman Lab interview.
If you'd like a quicker scan, here are some questions and answers I prepared ahead of a panel about bots organized by the New York City chapter of the Online News Association. Here are those notes, and some links, too:
What should you consider before you start working on a bot?
- You're entering uncharted territory! Have fun, explore, try new things.
- There are no obvious places to find your bot. Bot makers talk about "discoverability" of bots, which is pretty problematic everywhere at this point.
- So consider building where people already are interacting with you.
- Make a "worker bee" bot that does a particular task well -- not a "know it all" bot like Siri or Alexa.
- Is the information being exchanged sensitive? If so, think carefully. Making bots often means sending conversations through one or more 3rd-party services.
What are the specific design questions you need to keep in mind?
Platform The conversation has to take place somewhere. (Alexa, Messenger, Twitter, your own chat window.)
Natural Language Processing Spoken or written words need to be analyzed to determine the human's intent and pick out the key components needed to provide an answer. "Do I need an umbrella in Paris tomorrow?"
- intent: weather forecast
- key components needed:
- location: Paris (probably France)
- time: March 16, 2017 (probably in Paris time)
Action Use the intent and key components to accomplish the task: go get a weather forecast, turn off the lights in the living room.
Reply Craft and send a response that has voice/character that fits you, your brand or your service.
What are the major differences between audio and messaging-based bots?
- In voice systems, there's the added step (and challenge) that the human's speech has to be recognized correctly.
- Messaging bots can also provide images, buttons, links. Voice systems can't.
Do I need to hire developers to work on my bot?
- If you want to play and tinker, no. Even bots you'll use internally (like in Slack) can be made DIY. It can be challenging and confusing, but totally doable and often free. You can even make "skills" for your home Alexa or FB messenger.
- If you want to make your bot public -- publishing it on a platform like FB Messenger or Alexa -- then yes. These require approval by the companies you are using, and you probably can't leap all of the hurdles without some coding experience.
Are there specific types of stories that work well in bot format?
We put all sorts of Quartz stories into our bot-like app. The key, for us, is to keep it brief, smart, plaful and well-written. There is a team of writers working app, and they try hard to keep a consistent voice.
Do I need to partner with specific platforms or companies (Amazon, Messenger, Kik, etc) to get started or can I DIY?
Even with those partners, you can do it DIY -- for a bot that only you can play with. A good example is this very clear Alexa getting-started guide.
Once you want to make it public, though, you have a bigger hill to climb. Your bot will need to be approved by the platform people -- and that can be quite a process.
Slack bots don't require approval, and are also great to experiment with (on your co-workers!)
How can I tell if my bot is working? What do you consider success?
Some possible metrics:
- Is the bot understanding the requests?
- Are people getting the information they want?
- Are users having fun?
- Does anyone come back?
How can internal bots (non-user facing) help in the newsroom?
A lot. From simply watching channels for the words "wifi password" (and responding accordingly) to more sophisticated tricks.
How can I get started on a bot today?
- Get started building a Slack bot with this beginner's guide.
- Here's a (slightly dated but still valid) demo of using IFTTT to make an automated tip bot, watching Federal court feeds.