Why I Started Taking AI Seriously

My first reaction to AI was skepticism. My feed was full of “this tool just disrupted everything” posts, and the repetition made me numb. I don’t have a technical background, and I didn’t trust the “one prompt fixes all” narrative. For me, any tool has to answer one question: can it actually help me do the real work I’m doing right now?

What made me take AI seriously wasn’t a trend. It was pain. I was handling renovation consulting and project reviews while writing content simultaneously. My information was scattered everywhere: client conversations on WeChat, site notes on paper, ideas in voice memos, article outlines in a third document. Every day was busy, but I felt like huge amounts of time went into “moving and repeating” rather than into judgment and creation.

I made the typical mistake at first. I collected “universal prompts” that looked impressive but felt off in use. They produced complete sentences — just not in my voice. They generated suggestions — just not for my situation. Eventually I understood: the problem wasn’t the tools. I was treating them as an “answer machine” when they’re actually a “thinking partner.”

The shift happened when I started breaking my own workflow into stages. I mapped a complete task across phases: information gathering, structural organization, risk identification, written output, and post-mortem. Then I only brought in AI at the phase where it actually fit.

After a client consultation: I write my core assessment first, then use AI to check for gaps. When writing an article: I nail down the core observation and experience first, then let AI help sharpen the structure and readability. After a project finishes: I feed in my notes and ask for a “reusable checklist for next time.” Working this way, efficiency improved — but more importantly, my thinking became more stable.

People ask me what the biggest change from AI has been. Not “speed.” “Clarity.” Before, I had lots of experience but it was scattered. Now I’m better at converting experience into method, and method into reusable assets. This matters especially for people like me who came up through sites and real projects. We’re naturally strong at practice — but without a system, experience never compounds.

The limits are just as important to state clearly. AI won’t go to a site for you. It won’t own the consequences of a wrong call. It won’t build trust with an anxious client. It’s good at processing information; not good at carrying relationships. Good at generating options; not good at making value judgments. My current stance: use it when it’s useful, don’t mythologize it, improve efficiency where possible, but don’t outsource thinking.

When I say “take AI seriously,” I really mean “seriously upgrade yourself.” In traditional industries we’ve always relied on experience — and we still will. But how experience gets organized is changing. Whoever can structure their experience, clarify their expression, and systematize their workflow will have more agency. This isn’t a threat to practitioners — it’s an opening. The ones who understand both the domain and the tools become the rarest people in the room.