6 reasons why scheduling is a machine task

There are many things humans do better than machines. Scheduling meetings isn’t one of them. When we set out to build x.ai and fix the time and productivity suck that is meeting scheduling, we decided early on that our agent needed to be fully autonomous. Humans certainly play a vital part of labeling our data, building, supporting, selling, and improving everything that is x.ai, but the actual process of scheduling meetings as done by our intelligent agents, Amy and Andrew, is machine-driven.

Obviously, I’m a bit biased here. Let’s see if the six primary reasons that I believe scheduling meetings is a task for machines convince you as well.

1. Speed

Machines are fast, and getting faster. With practice, humans can get pretty quick at many tasks, but there is no Moore’s law for humans that dictates a doubling of our internal processing speeds every 18 months. Computers, on the other hand, can read and parse a volume of text in milliseconds that would take even the fastest human speed reader hours. Oh, and machines don’t need sleep, or food, or take bathroom breaks—they can operate 24/7/365 without complaint.

By taking humans out of the loop, intelligent agents like Amy and Andrew can be there for you literally whenever you need them, right when you need them.

2. Accuracy

It’s not that machines can’t make mistakes. They can. But they make them far less often than people while juggling much more data. By putting meeting scheduling fully in the hands of machines, Amy and Andrew can instantly and accurately recall your preferences, they can look up addresses, account for time zones, or remember special schedules. Humans can do all that too, of course, but they’re much more liable to make a mistake along the way (I’m sure you can remember at least one mistaken “reply all” moment and the ensuing mortification), and they won’t be able to perform as quickly. We might still be a few clicks away from machines overtaking some humans on this point. There’s no doubt in my mind that we’ll inevitably get there.

3. Skills

There’s a limit to the number of things a human can learn. Machines are only limited by their processing power—that thing which keeps doubling—or their storage capacity. That means machines can do things that humans could never feasibly do, like understand meeting invites in dozens of languages along with local mores and etiquette. Amy and Andrew can’t do that yet, but they could (and they will), and that’s what separates them from their human counterparts.

4. Price

Scaling a machine-only operation is way more cost-effective than scaling one that relies on people. Hardware is cheap and easily replaced, and as I mentioned before, it can operate around the clock. And machines can handle the requests of hundreds or thousands of people at once. Human-in-the-loop outsourcing operations aren’t even in the same ballpark. The efficiencies realized in having machines schedule meetings are so great that we’re able to offer a free version. No human-in-loop out-sourcing operations can EVER offer a free forever scheduling service.

5. Security

Machines are also more secure. We know that you trust your assistant with valuable, often proprietary information. Who you meet, your location and schedule, conference call access codes, etc. are all things your assistant might need to properly schedule meetings for you. Keeping that information safe and secure is another reason we believe machines are better suited to meeting scheduling.

Machines don’t gossip, they don’t make reply-all or auto-complete errors, and they can’t be socially engineered (there is not enough charm in the universe for someone to be able to persuade Amy to give them your schedule!). If you decide you no longer need a machine and “fire it,” your data remains just as secure as ever; when you part ways with a human you are, to a large degree, relying on their kindness, which can’t always be relied upon.

6. Network effects

Even if an individual human could could pull off the first five tasks on this list, the effect of running and optimizing the task as a network makes the transition to machines for meeting scheduling inevitable. The network effects possible by using machines to schedule meetings will unlock true scheduling nirvana.

When machines are talking to other machines to schedule your meetings, and all humans—even your guests—are out of the loop, everything becomes as fast, efficient, and painless as possible. For example, let’s say you have three meetings on three separate days uptown next week. Each will require about an hour of commuting, so scheduling them all on the same day would save you a lot of time. Right now, Amy is at the mercy of guests for whom she has no visibility into their calendars. In a network, Amy could sort through everyone’s calendars and figure out the best way to maximize everyone’s time, and minimize your commute.

Or take rescheduling, which is hugely painful when you don’t have an assistant. If we all have an AI assistant, then rescheduling pain disappears. Amy could reschedule a meeting 17 times and no one would even notice. Amy could move an unimportant coffee scheduled for sometime this Autumn, to any day in any month, and she can keep moving it around, to accommodate more important meetings. When time frees up that coffee lands (and sticks) on my calendar. No biggie for either party.

Tasks for humans

None of this is meant to sound anti-human. I work with dozens of amazing people every day building x.ai, and they’re among the smartest and most talented people I know. Amy and Andrew wouldn’t exist without them, and none of are us keen to live in the machine-only world of The Matrix any time soon.

But for specific tasks, the ones that drain productivity, provide no one with any enjoyment, and where computers clearly outperform humans, we should be moving toward a 100% machine-driven future. Meeting scheduling is one of those tasks.

This post originally appeared on Linkedin.


Want to hire Amy + Andrew? Start your free trial HERE