How a prototype became two tools for self-understanding.

I'm an IT business consultant. I work with teams, systems, and the space between them. One thing that's always bothered me is how poorly we understand each other in professional settings — we reduce people to job titles, DISC profiles, or a line on a resume.

So I started experimenting. What if AI could have real conversations — not quizzes — and actually surface the patterns that make someone who they are?

The first version was a Claude artifact called the Multi-Perspective Profile. Four AI-guided conversations, each inspired by a real thinker: Gabor Maté for emotional drivers, Esther Perel for identity tensions, Simon Sinek for purpose patterns, and Yuval Noah Harari for narrative assumptions. It used Big Five, DISC, CliftonStrengths, and Ikigai to map what came out.

It worked better than I expected. People who tried it said it surfaced things they hadn't articulated before. That was the signal to push further. The original is still online.

When I stress-tested the concept, three things broke.

The borrowed names created risk. Modelling conversations on real, living thinkers was great for prototyping but unsustainable. It implied endorsement that didn't exist and created legal exposure I couldn't manage as a solo builder.

Four lenses had a blind spot. The original four covered drives, identity, purpose, and narrative — but said nothing about how someone actually behaves in real situations. Emotional intelligence, stress responses, the gap between intention and impact. That was a whole territory missing.

The assessment stack had issues. CliftonStrengths is trademarked by Gallup. MBTI has weak scientific backing. I needed frameworks I could use freely and defend academically. So CliftonStrengths became VIA Character Strengths, and I added Schwartz Values — validated across 82 countries.

The rebuild gave every lens an original name: The Root, The Mirror, The Thread, The Orbit, and a new fifth lens — The Field — which covers behavioral patterns and how others experience you.

But the biggest shift was the output. I'd originally put a "Tension Meter" at the bottom of the profile as an afterthought — where your patterns pull in different directions. Turns out, that was the most distinctive part of the whole thing. No other personality tool leads with your contradictions. So the tensions moved to the top and became the centerpiece.

The concept became the Convergence Profile. And then it needed to become something people could actually use — not just a prototype I controlled. That meant building a real app.

I'm not a developer by trade. I'm an IT consultant who understands systems, architecture, and what good looks like — but I don't write production code day-to-day. Building a real web app meant using AI not just as the product, but as the builder.

The app is built with Next.js, React, and the Claude API. The system prompts that guide each conversation went through dozens of iterations — finding the right balance between structure and natural flow, between being warm and being perceptive, between asking good questions and making real observations.

That's The Overview Profile — five conversations, five lenses, one synthesized profile. Live at profile.theoverview.eu.

The Profile works. But five separate conversations is a commitment. Most people want to understand what this is before they invest that much time. I needed a way to deliver the same depth in a single sitting.

That's The Overview Flow. One continuous conversation that moves naturally through all five territories. Instead of five separate system prompts, there's a unified prompt that shifts focus as the dialogue unfolds — starting light, building trust, then gradually going deeper.

You choose a mode: deep (the full exploration), efficient (focused and faster), or surprise (let the AI decide). All three produce the same tensions-first profile at the end.

Building this taught me something about AI design: the bottleneck often isn't the model — it's the format you wrap around it. The same methodology, restructured as a single flowing conversation, changed the experience completely. Live at flow.theoverview.eu.

This project taught me more about AI than any course or article could. Not the theory — the practice. How to write system prompts that produce genuinely good conversations. How to structure an app when you're not a traditional developer. How to make decisions about frameworks, legal exposure, and user experience when there's no team to defer to.

The biggest lesson: the interesting part of AI isn't what it can do — it's what happens when you design carefully around what it can do. The technology is the enabler. The design decisions are the product.

Both tools are free. Both are experiments. And both are honest about their limitations — they surface patterns worth exploring, not verdicts. That's deliberate. I'd rather build something that's upfront about what it is than something that overclaims.