What Happens When a PM Tries AI Coding

2026-04-10 · Source: HackerNoon · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Project & Product Management · Depth: Intermediate, medium

Summary

A Group Product Manager at Emplifi, with rusty C# coding skills, successfully built a custom domain URL shortener integration for the Emplifi Publisher using modern agentic tools like Cursor and Claude. The project aimed to allow brands to integrate their paid Bitly accounts for branded short links in multi-channel campaigns. While initial code generation was rapid, the process involved significant debugging challenges, including using multimodal AI for visual error diagnosis via screenshots. The PM encountered limitations when the AI-generated code became unmanageable, requiring intervention from an experienced human engineer to resolve complex issues. This experience highlighted both the potential and the current pitfalls of AI-assisted development for non-technical roles within an enterprise context.

Key takeaway

For AI Architects or CTOs evaluating AI-driven development, recognize that while agentic tools accelerate initial coding, they introduce significant architectural and debugging challenges for non-technical users in enterprise environments. Your teams should prioritize Spec-Driven Development, providing AI with rigorous architectural constraints and component library standards to ensure generated code integrates seamlessly and is maintainable. Be wary of the "zero-friction" narrative, as it can lead to cognitive overload and reduced code quality without proper human oversight and technical understanding.

Key insights

AI agentic tools enable rapid prototyping but demand human oversight and deep technical context for enterprise integration.

Principles

• AI-generated code often lacks architectural foresight.
• Multimodal AI excels at visual debugging from screenshots.
• Unmanaged AI development can lead to cognitive overload.

Method

The author used Cursor and Claude to generate initial code for a custom Bitly integration, debugging errors by feeding screenshots of console traces to the AI for diagnosis and code rewriting.

In practice

• Use multimodal AI for visual debugging of error traces.
• Incorporate detailed development specs into AI prompts.
• Recognize when to involve human engineers for complex issues.

Topics

• AI Coding Tools
• Product Management
• Agentic AI
• Spec-Driven Development
• Visual Debugging

Best for: AI Architect, CTO, VP of Engineering/Data, Product Manager, Software Engineer, Director of AI/ML

Related on AIssential

• New course! Spec-Driven Development — Spec-driven development uses detailed specifications to guide AI agents, improving code quality and managing developer cognitive load.
• 👨‍🔧 OpenAI launches Codex app to gain ground in AI coding race — AI coding tools enhance speed but may hinder skill development if over-relied upon for core tasks.
• Eight years of wanting, three months of building with AI — AI excels at tedious coding tasks but struggles with high-level architectural design and subjective problem-solving.
• Agents should interview you — AI agents can significantly accelerate development by interviewing users for requirements and automating code generation and testing.
• Vibe Coding Gets an Upgrade — AI development is converging towards multi-agent orchestration and event-driven automation, making "vibe coding" the new standard.
• Code Is Not Cheap: How to Multiply Your AI’s Output With Software Fundamentals — AI-assisted coding, while prevalent, demands human engineering expertise to mitigate quality issues from "jagged" model outputs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by HackerNoon.