Notes from building systems
What works, what breaks, and what we learned building automation systems, Agentforce agents and products. Working notes, not content written for traffic.
- 31 Oct 2026
n8n + AI Agents | 30 Days of AI Agents, Day 12
How to combine n8n workflow automation with AI agents: architecture, the AI Agent node, a research assistant example, use cases, and best practices.
- 30 Oct 2026
AutoGen | 30 Days of AI Agents, Day 11
How Microsoft AutoGen works: conversational multi-agent teams, agent roles, group chat, code execution, and how it compares to LangChain and CrewAI.
- 29 Oct 2026
CrewAI | 30 Days of AI Agents, Day 10
Day 10 of 30 Days of AI Agents: how CrewAI organizes role-based agent teams, tasks, and processes, with a working example and honest trade-offs.
- 28 Oct 2026
LangChain | 30 Days of AI Agents, Day 9
Day 9 of 30 Days of AI Agents: what LangChain is, its core building blocks, a small chain and tool-calling example, and when plain code is clearer.
- 27 Oct 2026
MCP (Model Context Protocol) | 30 Days of AI Agents, Day 8
Day 8 of 30 Days of AI Agents: what MCP is, how hosts, clients, and servers fit together, a small Python MCP server, and the security caveats.
- 26 Oct 2026
Building Your First AI Agent | 30 Days of AI Agents, Day 7
Day 7 of 30 Days of AI Agents: build a small working agent in Python with tool calling, web search, a safe calculator, memory, and a step limit.
- 25 Oct 2026
AI Agent Lifecycle | 30 Days of AI Agents, Day 6
Day 6 of 30 Days of AI Agents: the six-stage agent lifecycle from perceive to evaluate and learn, with a worked example and what to log at each stage.
- 24 Oct 2026
How AI Agents Use Memory | 30 Days of AI Agents, Day 5
Day 5 of 30 Days of AI Agents: short-term, working, long-term, and semantic memory, where each lives, how agents use it, and the mistakes to avoid.
- 23 Oct 2026
Single-Agent vs Multi-Agent Systems | 30 Days of AI Agents, Day 4
Day 4 of 30 Days of AI Agents: single-agent vs multi-agent systems, when each fits, the real costs of coordination, and how to grow from one to many.
- 22 Oct 2026
Prompt vs Workflow vs AI Agent | 30 Days of AI Agents, Day 3
Day 3 of 30 Days of AI Agents: prompts, workflows, and agents compared, with clear rules for picking the cheapest level that solves the problem.
- 21 Oct 2026
AI Agent Architecture | 30 Days of AI Agents, Day 2
Day 2 of 30 Days of AI Agents: the six building blocks of agent architecture, the think-plan-act-evaluate loop, and a practical example stack.
- 20 Oct 2026
What Is an AI Agent? | 30 Days of AI Agents, Day 1
Day 1 of 30 Days of AI Agents: what an AI agent is, its six core components, the think-act-observe loop, and how it differs from a plain chatbot.
- 19 Oct 2026
What an AI agent can automate in a business (and what it shouldn't)
One agent on the inbox reads the request, pulls the CRM record, drafts a reply in your tone and stops. Triage, follow-ups and status reports follow the same pattern.
- 19 Oct 2026
The Rule Behind Every System We Build: If It Repeats, Automate It. If It Repeats as a System, Productize It.
Avnish Yadav on the rule behind CodeITronics: spot repetition, climb from manual to script, workflow, system, and product, and avoid automating too early.
- 18 Oct 2026
E-commerce Automation That Actually Reduces Support Tickets, Not Just Orders
Design e-commerce automation around fewer where-is-my-order tickets: Shopify webhooks, proactive delay alerts, inventory sync, and idempotent retries.
- 17 Oct 2026
n8n vs. Custom Code: When Visual Automation Beats a Backend Build
A balanced guide to choosing n8n or custom code for automation, based on state, change frequency, failure cost, testing, scale, and multi-tenant needs.
- 16 Oct 2026
Migrating Millions of Rows With Zero Downtime: Lessons From a Mongo-to-Postgres Cutover
How we moved millions of rows from MongoDB to normalized Postgres with zero downtime: schema mapping, streaming backfill, dual-writes, and safe cutover.
- 15 Oct 2026
From Internal Tool to Micro-SaaS: A Founder's Playbook
How to spot a repeated pattern, abstract it into modules, and decide whether an internal tool should become a micro-SaaS, with the trade-offs we learned.
- 14 Oct 2026
Stop Logging Into 5 Dashboards: Building a One-Slack-Message KPI Bot
How we replaced five daily dashboard logins with one Slack message: a cron worker that polls Stripe, PostHog, and SQL, computes KPIs, and flags anomalies.
- 13 Oct 2026
Auto-Resolving Routine Support Tickets: Inside an AI Triage System
Inside our AI support triage system: knowledge-base grounding, intent detection, and escalation logic that auto-resolves a large share of routine tickets and routes the rest.
- 12 Oct 2026
Your Inbox Doesn't Need Another App — It Needs an Agent
How we built an AI inbox agent that classifies, summarizes, and routes email into tasks, with confidence thresholds and human review to keep it trustworthy.
- 11 Oct 2026
How an Autonomous Lead-Gen Agent Multiplied Outbound Volume
How our lead-gen pipeline with Puppeteer, AI agents, HubSpot, and Postgres removed manual SDR research and delivered a several-fold increase in outbound volume.
- 10 Oct 2026
Why Most AI Agent Projects Fail Before They Launch
AI agent projects rarely fail on model quality. They stall on vague scope, missing evals, bad data, and no observability. Seven patterns and how to fix them.
- 9 Oct 2026
The Hidden Cost of Skipping a POC Before Agentforce Rollout
Skipping an Agentforce POC moves cost into production. The failure modes it causes, the rework that follows, and what a 2-4 week POC should prove.
Subscribe to the edition
One useful systems note a week from real Salesforce and automation builds. Unsubscribe in one click.