The Team That Wasn't

I was brought in as a senior architect to build a brand new AI product. On paper, the team existed — developers, a tester, a client, system integrators. Everything was in place.

The requirements I was handed were summaries — another customer's docs fed into ChatGPT, repackaged, and thrown on the wall. When I asked for a call to understand the rationale, there was resistance. So I figured it out myself — talked to the customer, spoke with the system integrators, and pieced together a rough design.

The offshore developers hadn't been onboarded to any requirements. I gave them recordings, architecture docs, sequence diagrams, a step-by-step plan. Some didn't show up after the first week. Others couldn't code to standard. The most senior engineer spent days with an AI assistant and produced code that never worked. I let people go and started over.

I rebuilt everything from scratch — in stages, testing every piece myself, using Claude Code, working weekends. 300 hours in one month. The product shipped. But I wasn't proud of how we got there.

The Trap

Looking back, I should have paused and owned the hiring problem myself. But when the deadline is staring at you and the client is waiting, your instincts take over. My anxiety said deliver this now, not fix the team first. I was already the architect, the developer, the BA, the delivery lead, and the tester. I couldn't add recruiter to that list.

That's the trap nobody warns you about. When the team is broken, the strongest person absorbs all the work. And the more they absorb, the less time they have to fix the real problem — which was never the product. It was the people.

The Broken Funnel

Interviews should be fun. When the funnel works, you sit across from someone sharp, talk through a real problem, and walk away energized. But when contracting firms are sending whoever is available — focused on their percentages, not on quality — every interview becomes draining. Candidates who couldn't think independently, never used an AI tool, didn't ask a single follow-up question. Eventually you stop looking and just do the work yourself.

What I Learned

The product shipped. But it didn't have to be this hard. One good team and it would've been a completely different story.

The gap wasn't technical — AI tools were right there for everyone. The gap was human. Nobody took ownership. Nobody translated a requirement into action without being walked through every step. Nobody asked a second question.

In the AI era, the skill gap isn't about whether someone can code. It's about whether someone can think, react, own their work, and engage. AI can help anyone write code. It can't make someone care about the outcome.

We did finally hire one person who was different. It wasn't his skillset that stood out — it was his curiosity. He asked questions. He dug into the problem before jumping to the solution. That curiosity turned into ownership, and that ownership turned into trust. One person. That's all it took to see what the team should have looked like from the start.

Ownership isn't something you measure on a resume. You measure it by how someone reacts when they don't understand something — do they sit quietly, or do they lean in?

If you're building a team right now, don't ask "can this person do the job?" Ask: "will this person own it like it's theirs?" That's the only question that matters.