As a consultant, I participate in a lot of meetings. While I do enjoy meeting with people, I don’t love the task of scheduling meetings. The inevitable email ping pong that happens as you go back-and-forth about time and place can be time-consuming. How about this time? How about this day at this time? How about this time? Often, after finally working out the details, a meeting cancels or has to reschedule. This can all lead to a lot of time lost to administrative work.
I found out about X.AI from a twitter connection, and immediately signed up for the beta program. After waiting for a year, I was finally able to get my AI assistant…and the experiment began.
X.AI is a company that developed an AI agent that schedules meetings. The agent is named Amy Ingram or Andrew Ingram (they both work the same, you can choose which one to use). While the app does have a web interface for setting up the service and specifying a few preferences, when I first began using the AI, interactions happened in natural language over email.
To use this AI, first, get an X.AI account. Second, connect it to your email (currently works with Office 365, Gmail, Microsoft Exchange). Once you are set up, whenever you get an email request for a meeting, copy Andrew@X.ai. At this point, the AI takes over by having a conversation with the other party, using natural English language to negotiate the time and place, based on your calendar availability and specified preferences. Because I have specified standard preferences for things like web meetings, lunch meetings, and coffee meetings, the AI knows how to set up each type of meeting — taking care of those details.
What Worked for Me?
This AI app has saved me over 40 hours a year, scheduling over 400 meetings without my intervention. The company, X.AI, continues to add skills to the AI, releasing new features, settings, and interactions. The app has evolved beyond email-only interactions, to include a mobile app, chrome browser plugins, slack commands, and meeting types with publishable links (similar to Calendly or other scheduling tools).
The X.AI app can successfully organize meetings with up to five participants. For example, one meeting that Andrew set up for me included people from five different companies. Over 31 emails were involved by the time the meeting was scheduled. I didn’t’ have to be involved in any of that detail.
Positive customer experience is an important consideration whenever you are implementing technology that is customer-facing. My clients and collaboration partners embraced this experience. Some even raved about it, “Amy is soooooooooooo cool” or “That’s the best thing that ever happened to me over email!” One client said, “You have an assistant? Am I paying you too much?” Now, he expects to interact with the bot and is totally comfortable with it. People seem to adjust to this app quickly. As far as my customer base is concerned, it appeals to all ages and technical comfort levels. Maybe because it is so familiar: just an email conversation.
X.AI was easy to adopt. Since I started using the AI in 2016, I’ve found it incredibly easy to adjust to a new habit in my workflow. I had to remember to copy Amy@X.ai or Andrew@x.ai on meeting requests, but after mastering that habit, it became a normal part of my daily work. After the first few meetings, I was hooked and decided I wouldn’t go back to doing this manually.
The AI has a narrow skillset. It does one thing very well: schedule meetings. It won’t make the dinner reservation related to a scheduled meeting. Additionally, some meeting tasks require human judgment. For example, if you like to stack meetings of a certain type, the AI doesn’t do that…yet. It also doesn’t schedule recurring meetings.
Some scenarios are too complex for the app. In the cases where the AI doesn’t understand the conversation in an email interaction, it will ask me for instructions. The AI understands English, but as we found in one experiment, it doesn’t yet understand Australian slang.
Anyone who has done business with me over the past three years has probably encountered Andrew or Amy, the AIs that help me with meeting scheduling. With this AI app, I am beyond experimentation. It became a normal part of my workflow within the first month and has maintained that status for nearly three years.
Along the way, I learned that it does take some time to learn how to collaborate with clear communication, and to trust the AI to do what it is intended to do. When I began using the app, I scheduled meetings with my husband, so that I could test what the experience would be like for the other party. Once I realized what the experience was like on both sides, I was more comfortable trusting it, and I began using it with my customers.
Today, I sometimes forget that Amy and Andrew are bots, and catch myself beginning to write “thank you” messages to an AI, because it occasionally does things that seem “thoughtful”, such as reminding me of an upcoming meeting that I scheduled months ago, to confirm that the details are still set.
How could this type of AI application help you?
This AI application is a practical use of natural language processing and machine learning. It has a continually evolving set of skills, features, and user interactions completely focused on the business problem, “How might we make the job of scheduling meetings as easy as possible?” A lot of companies are adopting the technology, and it’s a practical consideration for any professional who needs to schedule a lot of meetings, calls, lunches, dinners and does not have the luxury of an executive assistant. With the ability to handle meeting tasks with ease, it’s a practical tool for any professional’s productivity toolkit.
To view short presentations on each of these experiments, see my video series Everyday AI.