The integration of the Copilot AI in Windows has been lackluster up to this point, if you ask me. It felt rushed and Microsoft is still scrambling to improve its use for users. The same is true for any other AI that is available as an application.
While there are plenty of options to use AI on Windows, the tools do not really offer much in terms of usability that is not also available elsewhere.
Microsoft seems to have forgotten a key ingredient regarding the successful launch of a new product: the benefit for the user.
The initial Copilot offered some of that, as it supported the tiniest amount of Windows-specific control options. It demonstrated what could have been, even though the implementation itself was cumbersome. Even Microsoft did not publish the full list of commands that Copilot supported back then.
Microsoft changed its strategy suddenly and replaced the integration with a copy of the web version. Windows-specific options were gone and the clear signal was that they would not be back anytime soon.
Two new Copilot for Windows features were shown this week. While I would not call them gamechangers, some Windows users may find them useful.
Number 1: Gaming
Microsoft describes Copilot for Gaming as “the ultimate gaming sidekick that helps players get to play faster, sharpen skills with expert coaching, and enjoy a more social gaming experience”.
The key feature is helpful coaching, not game recommendations, maintaining connections with friends, or seamless game setups.
One particularly exciting area which Kardar demonstrated on the podcast is a future feature of Copilot for Gaming: in-game assistance. In example scenarios, players were presented using natural language, where Copilot was able to provide situation-specific advice to assist players in overcoming challenges and improving their gameplay.
So, when you are stuck in a game currently, you may fire up a web browser to find a solution. Maybe it is a YouTube video or a tutorial that explains how to beat a particularly difficult boss in the game you are playing.
Clearly, it remains to be seen how useful this AI assistance is really going to be, but it is finally something specific and not just a feature that everyone is also offering.
Number 2: Press to talk
The main mode of interaction with AI is typing currently. Microsoft, Google, and other major players are evolving their AIs to allow different means of communication.
Talking offers advantages in some situations, for instance when you have your hands busy or do not want to leave an app or game that you are using at the moment.
Press to talk is a new feature of Copilot for Windows that is rolling out currently to Windows Insiders. Microsoft mapped the function to the Alt-Spacebar shortcut of Copilot. Press the two keys for two seconds to start a conversation with Copilot.
Ask anything over the microphone and you get an answer from the AI. Copilot ends the conversation automatically when there is no speaking for several seconds.
This may be useful in several scenarios, from gaming over other fullscreen applications to asking for contextual help without having to do a lot of typing.
Closing Words
Yes, these features may not persuade the majority of Windows users to give it a try or buy a new Copilot+ PC. What I like about them is that they introduce usefulness to Copilot when using Windows in particularly.
I expect more features like these to become available in the future.
What is your take on this? Do you see your “future you” using Copilot or another AI? Feel free to leave a comment down below.
I remain for the time being *very* cautious with integrated AI features, OS as browser wide, an attitude related to awareness regarding privacy settings set as opt-out, hence active by default.
My approach regarding services, digital as in life, has always been ‘if and when I choose to consult them’ and therefor any incentive, persuasive policy to adopt them (free of charge or not) generally initiates an irritation at best, a boycott at worst regarding the so-called ‘for you, for a better experience’ and you name it. Of course a product is to be publicized, people need to be informed of these but, as I see it, with raw and non- information such as, implicitly or explicitly defined as related to who I am, what I do and therefor what I should certainly be interested by. “Stop digging into my digital or not life, promote your product and forget me, if I ever need it I’ll know where to get it” is my credo. I detest proselytism and that includes commercial proselytism.
This is my stance as well, even though I test many AIs regularly for work. If you ask me, it is a question of usability and privacy. It needs to bring something to the tablet that improves something somehow, and do so without sacrificing my privacy.
@Martin, usability as well, indeed, usefulness. We see nowadays AI integrated everywhere, even when its pertinence is arguable; it’s like if the absence of AI would be relevant of an out-of-the-game product. Big words like hysteria (which have a strict definition) come to my mind. But some, many of us are cautious and I happen to encounter labels such as ‘AI-free’ : after the excitement of novelty come the times of rational thoughts, and those include indeed privacy and usability. The next in the row is the so-called ‘agentic AI’ : a most interesting article and video is that of ‘Signal President Meredith Whittaker warned Friday that agentic AI could come with a risk to user privacy.’ available at :
[https://techcrunch.com/2025/03/07/signal-president-meredith-whittaker-calls-out-agentic-ai-as-having-profound-security-and-privacy-issues/]
Really worth the read and sparing time to view the one hour video.
Maybe a sacrafice of privacy, but I had an MRI the other day; the radiologist issued his report, and I understood nearly nothing; the report didn’t really tell me what issues were evolving in the body. My GP gave me a super brief synopsis that I already knew.
A search online, and I found a site that reads medical records, MRIs included. A record can be anonymized. Copy and paste the “findings.” Enter. And the AI program gave me this wonderful explanation of what was going on, how it compared to a previous MRI several years ago, and what I should expect in the future–sort of a prognosis.
Even before the reading, it doesn’t claim to be 100% reliable and one should consult one’s healthcare provider before making any decisions, but it gave me such precise information that explained so many things, simple things that two doctors couldn’t put into normal language.
As of now, I see the medical field as most promising for the use of AI–gaming aside.
The sentence below is missing two words that makes it incorrect and misleading. You need to add “appear to” after the colon.
“Microsoft seems to have forgotten a key ingredient regarding the successful launch of a new product: the benefit for the user.”