AI Agent & Copilot Podcast: ChatGPT 5 in Copilot

In this episode of the AI Agent & Copilot Podcast, Shawn Dorward reviews the integration of ChatGPT 5 in Microsoft Copilot and how users should go deeper with adoption to experience the most benefits and get the most value out of the ever-evolving technology.

Key Takeaways

• Constant development: After trying the tools of ChatGPT 5, Shawn landed on a two-part opinion. First, it’s impressive yet intimidating how fast AI is changing and evolving. The speed and pace of improvements will continue, leading to the development of stronger, faster, more powerful, and more intuitive types of AI models.
• Buzz and benefits: The other part of the excitement, Shawn suggests, is that there’s sometimes buzz for the sake of buzz. Much of what users get out of these tools is based on what’s being prompted. For instance, it self-routes. However, is that really to the benefit of the end user or is it to the benefit of the demand that’s put on the engines that run the AI?
• Diving deeper: The evolution of AI is continuing, and those who are using it more deeply will see the benefits. If you’re just using it for tasks like summarizing meetings or chats, drafting emails and responses, or creating agenda notes, you’re not going to see much of a difference. This new iteration is faster with deeper reasoning prompts and inquiries. Shawn challenges users to try going deeper with AI use and adoption to experience the benefits of this release and future updates.
• Iterative accuracy: Shawn uses a process called “iterative accuracy.” He asks small questions, then builds on them. When he’s satisfied with the results, he’s able to compile multiple thoughts, components, or ideas into a single prompt to get closer to his ideal output. Now, with ChatGPT 5, users don’t have to be as much of a prompt engineer.
• Why leaders should care: If your end users aren’t seeing the value, then you should understand that and encourage that user adoption to go bigger. “What that tells me is that folks probably aren’t trying to stress or test the model hard enough to improve what they’re doing on the daily,” Shawn notes. “If they’re not using it, in the end, for the organization stuff, they’re probably not using it for the out of organization stuff that really is why you want to use AI.”

Stream the audio version of this episode here:

---