Posey's Tips & Tricks

The Practicality of Conversing with Copilot, Part 1

Microsoft’s push to add real-time speech, dictation and conversational responses to Copilot revives long-standing questions about voice-driven productivity, drawing on lessons from legacy tools to assess where verbal interaction is likely to help -- and where it may fall short.

• By Brien Posey
• 01/21/2026

At Microsoft Ignite last year, Microsoft spent quite a bit of time talking about all of the speech interaction capabilities that are being integrated into Copilot -- particularly Microsoft 365 Copilot. Not surprisingly, many have been asking whether these new capabilities are little more than a gimmick, or if they have practical, real-world use. While it's too early to say for sure, my guess is that verbally interacting with Copilot will be great for some tasks, but not so good other tasks. Interestingly, I am basing this opinion on real world experiences involving similar, but legacy technologies.

So before I tell you how my past experiences have shaped my opinion about what we can expect from Copilot, let me give you a quick rundown of some of the voice features that Microsoft is adding to Microsoft 365 Copilot. You can access the full list of features here.

One of the core capabilities that Microsoft is offering through Microsoft 365 Copilot is dictation. You will be able to use your voice to write email messages, create documents and that sort of thing. It is worth noting, however, that Microsoft Word's native dictation feature predates Copilot.

A second capability that Microsoft is offering through Copilot is the ability to read text aloud. Again, Microsoft Word has included a Read Aloud feature for many years, but Microsoft has definitely made some improvements. Whereas the old Read Aloud feature produced robotic sounding speech, the new Read Aloud feature leverages Microsoft's Ava voice, which sounds far more natural than what was available in the past.

Of course, Copilot can do more than just verbally read a Microsoft Word document. When you speak to Copilot, it can verbally respond to you in real time, making for a completely interactive experience. While there are potential productivity benefits associated with such interactions (more on that later), there are some other advantages to being able to verbally interact with Copilot. As an example, Microsoft has created a tool that you can use to practice for a job interview so that you hopefully won't get tripped up by tough questions. Copilot can also help you to rehearse presentations.

So now that I have spent some time talking about some of Copilot's basic verbal capabilities, I want to take a step back and tell you about my own experiences using verbal inputs and text to speech capabilities. As I do, I want to talk about which types of Copilot interactions will likely be similar to my past experiences, as well as how Copilot is most likely to improve on these past experiences.

I have actually been experimenting with computer-based speech since I was a kid back in the '80s (this isn't my video, but it's what I used back then). However, the vast majority of my verbal interactions with a PC stem from my long term use of a program called Dragon Naturally Speaking .

I used Dragon Naturally Speaking for many years, until Windows compatibility issues eventually forced me to stop. For those who might not be familiar with Dragon, it was a software application used for dictation. As a professional author, I found that being able to dictate my books and blog posts often helped me to produce content more quickly. I also found that dictating kept me more focused on the task at hand and that my mind wasn't as prone to wandering the way that it so often does when I type a blog post.

In addition to being a dictation tool, Dragon Naturally Speaking supported text to speech (read aloud) functionality and it could also be used to perform other types of common tasks, such as launching an application, composing and sending an email, and that sort of thing. These are all capabilities that we find in Microsoft Copilot today!

I am going to delve into a discussion of lessons learned in Part 2. Before I do however, I wanted to point out two very key differences between the old Dragon Naturally Speaking software and Microsoft Copilot.

First, at the time that I was using Dragon Naturally Speaking, I knew quite a few other authors who had also tried using it, but I don't think any of them were able to make it work well enough to make it practical for use in a production environment. I suspect that the reason why ultimately came down to training. I spent an insane amount of time training Dragon to recognize my voice, building a custom vocabulary of tech related terms that I used frequently, and providing Dragon with thousands of articles that I had written so that it could learn my writing style. Hence, my experience was that Dragon worked extremely well, but I had to invest many hours into training the software before it ever got to that point. Conversely, Copilot is ready to go on day one, with no training required (although Microsoft concedes that Copilot may have trouble with certain accents or speech patterns).

Another major difference between Microsoft Copilot and Dragon Naturally Speaking is that Copilot is not purely a speech to text tool – it's speech to intent. The old Dragon software that I used typed EXACTLY what I said, even if what I said was wrong or unrefined. Copilot lets you get by with messy rough speech or messy ideas. Based on what I have seen so far, Copilot will clean up some of it on the fly (such as removing unwanted umms and uhhs), and it can help you to clean up the rest later on.

So now that I have given you an overview of both Copilot and the old Dragon software, I want to continue the discussion in Part 2 by talking about some of the lessons that I have learned in the past, where I think Copilot will shine, and where I think that verbal Copilot interactions just won't help.