Bots are undeniably at the top of the techie mind these days. Wherever I turn online, I seem to run into blog posts about conversational UX, invisible software, and bots in general. VCs are pumping up bots in their conversations as well- looking to get in on the action. For example, Aleph VC recently started a bot challenge, because they believe that bots residing on messaging platforms will change most of human computer application programming mode.
I hope that this post will help you better understand what bots are, how bots can be used and how bots can contribute to your business. Before we dive into the bot applications of today, I’ll cover some of the historical contact we’ve had with bots over the last 20 years.
My personal history with Bots
I started playing around with bots approximately 20 years ago. It began long before Slack existed. Long before Siri, Cortana, and the others members of the bot gang took over our lives, there were bots we interacted with- unbeknownst to the fact that they were actually bots. The only problem was that I was 10 years old, so I just played around with them, had some fun and didn’t really think about their business potential. The first cool bot I encountered was a relatively sophisticated one : Dr. Sbaitso, an artificial intelligence speech synthesis program released in 1992 by Creative Labs for MS DOS-based personal computers. (Wikipedia)
A few years after, Microsoft started using their own new bot, which became quite famous. Clippy, an animated paper clip with great round googlie eyes, was the Microsoft ‘97 assistant available for everyone using Microsoft office. Clippy is arguably one of the most famous bots for those of us who are 25 years and older at this time. My wife, who previously worked for Microsoft, told me that she thought Clippy was really cute, though she couldn’t remember actually using his assistance. In addition to Clippy, Microsoft released another feature/software called Microsoft Agent, which was a collection of different bots that use speech to text technology. The most recognisable of them was “Merlin”.
How to build good bots
As can be learned from history, bots require a certain degree of human attributes, for humans to naturally interact with them. In the world of bots, personality is the new UX. When we interact with humans we usually prefer to interact with people who are interesting, funny and even charming. The same applies to our interaction with Bots. The way to build a successful bot is basically to build a personality.
Companies like X.ai, Claralabs and Howdy build office assistant bots. These companies create their artificial bots with the help of performers and writers from performing arts backgrounds, in order to design appealing personalities for their bots. I recommend reading this FastCompany article by John Pavlus The Next Phase Of UX: Designing Chatbot Personalities where you can learn more about some exciting new bot personalities that work you can within Slack’s platform.
Slack = the bot platform
Two weeks ago, I had the chance to hear a lecture from Amir Shevat, Director of Developer Relations at Slack. The lecture was hosted at a local bar in Tel Aviv, in conjunction with The Junction accelerator.
Amir introduced the Israeli audience to the bot revolution taking over Silicon Valley. He explained how Slack provides a platform for businesses to leverage their activity by building communication bots on top of Slack. You can read more about the conversational UX revolution in Amir’s blog post. If bots still sound too ambiguous to you, let me break it down.
On a high level, the idea of bots is fairly simple- and shares much in common with Slack’s motto:
“To make people’s lives simpler, more pleasant, and more productive.”
A bot is basically a software that interacts with the user using natural language. As a user, you can interact with your preferred bot using natural language without the need to learn a specific graphical interface. If you are a product manager of a Saas solution, Slack is a good place to start experimenting with bots.
Slack provides several interfaces for building bots. The bots can focus on one or more of the following actions;
- Notifications. For example, if a server crashes a bot can alert you by saying “Hey, Magma server is down. What should I do?” Then a conversation can start around that topic in the right channel.
- Slash command. Slack provides an easy way to invoke specific actions on your service and quickly return the result to the user. A slash command sends a specific url to your server and returns the message from the server directly to the Slack user. Using the slash command allows users to access apps without logging/installing the apps. Having this kind of embedded integration saves money that would usually be spent on distribution. For example, having the ability of getting a Lyft ride without installing the app provides a way of distribution that is faster and cheaper for Lyft, since the user did not need to go to the app store, download, install, login and only then actually use the service. With a bot, the user is able to access the application’s service from the Spack entry point.
Bots will disrupt the productivity market
Slack focuses on productivity bots. The bot goal is to replace old school corporate tools that manage processes such as PTO, timesheets, travel requests, support tickets, and other administrative tools. You can find a list of interesting bots at slackbotlist.com. In two years’ time, I project that many more people will be using bots for productivity assistance. A popular example of an assistant bot is Amy Ingram by X.ai. Amy is an office assistant bot that saves precious time by coordinating meetings for you. Amy does this by sending and receiving emails, until all parties are happy with the time and place of the meeting. I recommened listening to X.ai CEO and founder, Dennis Mortensen’s interview in the “This is Product Management” podcast.
There are many other examples of bots that use conversational UX in Slack, and one of my favorites is Envoy. Envoy is a startup that greets guests and registers their arrival for office meetings. Through the Envoy slack command, you can send a meeting invite to a guest. When the guest arrives, you’ll receive a notification of their arrival- straight into your slack channel.
Slack recently partnered with howdy, a bot development kit, to build an open source framework for building bots. The idea is to create a solution that can facilitate complex conversations. If you look at the average corporate environment today, people work with an average of 14 different service owners in order to get their jobs done. This can include your company travel agent, someone in the benefits department, invoice processing, expenses, and so on. The Slack idea, is to create bots that can help you with each of these 14 streams of work. We’re able to cognitively process and tell the difference between this amount of people, so mapping this to virtual assistants should be possible for us to process and use for our advantage. In the near future we’ll have bots for food delivery, a bot for PTO reporting, a bot for timesheet reporting, and much more.
Slack is working with with IBM Watson in order to deepen their bot’s sentiment analysis. It’s important for the bot to understand when it pisses off the user. Shevat joked in a tongue in cheek way- that this may be even better than a husband’s ability to understand why he annoyed his wife! Friendly bots that can understand sentiment are an important step for a brighter bot future.
The future of bots
Bots are undoubtedly going to become a significant part of our lives. I found the vision of Dennis Mortensen, CEO and founder of X.ai, especially inspiring. Mortensen shared his view of the future of invisible software in a podcast. Imagine you email Amy Ingram, your office assistant bot to schedule a meeting for you after you come back from your business trip in San Francisco. You add your travel agent bot Robert to the cc and let Amy coordinate with Robert the details of the meeting. After introducing the two, you unsubscribe from the email thread and Amy and Robert work together all of the details. Amy and Robert are two softwares, that communicate with each other using natural language. And the kicker? They’re not aware that they are both software. The most amazing part is that future bot software will connect using natural language rather than APIs. Using natural language processing to connect software will make software integration significantly cheaper, faster, and more accessible.
Call to action
If you want to take the your first steps with bots, and not sure where to start, here are a few useful action you take:
- Start experimenting with Slack bots
- If you are not using Slack yet, it’s a good time to start.
- Install Nikabot, a project management bot that asks team members what they were working on.
- Register to other bot services such as Amy ingram by X.ai, Clara by Claralabs, M messenger by facebook, and challenge your Siri/Google Now/Cortana more often than usual.
- Explore the slack app directory and look for what is NOT there.
- Get familiar with Howdy’s botkit to learn how to start building your own bot.
- Push yourself out of the comfort zone and sign up for the The Aleph Bot Challenge #1 (If your live in Israel)
If you work on an amazing bot, please feel free to share with me 🙂