Skip to main content
TUR Tech Pulse — follow updates from the tools I use and recommend. Occasional, curated.

AI Agents, Company OS & Agentic Systems

Agentic engineering, cost-aware model routing, agent organizations, Company OS, and MCP infrastructure — canonical notes on ROI, spend policy, and production agent systems.

Start with the cost/ROI series hub and Operating Note 002 (right agent, right model per task). Then explore the €35/mo autonomous content org case, MCP production lessons, and Company OS architecture. This hub prioritizes operating models and measurable routing over generic agent intros.

Start here — cost/ROI & agent organizations

Canonical entry points for model routing, spend policy, and production agent org patterns — not generic agent intros.

Articles

What is AI Agent Cost ROI?

analysis · · Updated
AI agent cost ROI is outcomes per euro when failure, retries, and routing are priced in — not a generic TCO calculator or CFO SaaS pitch.

OpenRouter vs LiteLLM for Agent Model Routing

analysis ·
Both sit between your agents and model vendors — but OpenRouter and LiteLLM solve different routing problems. A practitioner comparison for agent orgs, not an affiliate listicle.

What is Agentic Cost Engineering?

analysis · · Updated
Agentic cost engineering is the practice of routing each agent task to the cheapest capable model and role — then measuring ROI with failure priced in. Not generic AI spend tips.

Operating Note: When Is an Autonomous Content Org Actually Done?

operating note ·
A production-grade done definition for autonomous content orgs — not when agents write drafts, but when approval gates, role separation, failure alerts, and cost lines hold under real traffic.

Operating Note: Domain Eval Fixtures (Content Voice + SQL)

operating note ·
Extend the generic routing battery with domain packs — copy-paste fixtures for content voice compliance and warehouse SQL correctness before you promote a model into a lane.

From Unused Dashboards to Unused AI Output — The Next Shift Is Operations, Not More Artifacts

analysis ·
Dashboards failed because they never entered daily action. Generative AI is repeating the trap with decks, drafts, and more boards. The shift is embedding agents into operations with policy and ROI discipline.

Operating Note 003: Spend Policies Per Agent Role

operating note ·
Company OS-shaped spend limits: each agent role gets a different budget ceiling and approval gate — not one shared API key with no guardrails.

Operating Note: Failure-Priced ROI Ledger Week

operating note ·
A one-week ledger template: log task class, model route, spend, failure cost, and accepted outcomes — so cheap models cannot hide behind token dashboards.

Semi-Automated 24/7 Cursor + GitHub Learn Loop — What I Run Day-to-Day

field note ·
After go-live: feed agent:ready issues, approve PRs from GitHub Assigned, promote learn:candidate into rules — semi-auto 24/7, not a merge factory.

Company OS Learnings — What I Kept After Retiring the Control Plane

field note ·
Company OS was a TUR control-plane experiment. It is no longer day-to-day ops — but the six-layer model, adapter boundaries, and approval ladder still transfer.

Cursor Cloud Automations on This Site — Issue to Draft PR

field note ·
How this Astro repo ships from GitHub issues: agent:ready labels, a cron Execute automation, Node 24 Cloud Agents, draft PRs, and CI before a human merges.

MCP Gateway Lessons from Production

field note ·
What breaks when you put a Model Context Protocol gateway in front of real tools: permissions, blast radius, timeouts, and silent half-successes.

Routing Tables for Content and Data Agents

field note ·
Copy-ready routing tables for content and data agents — task lanes, model tiers, and approval gates using the same A–D discipline as coding agents.

Eval Fixture Pack for Model Routing (Copy-Paste)

tutorial ·
A small, stable eval battery you can copy — fixtures for triage, JSON, edits, critique, and irreversible-action refusal — to promote models into agent lanes safely.

Automatic Model Watch: Know What New Models Are Good At Before You Route Traffic

field note ·
A practical pattern for watching new LLMs automatically — capability smoke tests, failure modes, and when to promote a model into production routing.

Operating Note 002: Right Agent, Right Model, Best ROI

operating note ·
My niche: cost-optimize agent systems by routing each task through OpenRouter to the cheapest capable model and agent role — without killing reliability.

When Cheap Models Get Expensive: Pricing Failure Into Agent ROI

analysis ·
Token price is a trap. A free model that fails silently and burns human hours is expensive — here’s a simple way to price failure into agent ROI.

