On March 2023 , Microsoft generative AI called Copilot into Microsoft 365. This technology is potentially as big a game changer for office productivity as Microsoft Office initially was.<\/p>\n\n\n\n
When I started in the technology business, secretaries were commonplace. I even shared one for a while, and it was a fantastic experience. Sadly, I was at the tail end of such assistance. It was incredible to have an employee whose job it was to have my back and who would supplement and assure my work.<\/p>\n\n\n\n
As products like Office advanced, companies concluded that secretaries were no longer necessary, and executives were given tools that supposedly automated the secretary\u2019s role. They didn\u2019t, though, because those tools weren\u2019t smart.<\/p>\n\n\n\n
The automated tools made us more productive but didn\u2019t fix the shortcomings we had like human assistants could, such as writing a good letter, maintaining correspondence, and keeping our inability to manage our temper a secret (something social media blew up).<\/p>\n\n\n\n
This initial use of generative AI in Copilot, while far smarter than Office without it, won\u2019t initially overcome all of our shortcomings, but it will get smarter with time. I believe it will evolve to embrace the old secretary role fully. But initially, it will increase our productivity by turning our ideas and concepts into completed works.<\/p>\n\n\n\n
We need to keep in mind that this technology, already in its 4th generation, is still very young and immature. We can\u2019t depend on it like we would an experienced human assistant, which means it is more about productive quantity than quality. We\u2019ll need to both learn this tool and be aware of its weakness.<\/p>\n\n\n\n
While it\u2019s far more intelligent than Office was, it isn\u2019t as smart as a good human assistant. It can\u2019t pick up on context, it won\u2019t always understand the impact of tone, and it won\u2019t be able to protect you from a mistake.<\/p>\n\n\n\n
The tool isn\u2019t yet a peer, though I expect it will evolve to become one. In a way, it is like a young child that looks to their parents as if they are gods and doesn\u2019t question anything those parents say. It doesn\u2019t understand nuance or perceive that those parents aren\u2019t perfect.<\/p>\n\n\n\n
This tool will not only do what you tell it to do, even if that isn\u2019t in your best interest, but it will also take the data it is fed, fact or fiction, and treat it like truth. As a result, you will need to supply oversight and recognize that it can do things at a speed that could be dangerous.<\/p>\n\n\n\n
One of the things that makes Twitter dangerous is that you can see something and react to it without thinking through the implications of your reaction. You see a funny post and repost it without realizing that the post is sexist or racist, and your career is suddenly done.<\/p>\n\n\n\n
You get angry and post your feelings about something your CEO has done, you misread a post or just poorly word a response to a post and, again, your career is over. The speed at which you can present on Twitter makes it one of the biggest dangers to your future.<\/p>\n\n\n\n
Copilot is initially a speed tool. If you are upset, it will draft an email that could end your career, so you must use Twitter-like restraint to keep from sending it. One of the advantages of writing something yourself is that it gives you time to think through what you are doing and maybe rethink it.<\/p>\n\n\n\n
I\u2019m not suggesting the tool is bad, but like any tool that initially focuses more on speed than quality, the risk of acting without thinking increases significantly. Fortunately, for those of us who haven\u2019t lost our jobs due to a poorly thought-through Tweet (my fix was to stay off Twitter mostly), we know how to pause and think before acting.<\/p>\n\n\n\n
Still, as awesome as it is (and it is awesome!), we need to realize this new technology is in danger of acting first, causing us to regret that action later.<\/p>\n\n\n\n
After reading that heading, I can practically hear a bunch of folks throwing things at me and screaming, \u201cNO!\u201d While the concept of Clippy was good, the technology and execution were admittedly horrid. But as Copilot evolves, it will gain the ability to help you perform certain tasks better because it will learn how best to accomplish them. In other words, even initially, Copilot is massively more helpful than Clippy could be.<\/p>\n\n\n\n
Where I\u2019m particularly interested in seeing this tool work is with PowerPoint. Most of us, including me, suck at PowerPoint, which is ironic since my initial success as an executive was creating great presentation slides. But as time went by, I got lazy and started creating word-heavy slides that didn\u2019t properly use the visual aspects of that tool.<\/p>\n\n\n\n
Copilot for PowerPoint takes that text approach and builds visually compelling slides from it, so I can still be lazy and get far more visually captivating presentations. It fixes the bad practice that many of us have developed over the years and creates a forcing function where we can iterate the content until the slides tell the story we want told.<\/p>\n\n\n\n
To me, this version of Copilot provides the greatest benefit with the least potential risk. Given my degrading presentation history, this could be life-changing. This video shows Copilot in action.<\/p>\n\n\n\n