Asking an AI chatbot to fix a broken PC often feels like shouting into a void. You get generic advice, hallucinated registry keys, or loops of unhelpful suggestions. But the problem usually isn’t the AI—it’s how you ask it for help. After testing various approaches with Microsoft Copilot and ChatGPT, we found that specific prompting structures dramatically improve accuracy.
Context is king
The biggest mistake users make is asking vague questions like “Why is my PC slow?” Without details, the AI has to guess. Instead, provide concrete context. State your operating system version, recent changes, and exact error messages. For example, specify that you are running Windows 11 on a specific build number if known, or mention that the issue started after a recent update.
Use step-by-step instructions
AI models excel at following structured logic. Rather than asking for a solution immediately, ask the AI to guide you through a diagnostic process. Request that it asks clarifying questions before offering fixes. This mimics how a human IT professional works, ensuring the advice is tailored to your specific hardware and software configuration.
Verify every command
Even with better prompts, AI can still hallucinate commands. Never blindly copy-paste PowerShell scripts or registry edits suggested by Copilot or ChatGPT. Always read through the code to understand what it does. If a command looks suspicious, ask the AI to explain each line before you run it. This extra step prevents accidental system damage.
What this means for you
You don’t need to be a coding expert to use AI for troubleshooting, but you do need to be precise. By treating Copilot and ChatGPT like junior technicians who need clear instructions, you can save hours of frustration. The key is patience and specificity—give the AI the data it needs, and it will give you better answers.
Source: ZDNET
Over to you: Have you ever received dangerous advice from an AI chatbot while trying to fix your PC?