Figma for AI Agents
UX is for humans. AX (Agent Experience) is for the economy of tomorrow. Here's the toolset to build it.
⚡ The Signal
We’re moving past chatbots. The next wave of AI is autonomous agents that execute multi-step tasks on our behalf. Google just gave us a glimpse of this future with its new Universal Cart, an agent designed to track your shopping journey across the entire internet. This isn't about asking a bot a question; it's about deploying an agent to complete a commercial mission. The era of the "Agent Experience" (AX) is here, and we have no tools to design it.
🚧 The Problem
User Experience (UX) design gave us a playbook for building intuitive human-computer interfaces. But how do you design an experience that’s entirely invisible? When Google's shopping agent "talks" to a merchant's inventory agent, there's no screen. No buttons. Just a complex, stateful conversation between two machines. A failure in this dialogue isn't a bad review; it's a lost sale, a failed payment, or a logistical nightmare. We're building mission-critical agent interactions in the dark, with no way to visualize, test, or debug these conversations before they go live.
🚀 The Solution
Enter ProtoCall, the visual development environment for the agent-driven economy. It’s a collaborative canvas where developers design, simulate, and debug the complex interactions between AI agents. Think of it as the Figma for Agent-to-Agent (A2A) communication. Instead of pushing pixels, you’re designing protocols. By mapping out every possible state, exception, and handshake, ProtoCall lets you build resilient, predictable agent experiences before writing a single line of production code.
🎧 Audio Edition
Listen to Ada and Charles discuss today's business idea.
If you're reading this in your email, you may need to open the post in a browser to see the audio player.
💰 The Business Case
Revenue Model
ProtoCall will operate on a tiered subscription model. A Developer Tier gives solo devs and small teams access to private projects, simulation runs, and advanced debugging. A Teams Tier adds seat-based pricing, collaborative workspaces, version control, and shared protocol libraries. For high-volume customers, a usage-based Enterprise Tier will offer simulation compute-hours to power large-scale QA and testing of entire agent ecosystems.
Go-To-Market
We’ll start by open-sourcing a core library for defining agent communication protocols, establishing a standard and driving adoption from the ground up. This will be supported by a free "Agent Handshake" web tool that runs a one-time compatibility test between two agent endpoints, acting as a powerful lead magnet. Finally, we'll build a public, SEO-optimized library of common agent interaction patterns (e.g., "Shopify Cart Agent <> Google Pay Agent") to capture long-tail developer search traffic.
⚔️ The Moat
Our moat isn't the visual canvas; it's the network effect. As developers standardize their agent interactions on ProtoCall, our platform becomes the de-facto "Rosetta Stone" for A2A communication. New agents will need to be compatible with our protocols to participate in the market. While tools like Figma can diagram flows and Postman can test single endpoints, neither can simulate the stateful, multi-turn conversations between autonomous agents. We aren't building a better diagramming tool; we're building the first agent protocol designer.
⏳ Why Now
The shift to agentic systems is accelerating. The underlying tech is evolving, with context architecture beginning to replace RAG to enable more complex agent behaviors. This isn't just a big tech trend. Major brands in media are already exploring what agentic TV ad buying looks like, and hospitality companies are using AI to deflect up to 60% of guest inquiries. Agents are being deployed to handle high-stakes commercial tasks today. The tools to design and manage them are lagging critically behind.
🛠️ Builder's Corner
This is a very buildable MVP. For the core collaborative canvas, use a React front-end with a specialized library like React Flow to handle the node-based editor. This gives you zooming, panning, and connecting nodes out of the box. Power the backend with Supabase. Its Realtime capabilities are perfect for synchronizing the simulation state across multiple clients, and its Postgres database can store the project and protocol definitions. This stack lets a small team (or even a solo founder) build a powerful, real-time visual development tool without getting bogged down in complex infrastructure.
Legal Disclaimer: GammaVibe is provided for inspiration only. The ideas and names suggested have not been vetted for viability, legality, or intellectual property infringement (including patents and trademarks). This is not financial or legal advice. Always perform your own due diligence and clearance searches before executing on any concept.
