Most AI development tools operate as black boxes. You send a message, something happens behind the scenes, and you get a response. Scott Vance built the opposite approach. Convo Lang is a plain text framework that exposes every part of the AI interaction process, making context engineering both precise and visible.

In this episode of Lead with AI, host Dr. Tamara Nall speaks with Scott, a creator and developer from just north of Cincinnati who has spent nearly 20 years building software. At his core, Scott loves creating things, whether on computers or through woodworking. This maker mentality drove him to constantly find more efficient ways to work, especially as AI began transforming development. Scott noticed massive repetition in AI work. Basic tasks like formatting prompts and converting them into structures that models could process created constant friction. This sparked Convo Lang, which started as a small library and evolved into a comprehensive framework for developing AI-native applications.

 Context Engineering Made Visible 

When most people use ChatGPT, they see message bubbles appearing on screen. Behind that interface, the system takes your message and adds it to a list of all previous messages plus responses. This list becomes the context. The context window determines how many messages the AI can process simultaneously.

Here's what most people don't realize: every time you send a message to an LLM, every single previous message gets reset and reprocessed. Eventually, you run out of room. Convo Lang helps you get the most out of limited context by being exact in how you feed information to models while exposing what's happening in the background.

Convo Lang files are just text files with simple formatting. Each line indicates whether a user, AI, or system is talking. That full list of messages gets stored in a plain text file you can read. You can combine it with traditional procedural code through an embedded scripting language. Complex features like tool calling and RAG all get written in text files that humans can easily read. When you have that file, you feed it into the Convo Lang CLI, which reads, executes, sends formatted data to your chosen AI, and returns the response.

 Week's Work in Hours   

Scott recently experienced a moment that validated his entire approach. He's been developing Convo Make, a tool that builds on Convo Lang and provides spec-driven development. Instead of writing code line by line, you write down what the application should do and feed it to an LLM.

Scott admits he's been spending more time on Convo Lang than his main job. He needed to redo the interface for MindArk.ai, his application project. Using Convo Make, he set up the spec and generated the entire frontend and backend in one shot. This included security rules and all the components that many no-code tools miss. Besides small tweaks afterward, the application functioned completely. What would have taken a week happened in a few hours.

This moment crystallized Convo Make's potential through what Scott calls precision context engineering, where every specification translates directly into working code because context is managed explicitly rather than hidden.

 Ethics and the Human Interface for AI   

When asked about ethics, Scott acknowledges AI will empower many people, but unfortunately, many others won't be needed anymore. However, he believes Convo Lang provides a more reliable human interface for working closely with AI. The framework helps developers understand what's happening under the hood rather than using their behavior to train further models, which many tools do behind the scenes.

Scott works closely with a developer at Syntax Data who uses Convo Lang for early prototyping. The breakthrough came when this developer realized every interaction with an LLM gets stored in a local file you can review, creating complete history of how AI reached specific decisions.

 The Future Where AI Interprets 

Scott's boldest prediction centers on language models that are also code interpreters. Currently, we tell AI to generate the same code humans have been writing for decades. Scott envisions AI models that can internally evaluate code written specifically for AI rather than for human developers.

This represents a fundamental shift. Right now, AI writes Python and JavaScript designed for human readability. But what if AI could interpret code optimized for machine processing? Scott believes this would create a paradigm shift, potentially accelerating model progress dramatically.

In the short term, Scott is working on features that integrate with MCPs (Model Context Protocols) for deployment-ready applications. He's also developing a SaaS service where you can describe your application using Convo Make and build production-ready apps in one click and a couple of prompts. Long-term, Scott hopes Convo Lang becomes something that lives in the background that users never know about, silently managing context.

 Getting Started with AI Development   

For developers curious about Convo Lang, Scott offers multiple entry points. The easiest way is visiting learn.convo-lang.ai, which provides interactive examples that function as both documentation and tutorial. For more technical developers, the Convo Lang VS Code extension allows writing Convo Lang directly in the editor with support for all major models. Scott maintains a sub-reddit for Convo Lang where the community discusses use cases and shares examples. He occasionally posts on LinkedIn but considers Reddit the primary platform.

Scott's approach reflects his identity as a creator who learns by trying things rather than reading books (he's the first Lead with AI guest to admit he doesn't have a book recommendation). This hands-on philosophy permeates Convo Lang's design: expose everything, make it readable, let developers control every part of the process.

Ready to try plain text AI development? Visit Learn.Convo-Lang.ai for interactive examples or install the VS Code extension to start building AI-native applications.

For more insights on how AI is transforming business and society, Subscribe to the Lead with AI podcast, where we explore the frontiers of artificial intelligence with the innovators who are shaping its development.

Follow or Subscribe to Lead with AI Podcast on your favorite platforms:

Website: LeadwithAIPodcast.com | Apple Podcasts: Lead-with-AI | Spotify: Lead with AI | Podbean: Lead-with-AI-Podcast | Email: Tamara@LeadwithAIPodcast.com

Follow Dr. Tamara Nall:

LinkedIn: @TamaraNall | Website: TamaraNall.com | Email: Tamara@LeadwithAIPodcast.com

Follow Scott Vance:

LinkedIn: @scott-vance | Convo Lang AI: Learn.Convo-Lang.ai | Email: scott@convo-lang.ai