Podcast Summary

Scrum Master Toolbox Podcast: "AI Assisted Coding: Stop Building Features, Start Building Systems with AI"

Guest: Adam Bilišič (CEO, Nodion Labs, former CTO)
Host: Vasco Duarte
Date: February 17, 2026

Episode Overview

This special AI bonus episode explores the evolving landscape of AI-assisted software development. Adam Bilišič, an industry expert on building and educating around AI-powered developer tools, joins Vasco to discuss the nuances between “Vibe Coding” and AI-augmented coding, practical strategies for leveraging coding agents, best practices, and future trends. The central message: stop obsessing over building features, and start thinking about building flexible systems with and for AI agents.

Key Discussion Points & Insights

1. Vibe Coding vs. AI-Assisted/Augmented Coding

[02:26][Adam]

• Vibe Coding is a new but loosely defined term, referring to building applications without deep technical knowledge of fundamentals like database structure or security.
  • Eases product owners into offering clearer feedback by letting them “sketch” digital ideas directly.
  • Useful for quickly shaping product interactions, but limited by lack of deeper tech understanding.
  • “The person creating the app doesn't have to have in-depth overview of how the app works in the background... They’re essentially a manual tester of their own application.”
• AI-Assisted/Augmented Coding involves equipping professional developers with AI “coding agents” that handle repetitive coding tasks while the developer steers architecture, best practices, and systemic thinking.
  • Shift is from feature implementation to system design and reusable patterns.
  • “Instead of now building features, change your thinking and start building systems.” [06:40][Adam]

2. Features vs. Systems Mindset

[06:40–09:17][Adam & Vasco]

• Building features = manually coding each individual requirement; building systems = creating re-usable recipes, commands, or agents that automate recurring needs across a codebase.
  • Example: Instead of prompting the agent to make a web form each time, create an agent or command that generates all forms with required standards.
  • “You’re essentially trying to get your thinking into a reusable, like recipe.” [08:31][Adam]
  • Systems codify best practices, design choices, and team knowledge into tools the AI consistently applies.

3. Practical Tips for Success with AI Coding Agents

[12:37–17:23][Adam]

• Adam’s journey: Early experiments building a custom GPT-based coding agent revealed the crucial importance of managing context (what info the agent “remembers” and processes) and writing effective prompts.
• Context Management:
  • Beware of overloading AIs with too many MCPs (plugins/tools). Check context usage (e.g., /context in Claude Code).
  • Use sub-agents so each handles only what is needed (e.g., frontend agent with Playwright; backend without).
  • “Take a look at what kind of MCPs you have installed and how much context they are eating...” [17:23][Adam]
• Prompting:
  • Use reusable prompt templates, not ad hoc instructions every time. This standardizes behavior, improves collaboration, and reduces errors—especially in a team setting.
  • “You should create specific template of the prompt...It helps you define what these either commands, skills, sub-agent...are doing.” [19:16][Adam]
• Scalability:
  • Structure, unification, and standardization are essential as companies scale AI-assisted development.

4. The Five Levels of AI Augmented Coding

[21:04–32:39][Adam] Adam shares a developmental “ladder” for AI-augmented coding skills:

1. Master of Prompts:
   • Learn to write effective, clear prompts. Limitation: all architectural/context knowledge must be constantly retyped.
   • “If we stick to the prompt level...past information...we need to keep that all in our memory and remember to type it every time.” [25:41][Vasco]
2. Configuration Mastery:
   • Use agent configuration files (e.g., cursor.rules, claude.md) to auto-load recurring rules or styles. Caution: context pollution as files grow.
3. Context Mastery:
   • Understand/manage what is being loaded into the LLM’s context window; modularize info (separate markdowns, targeted plugins), and distribute across agents.
4. Automation Mastery:
   • Automate recurring workflows by creating custom agents, commands, skills. Build a “virtual team” of specialized agents delegated by an orchestrator.
5. Orchestrator:
   • Develop/thinking in systems: an orchestrator agent calls specialized sub-agents as needed, compartmentalizing context and automating full workflows.
   • “You are now creating systems and orchestrating. Obviously it’s still important you check what the agents are doing and monitor them…” [31:07][Adam]
   • Emphasized: always use version control (e.g., Git) for backtracking.

5. Trends and the Future of AI Coding

[21:04][Adam]

• The field is rapidly evolving.
• Developers who refuse to adopt AI tools risk becoming obsolete; companies increasingly require these skills.
  • “If they didn’t change their opinion in the last two, three years ... it’s kind of last chance to do so” [22:52][Adam]
• Adam wrote a PDF guide on these levels, available in the show notes.

6. Recommendations and Resources

[33:37][Adam]

• Official Claude Code documentation
• Claude Code Subreddit
• YouTube: Indie Dev Dan ("videos and tips...build very good prompts and inspires you what kind of agents and skills and commands you can build")
• Connect with Adam on LinkedIn for regular posts and discussions, or visit Nodion Labs for more resources.

Memorable Quotes

• On Vibe Coding:
  • “The person who is actually creating the app doesn’t have to have in-depth overview or understanding of how the app works in the background... They’re essentially like manual tester of their own application.” [02:26][Adam]
• On Shifting Focus:
  • “Instead of now building features, change your thinking and start building systems. And that’s the fundamental change...” [06:40][Adam]
• On Practical AI Use:
  • “Managing the context is everything...The more of the context is filled, the higher the likelihood is of hallucination...It’s good to understand how it works in the background and how to write the good prompts which yield you better results.” [13:40][Adam]
• On Learning AI Augmented Coding:
  • “If they didn’t change their opinion in the last two, three years, I would say that we are getting to the point where it will be kind of last chance to do so because the technology is evolving so fast and the companies usually already don’t hire people who don’t have these skills.” [22:52][Adam]
• On Systematization:
  • “You start to create your own team of virtual developers with specific focuses.” [29:55][Adam]
• On Orchestration:
  • “Orchestrator needs to give him enough information...But once you master this, then it’s amazing. They do their job, they return, and then the orchestrator can be like, okay, now let’s spin up the tester...” [31:09][Adam]
• On Reliability:
  • “Always backtrack when something goes wrong...that happens to humans as well. Right? We get into a change path that eventually becomes so deep and so complicated that you have to just backtrack and get back to where you were and try again.” [32:07][Vasco]

Timestamps for Important Segments

• [02:26] – Adam’s definition of Vibe coding
• [05:22] – Distinction: Vibe coding vs. AI-augmented coding
• [06:40] – Mindset shift: Features vs. systems
• [12:37] – Adam’s journey: Practical context management and prompting
• [17:23] – Actionable tips: Managing context and reusable prompt templates
• [21:04] – Resistance to AI and “the five levels” framework
• [25:41]–[32:39] – Deep dive: The five levels, orchestration, and automation
• [33:37] – Resources and learning recommendations
• [34:23] – How to connect with Adam and further learning

Conclusion

This episode is an essential listen for anyone navigating the fast-changing world of Agile development and AI-powered coding. Adam Bilišič offers both strategic perspective and actionable advice—urging teams to move beyond prompting and towards system-building, orchestrating a team of coding agents that embody organizational knowledge and best practices for sustainable, high-quality development.

Resources Mentioned:

• Official Claude Code Docs
• Claude Code Subreddit
• YouTube: Indie Dev Dan
• Nodion Labs
• Adam Bilišič on LinkedIn