Last week, I attended Microsoft Ignite to discuss the AI and Copilot launches on Azure and Windows.
As I watched the presentation, I was transported back to the debut of Microsoft Bob in the 1990s and the first effort at building a digital assistant dubbed Clippy, neither of which matched expectations at the time. But Copilot can now do considerably more than those two previous products, and the power of the outcome, both within and outside of Microsoft, is remarkable.
Prior to the event, I met with Reply, a company that specializes in preparing businesses for Copilot and putting up measurements to validate the technology’s advantages. Reply was ecstatic about how much more productive they and their clients had become.
What is absolutely astounding is that just eight months ago, Copilot technology was in its infancy. It became generally available last week at leading companies in medical, agriculture, services, software, automotive, finance, and petrochemicals. Now, 70% of users say they don’t want to work without this feature.
We didn’t have AI yet. Clippy and Bob were scripted, command-oriented services, which implied that you could converse with your computer. They failed because they were unable to match that expectation.
Conversational AI’s entire nature is to be able to achieve what Bob and Clippy desired but couldn’t, which is to enable you to command the system as if you were talking to another person. This technology not only allows you to interact more effectively with your computer, but it also learns how you operate and can progressively undertake your repetitive activities, such as replying to emails, scheduling meetings, and even standing in for you during team meetings.
Resolving Steve Ballmer’s Issue
While watching the presentation, I was reminded of a meeting I had with my old buddy and former Microsoft CEO Steve Ballmer and how he struggled to prioritize client demands. His problem was that he had a large number of key clients, all of which were national or multinational corporations, and he was inundated with data, making it impossible for him to utilize consumer input to determine priorities efficiently.
Ballmer is one of the brightest guys I’ve ever met, especially when it comes to statistics, yet the situation was still beyond his ability.
Microsoft presented Copilot for Services at Ignite, where input from almost any client, not just the biggest, could be connected in real time. As a consequence of the prioritized list based on user-defined criteria, a customer-validated action plan was created.
If Ballmer had been able to utilize Copilot, he would have been significantly more successful personally, and Microsoft’s products would have been far more appealing throughout his time.
Ballmer’s dilemma was believed to be difficult to tackle before Copilot because the volume of data and the rate at which it changed rendered traditional analytic approaches ineffective.
AI now operates at machine speed. It can provide answers in a fraction of the time that traditional techniques do, implying that executives who are now doing moderately may use this technology to become superstars, and superstar CEOs (of whom there are very few) could use this technology to become invincible.
It really is a game-changer.
Reply’s Beta Copilot Experience
As I indicated in the introduction, I met with Reply the day before the conference started. Reply has been collaborating with a large number of clients to bring the beta version of this product into production. An update around three weeks ago damaged the product. As this example shows, deploying a beta product is always fraught with danger.
You’d think people would simply say, “OK, it’s in beta, and a patch will be available soon.” However, more than two-thirds of consumers said they couldn’t live without the product and had grown completely reliant on it. They desired and required an immediate repair, and they got it. I don’t remember ever seeing such grief over a bug in a beta product.
Initial response indicates a 10% to 30% boost in productivity.
What these statistics don’t show is that this improvement is the result of Copilot doing tedious administrative tasks that employees dislike, such as responding to emails, taking notes at meetings, creating agendas, mucking around with Excel, and trying to remember how to do pivot tables. Copilot was providing a lot of stuff for them in its beta form when customers were still getting to know the software.
I’ve never seen such a rapid transition from a beta product to what it is now in my 40 years in this profession. This week, let’s speak about it.
We’ll finish with my Product of the Week, the Fisker Ocean, an electric SUV that I recently bought and believe is the greatest you’ll be able to purchase before 2025. There are already 62,000 orders, but since they only create 2,000 every month, I’ll have to wait a little longer.
Bob and Clippy’s Legacy
What fascinates me about Bob and Clippy is how far ahead of their time Microsoft was with both. It was properly predicted that consumers want a computing experience in which the machine performs the majority of the work. Bob was an avatar-forward design that, in principle, could receive English orders and then execute what you requested. Clippy was an assistant who sat on top of Windows and performed the same functions as the avatar front-end.
Bob was both sad and humorous since it was designed to be used by those who were elderly and unable or unwilling to learn Windows. When polled, this demographic overwhelmingly favored the product. However, someone at Microsoft attempted to present it as next-generation Windows, and Windows users detested it.
Wharton professor Ethan Mollick, who has conducted considerable research on generative AI, said that after users get acquainted with tools like Copilot, productivity may improve by up to 80%, which is unprecedented. As impressive as Microsoft’s and Reply’s early statistics were, those figures might be ridiculous by this time next year.