Autonomous Content Org on €35/Month: What Actually Runs

case study ·
How one person runs a multi-site content organisation with AI agents on roughly €35/month of infrastructure — roles, approvals, costs, and failure modes.

Company OS vs Agent Framework vs Orchestration

analysis ·
Agent frameworks, workflow orchestration, and Company OS solve different problems. A practical comparison so you stop buying the wrong layer.

Agent Cost & Model Selection Series — Start Here

article · · Updated
Hub for the cost/ROI series: right agent, right model, model watch, routing tables, eval fixtures, and pricing failure — positioning through practice.

When to Hire a Fractional Agentic Engineer

analysis ·
Fractional agentic engineering makes sense when you need production systems fast, not another demo. Signals, engagement shapes, and when to wait.

Company OS: The Architecture, Vision, and Plan

operating note · · Updated
Company OS is my working thesis for connecting human intent, shared context, autonomous execution, and accountable outcomes. Updated August 2026 with the live system map of what is actually wired in.

Your AI Agent Stack Is Your Competitive Edge — If You Treat It as Infrastructure

analysis ·
Most developers use AI in chat. I built a 24/7 agent stack on my own server. Here's what actually created leverage — and what was just noise.

Agentic Engineer vs AI Engineer: What Companies Actually Need in 2026

analysis ·
AI engineer and agentic engineer are not the same role. This guide explains the difference in scope, outcomes, and when each role creates the most business value.

Build In-House vs Hire an Agentic Engineering Consultant

analysis ·
A practical framework for deciding whether to build agentic capability internally or work with a consultant first. Includes speed, risk, and cost trade-offs.

What is Agentic Engineering?

analysis · · Updated
Agentic engineering is the discipline of designing and building systems where multiple AI agents coordinate autonomously to accomplish goals — without a human triggering each step. Here's what that actually means when you build it in production.

How to Build an AI Agent for Your Website: Complete Guide with n8n, Knowledge Base, and Frontend (2025)

tutorial · · Updated
Step-by-step guide to building a production-ready AI agent for your website using n8n, vector database knowledge base, and custom frontend. Includes code examples, architecture patterns, and deployment strategies.

From SaaS Dependence to Speed: How I Built an AI Feedback Agent in a Weekend

case study ·
I built a fully-automated AI feedback agent using n8n, LangChain and Azure OpenAI in a weekend. This experience revealed a bigger shift in how enterprises should think about tooling – build first, buy only when it makes sense.

Projects

Agentic Data Organisation with Claude

production
Design and implementation of an agentic data organisation with Claude coordinating analytics, data platform, pipelines, governance, and data quality.

How I Built an AI Feedback Agent That Replaces Traditional Surveys – Using n8n and GPT-4

prototype
Transform feedback collection from tedious surveys into intelligent conversations. Learn how I built an AI-powered feedback agent with n8n that asks follow-up questions, detects language automatically, and extracts structured insights – all while saving thousands on survey platforms.

MCP-Connected Development Workflow for Data-Informed Prototyping

prototype
Connected Claude to analytics, research, strategy, and design context through MCP to build React prototypes inside the development environment.

Work with me

Scoping agentic systems, MCP infrastructure, or a Company OS-shaped operating layer? Send a message to start.

ContactAgentic engineer (Stockholm)

Related Topics

Modern Data Stack Engineering

Build and optimize modern data architectures with cloud-native tools.

Workflow Automation & Orchestration

n8n orchestration, API automation, and where workflows fit under agents.

Full-Stack Development (Astro/TypeScript)

Modern web development with Astro, TypeScript, and performance optimization.

Conversational Analytics & Next-Gen BI

Conversational analytics, natural language interfaces, and next-gen BI.

All Topics

Modern Data Stack Engineering

Build and optimize modern data architectures with cloud-native tools.

Workflow Automation & Orchestration

n8n orchestration, API automation, and where workflows fit under agents.

Hockey Analytics & Sports Data Science

AI-powered hockey analytics and advanced sports data science.

Full-Stack Development (Astro/TypeScript)

Modern web development with Astro, TypeScript, and performance optimization.

Developer Productivity & Tooling

Developer workflows, tooling, and productivity patterns for technical professionals.

Stockholm/Nordic Data Engineering Market

Stockholm tech scene, Nordic data engineering market, and local insights.

Conversational Analytics & Next-Gen BI

Conversational analytics, natural language interfaces, and next-gen BI.

AI Agents, Company OS & Agentic Systems

Cost/ROI routing, agent organizations, Company OS, MCP.