AFAIK, Copilot (including Copilot for GCC) does not currently learn from individual user interactions in a persistent way. Each conversation starts fresh, meaning it does not adapt to personal preferences over time unless explicitly configured through settings (if available) or prompted each time. This means that if you prefer short and concise responses, you will likely need to specify that regularly.
That being said, Microsoft continuously updates Copilot's models and tuning, and feedback from users can help improve the overall product experience over time. However, this improvement happens at a broader level rather than being tailored to individual users automatically.
Given your initial impressions, it might be worth checking if Copilot GCC is running with any specific organizational policies or restrictions that could impact its performance. Also, experimenting with different prompting techniques might help achieve better responses. If Copilot’s adaptability is a key factor for your organization, you might want to explore AI solutions with more customizable and fine-tuned learning capabilities.
If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.
hth
Marcin