The Agento Suite: Why I'm Building AI Agents for Every Industry

I've been building products for nearly 18 years now. Aviation tools, travel platforms, AI experiments, the full spectrum. But sometime around late 2023, a pattern crystallized that I could no longer ignore: the most impactful software I could build in the next decade would not be traditional SaaS. It would be AI agents.

Not chatbots. Not wrappers around GPT. Agents: systems that understand a domain deeply enough to take meaningful action on behalf of a user.

That realization is what led me to start thinking about what I now call the Agento suite.

The Problem with Horizontal AI

Every week, a new "AI for everything" product launches. They promise to handle your emails, write your code, manage your calendar, draft your legal documents, and probably make your coffee. The pitch is seductive: one AI to rule them all.

But here's what I've learned from building products across aviation, travel, and now legal tech: domain expertise isn't a feature you bolt on. It's the foundation.

When I built Aviation Infinity, I didn't build a generic content platform and then add aviation topics. I built something that understood the specific needs of aviation enthusiasts, the specific content formats they consume, the specific community dynamics that matter. That domain-first approach is what made it work.

AI agents need the same treatment. A general-purpose AI assistant can draft a legal letter, sure. But it doesn't understand the difference between a demand letter and a cease-and-desist in the context of small claims court in Massachusetts versus California. It doesn't know which courts accept e-filing, what the local rules are for service of process, or how to structure arguments that a specific judge tends to find persuasive.

That level of domain depth is where the real value lives.

The Agento Thesis

The core thesis behind the Agento suite is simple: build vertical AI agents that are deeply specialized for specific industries, then connect them through shared infrastructure.

Each agent in the suite (LegalAgento, CodeAgento, and others I'm still designing) shares a common technical foundation. They all use the same streaming architecture, the same tool-use patterns, the same structured output formats. But their domain knowledge, their UX, their data models, and their workflows are completely different.

Think of it like this: a lawyer needs an agent that understands legal research, case law citation, document drafting conventions, and court procedures. A developer needs an agent that understands codebases, testing patterns, deployment pipelines, and code review workflows. The underlying AI infrastructure might be similar, but the surface area of expertise couldn't be more different.

Why "Suite" and Not "Platform"

I deliberately chose to build a suite of products rather than a platform. The distinction matters.

A platform implies that you're building infrastructure for others to build on. That's a valid strategy, but it requires massive scale to work. You need thousands of developers building agents on your platform to justify the investment.

A suite means I'm building the agents myself, for specific users, with specific problems. Each product in the Agento suite is a standalone business that can succeed or fail on its own merits. But they share enough DNA that building one makes the next one faster.

After building multiple products, I've learned that shared infrastructure between products is one of the biggest competitive advantages a solo founder can have. When I build a streaming chat interface for LegalAgento, that same component, with different styling and domain-specific affordances, can power CodeAgento. When I build a document generation pipeline for legal documents, the same architecture can generate code documentation.

The Technical Foundation

Every Agento product is built on the same stack: Next.js with App Router, TypeScript, Tailwind CSS, and the Claude API. I chose Claude over GPT for reasons I'll write about separately, but the short version is that Claude is better at following complex instructions and maintaining consistency across long interactions, both critical for agent-quality AI.

The shared technical patterns include:

Streaming responses. Every agent streams its output in real-time. This isn't just a UX preference. It's essential for tasks that take time. When LegalAgento is researching case law, users need to see progress. When CodeAgento is analyzing a codebase, developers need to see the reasoning unfold.

Tool use. Agents are not just language models that generate text. They use tools: they search databases, query APIs, read documents, write files. The tool-use infrastructure is shared, but the specific tools are domain-specific.

Structured output. When an agent produces a deliverable (a legal document, a code review, a research summary), it needs to be structured, not just a wall of text. Shared output schemas with domain-specific templates make this work at scale.

What I Am Building First

LegalAgento is the first product in the suite to move from concept to development. The access-to-justice crisis is real and urgent. Millions of people can't afford lawyers for problems that have legal solutions. I believe AI can bridge that gap, not by replacing lawyers, but by making basic legal services accessible to people who currently have no access at all.

CodeAgento is next. As someone who builds and maintains multiple products, I feel the pain of context-switching between codebases daily. An AI coding assistant that truly understands your entire product portfolio, not just the file you have open, would be a huge deal.

The Long Game

I'm not building the Agento suite to flip it in two years. This is a decade-long project. AI agents are going to fundamentally reshape how knowledge work gets done, and the companies that win will be the ones with the deepest domain expertise, not the ones with the most general models.

Every product I've built over the past 18 years has taught me something about building for specific users in specific industries. The Agento suite is where all of that experience converges.

The future of AI isn't one agent that does everything poorly. It's many agents that each do one thing exceptionally well. That's what I'm building.