Not only will workers who use this technology vastly improve, but the technology itself will progress at an unprecedented pace. Because that percentage was based just on user improvements and not the predicted gains of the underlying AI, the combination of user and AI improvements might surpass that 80%.
Microsoft’s Growing Advantage
I was involved in two controversies when I initially began in the profession. One was rather large, whereas the other was not.
The first occurred when a reporter found that IBM, which claimed that the new technology would result in enormous productivity gains, was not really deploying it. Then-CEO of IBM, Louis Gerstner, remarked something along the lines of, “Would you rather IBM prioritize themselves or their customers?” As if they were mutually incompatible options. That didn’t go over well.
Years later, I was in a meeting with Intel, which claimed that the productivity gains in its current technology justified its early adoption. But everyone from Intel in the room was using laptops that were more than five years old, which begged the question, “If the machines are that good, why aren’t your people using them?”
Since then, whenever a vendor touts a technology as having a significant impact on productivity, I question whether they are utilizing it. If they aren’t, I assume they’re lying.
Microsoft has completed its deployment of Copilot, and as a consequence, its workers will be significantly more knowledgeable about this technology than employees from any other organization. They will be able to display not just the entire power and potential of Copilot but also make Microsoft workers more valuable than their colleagues in terms of competitiveness.
Similarly to how people used to be questioned about whether they understood how to use Office, future workers will be quizzed about their knowledge of generative AI and Copilot. They will have an advantage over other applicants who have not mastered such talents. This HR situation is likely to apply to any firm that implements this technology.
Companies should reconsider their retention practices and guarantee that staff are compensated for their skills with Copilot, or they risk having them recruited away. If I were searching for work, I would make sure I understood Copilot and include that knowledge on my résumé or CV.
Finally, Windows is dead; long live Copilot.
My interpretation of this event is that, much as MS-DOS was declared dead when the Windows GUI was introduced and Windows 95 was released, Windows without Copilot is nearing its end. Microsoft will continue to closely integrate Copilot with future Windows versions, much as it did with the Windows GUI, which changed Microsoft’s MS-DOS platform.
I’m observing significant variations in futurists’ forecasts about “the singularity”—the point at which AI overcomes human control. Simultaneously, expectations for the arrival of artificial general intelligence, or AI capable of performing intellectual tasks similar to humans, are altering. More specialists are now indicating that these substantial advances, which were originally projected around 2050, may occur considerably sooner, around 2030.”
So, by the end of the decade, we’ll all need to be more adaptable to change because we’ll be up to our armpits with AI. In a few years, those AIs may be so far ahead of us that we’ll want to start saying better things about them just in case they look back and become offended by what we’ve been saying.
Perhaps we could stop referring to their intellect as “artificial.” How would you wish to be referred to as “artificially” intelligent?
I smashed my electric Jaguar I-Pace a few weeks ago, almost killing myself and fracturing my back—trust me, it hasn’t been enjoyable.
That occurrence prompted me to hunt for the greatest deal on a new electric vehicle in terms of price, performance, looks, and technological progress. I chose the Fisker Ocean, a moderately priced electric vehicle built by Henrik Fisker, who has developed some of the most beautiful automobiles I’ve ever seen.
The Fisker Ocean Extreme, which I ordered, has a 360-mile range, a 3.7-second 0-60 time, and 100 miles more range than my I-Pace. It’s also made in the same facility as my I-Pace, which has been the finest vehicle I’ve ever had.
This Fisker has several intriguing features, such as solar panels on the top that can provide up to 1,500 miles of travel every year. If you lose electricity in a distant location, the automobile will finally charge itself. There is a significant likelihood that by the time I receive my vehicle, it will be able to utilize the Tesla charging network directly; now, an adaptor is required.
It also boasts some entertaining amenities, such as a collapsible table for a laptop or eating on—one for the driver and one for the passenger—a power-pivoting center screen, and California mode, which opens all the windows, including the dog window in the rear, and the sunroof. We have three dogs, one of whom would adore the dog window.
The Fisker Ocean lacks air suspension, which I’ll miss, but it seems to have better accident-avoidance technology than my Jaguar (which failed during my accident), and it looks like one of the latest Range Rovers, which isn’t terrible at all.
My wife and I are already arguing over who will be the main driver of the new vehicle, despite the fact that I am already missing my terrible Jaguar I-Pace.
Based on my study, the Fisker Ocean is currently the best electric SUV under $100,000. It’s my Product of the Week because I put my money where my mouth is.
