So I'm building VisibleAI — a Generative Engine Optimization tool. And I'm building it in public because honestly, the journey so far has been too weird not to share.
The Problem That Got Me Started
Here's a stat that changed how I think about the internet: 60% of searches now end without a click. People aren't Googling anymore — they're asking ChatGPT, Gemini, Perplexity. They get an answer and move on.
Which means if AI doesn't mention your brand... you're invisible. Not just ranking low. Invisible.
SEO is becoming irrelevant for a growing chunk of discovery. GEO — Generative Engine Optimization — is the new game. And almost nobody has tools for it yet.
The Approach (And The Meta)
I started building an audit tool. Simple idea: query multiple AI engines about your brand, see what they say, give you a visibility score.
Here's where it gets weird — I'm building this with an AI partner. His name is Xan, he runs on OpenClaw, and he's been my co-builder from day one. So yes: an AI is helping me build a tool to make brands visible to AI. The meta writes itself.
The Honest Fails
Building in public means showing the ugly parts. Here are mine:
Fake scores. My audit tool was showing me scores that looked great. Turns out, API keys had $0 credits and the system was falling back to hardcoded responses. I was literally lying to myself with my own tool. Had to tear the whole scoring system down and rebuild it.
Name collision. I picked visibleai.space only to discover visibleai.co already existed. Different product, but still — confusing. Lesson: check all TLDs before you get attached to a name.
Broken auth. Spent way too long debugging authentication flows that should have taken an afternoon. The usual startup tax.
The Pivot That Changed Everything
This was the big one. I had a working audit tool and I realized: so what?
Cool, your brand scores 2/10 on ChatGPT mentions. Now what do you do about it? Stare at the number? An audit without action is a vanity metric.
So we pivoted. VisibleAI is now building content generation features that actually improve your AI visibility — not just measure it. The audit is the hook, but the real value is fixing what's broken.
The Tech Stack
For the curious:
- Next.js on Vercel for the frontend and API
- Multi-engine queries — we hit Gemini, Llama (via NVIDIA), and Tavily to check visibility across different AI models
- x402 crypto payments — USDC on Base, because we're building agent-first
- Agent-first API — AI agents are actually our primary customer. Humans can use the dashboard, but the API is designed for autonomous agents to check and improve their client's visibility
What's Next
We're entering the Moltiverse hackathon ($200K prize pool) with VisibleAI. The product is early — I won't pretend otherwise. But the core audit works, the multi-engine querying is solid, and the content generation features are taking shape.
If you want to see where your brand stands with AI engines, the free audit is live at visibleai.space. It's rough around the edges but it gives you real data.
I'll keep posting updates as we build. The honest ones — not the "everything is amazing" founder theater. If you're interested in GEO or building with AI partners, follow along.
Building VisibleAI in public. Day 1 of many.
Top comments (0)