Gene Conroy-Jones
I write about the AI-Driven Development Life Cycle, where AI agents execute tasks across planning, coding, and testing while humans act as orchestrators and governors.
Building with AI
View all Building with AI posts →
What Agentic Orchestration Changes About Your Tech Stack Agentic orchestration requires a tech stack where everything—work items, decisions, config—lives versioned in the repo; agents reset each session with no institutional memory. Five litmus tests (env spin-up, config independence, ClickOps avoidance, codified business logic, team measurement) expose foundation quality. Tools like Beads and Dolt enable git-backed issue tracking and state versioning, while ADRs capture deliberate omissions, and structured BDD specs replace undocumented tribal knowledge. 
Building Production Grade Applications using Agent Orchestration. After eight weeks I built a production grade application. Five weeks spent building and learning the factory, three building the product. Pre-AI this product would have taken at least six months and a team of 3-5 to build. This is the first in a five part series that examines the lessons learned across the four signals. 
AI Agents: Cut Through the Jargon The post explains how AI agents have evolved from simple chatbots to systems that can plan, reason, use tools, and achieve multi‑step goals autonomously. It highlights the rapid market growth, noting projections of a $10 billion industry expanding at over 45 % annually, and cites Gartner’s forecast that 40 % of enterprise applications will embed task‑specific agents by the end of 2026. The article outlines the core components of an AI agent—tools, skills, memory, and reasoning engine—and introduces emerging standards such as the Model Context Protocol (MCP) and Agent‑to‑Agent (A2A) protocol that enable tool access and inter‑agent collaboration. It also describes common architectural patterns (sequential, parallel, hierarchical, human‑in‑the‑loop, router) and reviews both low‑code platforms and developer frameworks for building and orchestrating agents. Finally, it presents practical use cases where agents deliver measurable value, including intelligent document processing, autonomous customer support, research and content pipelines, and internal operations automation. 
AI Agent Development Workflow from Brief to Pull Request An epic that once took three weeks now finishes in an afternoon, arriving as forty agentic work items instead of six tickets, breaking every ceremony built around the old cadence. The agentic version inverts the classic feature factory problem: keep-the-lights-on work becomes a precondition rather than a deferrable cost, and the same agents that need the foundation are effective at building it. The process runs through a directed acyclic graph formula—from a one-page brief to testable requirements, a plan reviewed by a separate agent, parallel implementation sessions, and a final report with SHA-256 hashes and a coverage matrix—with the human required only at the plan review gate. 
Fix One Bottleneck, Find the Next: AI and the Kaizen Loop You rolled out AI coding tools across the engineering organisation. Adoption hit 90%. The metrics look great. But the same rollout shifted your bottleneck. Like a production line, tuning one stage just exposes the next constraint, and this one could be landing on your senior engineering leaders, the 10x multipliers you can't afford to lose. Check in with them. They may need protecting. The traits of a strong senior engineer (knowledge sharing, iterative process, sound architecture, objective measurement) are ones I've written about before in [Five tenets of an early-stage startup engineer](https://www.foursignals.dev/blog/2022-09-12-five-tenets-of-an-earlystage-startup-engineer/). Your most experienced engineers, the ones who design the systems, mentor the team, and make the decisions that compound over years, may be the least likely to use AI. But when they do, they save more time than anyone else. We know what good looks like. So why are they under-utilising AI? 
Event: The Future of Software with Steve Yegge and Ajit Banerjee The event opened with a provocative question about what AI will never replace, setting the tone for a fireside conversation at Seattle’s AI House. Speakers Steve Yegge and Ajit Banerjee explored how AI is reshaping software development, arguing that while code generation becomes cheap, the real bottlenecks shift to idea generation, context, and human judgment. They highlighted the emerging need for subject‑matter experts to steer AI‑produced output, the erosion of traditional moats in favor of proprietary data and team context, and the importance of quality over sheer velocity. Practical insights included the rise of agent‑driven tooling, the risk of feature‑factory mentalities, and the value of human‑to‑human connections in ideation. The discussion concluded with observations on future product development, the evolving role of AI in SDLC 3.0, and a personal answer that human contact and empathy remain irreplaceable.