blog / AI
AI18 January 20233 min read

Six weeks with ChatGPT: what it's actually useful for in IT ops

The hype has died down slightly and I've been using ChatGPT daily for six weeks. Here's an honest inventory of where it earns its keep.

by Matt Roberts

Six weeks ago, ChatGPT landed and the internet briefly lost its mind. I wrote about the initial reaction. Now I've actually lived with it for a sustained period, and I want to give a more honest, less first-impressions account of where it's useful and where it's genuinely not.

This is aimed at people like me: IT professionals, not developers per se, who are trying to figure out whether this changes our work in any meaningful way.

Where it's genuinely useful

Scripting and automation

This is the biggest win for me. I write a reasonable amount of PowerShell: Intune management, Azure AD reporting, Teams administration, general IT ops automation. ChatGPT is now part of that workflow.

Typical usage: I describe what I want in plain English, it gives me a working starting point, I read it, adapt it, test it. It's not generating production-ready code; it's giving me a scaffold that's faster to complete than to write from scratch.

Last week: needed to pull a report of all Intune-managed devices with their last sync date and compliance status, filtered by a specific group, exported to CSV. Described it in two sentences. Got a working Graph API call with the right fields and proper pagination handling. Would have taken me 30-40 minutes to build myself. Took 10 minutes to verify and adjust what ChatGPT produced.

First-draft everything

Emails. Documentation. Meeting agendas. Proposals. Anything where the blank page is the enemy, ChatGPT breaks the block. I still write everything myself. I read what it produces critically and rewrite substantially, but having something to push against is faster than starting cold.

Explaining Microsoft documentation

Microsoft docs have improved, but they can still be dense, inconsistent, and written for an audience that already understands what they're asking. "Explain this to me like I'm going to have to present it to a customer who doesn't know Azure" is a prompt I use weekly.

Troubleshooting errors

Paste error message. Describe context. Get possible causes and steps to investigate. Hit rate is maybe 60-70%, but that's often enough to point me at the right place faster than a Google search.

Where it's not useful

Anything requiring current information

ChatGPT's training data has a cutoff. It doesn't know about things that happened recently. I've caught it giving me outdated Microsoft documentation references and, more dangerously, telling me features exist in products that have since changed. Always verify.

Compliance and legal questions

Never, ever rely on ChatGPT for anything with legal, compliance, or regulatory implications. It will give you a confident, plausible-sounding answer that may be completely wrong for your jurisdiction, your sector, or your specific situation.

Precise, factual claims

If you need to know the exact licensing requirements for a specific Microsoft feature, the exact syntax for an obscure cmdlet, or any specific factual claim: verify independently. ChatGPT hallucinates with confidence.

The workflow shift

The thing that's genuinely changed is the cognitive overhead of starting tasks. There's a certain resistance that comes with opening a blank document or a PowerShell window with a complex task ahead. ChatGPT reduces that resistance. Whether that's a meaningful productivity gain or just a psychological effect is something I'm still working out.

What I can say: I'm starting more tasks faster, and finishing them in less time. The quality of my output hasn't changed; if anything it's improved because I'm spending the time I save on review and refinement rather than initial construction.

That's a tool earning its keep.

#chatgpt#ai-productivity#powershell#it-ops
Share:X / TwitterLinkedIn

Related posts

Building my first AI-powered app: what I learned as a non-ML developer
AI

Building my first AI-powered app: what I learned as a non-ML developer

I built this website using Next.js, AWS, and Claude. I'm not a developer by trade. Here's an honest account of what that process was like and what surprised me.

14 May 20253 min read
Claude 3.7 and the rise of agentic AI — this is the inflection point
AI

Claude 3.7 and the rise of agentic AI — this is the inflection point

Anthropic released Claude 3.7 Sonnet in February 2025 with extended thinking mode. Combined with the MCP protocol, something important just shifted.

10 Mar 20253 min read
DeepSeek just changed the economics of AI. What it means for enterprise
AI

DeepSeek just changed the economics of AI. What it means for enterprise

DeepSeek R1 arrived in January 2025 and sent the AI industry into a brief panic. The dust has settled. Here's what actually happened and what it means for enterprise AI strategy.

28 Jan 20253 min read