Notes on building AI that ships.
Practical writing on building and shipping AI systems: the data underneath, the agents and RAG on top, the automation that runs it, the stores and sites it powers, and the search visibility that gets it found. What we learn building for clients, written down.
The range is deliberate, because we build the whole stack, not one slice of it. Some pieces cover AI search visibility (GEO), and they are structured the way we structure client content, with the answer first so assistants can quote them. The rest is simply what it takes to get data pipelines, agents, automation, and storefronts to production.
GEO vs SEO: what changes when the searcher is a model
SEO ranks a page for a person to click. GEO gets a passage quoted by a model. What actually differs, and what to do about it.
How to get cited by ChatGPT, Perplexity, and Google AI Overviews
A practical checklist for becoming the source an assistant quotes, platform by platform.
What llms.txt is, and whether it actually helps AI find you
A plain-text map for AI crawlers. What it does, what it does not, and an honest read on whether to bother.
RAG vs fine-tuning: which one your use case actually needs
Two ways to make a model know your world. When to retrieve, when to fine-tune, and when you need both.
Why your RAG chatbot gives wrong answers, and how to fix retrieval
When a RAG assistant makes things up, the model is usually not the problem. Retrieval is. The three levers that fix it.
What a production RAG system actually costs to run
The per-question price is the small part. Where the real cost of a RAG system lives, and how to control it.
AI agents vs chatbots vs automation: which one you actually need
Three terms used interchangeably, three different things. Clear definitions and how to pick.
AI agent or a simple workflow? When to use each
Agents are powerful and overused. A decision guide for when a plain n8n, Make, or Zapier workflow is the better build.
Automating an ecommerce store: from catalog uploads to weekly reports
The store tasks worth automating with AI, and the numbers from stores we built.
WhatsApp automation for business: what's possible, and what it costs
The two WhatsApps, what you can automate, and the costs and rules to plan for before you build.
Being findable now means being quotable by a model
Search traffic increasingly arrives pre-answered. What it takes to be the source an assistant cites, not the tenth blue link.
The gap between the demo and production is the whole job
Error handling, monitoring, and the 10,000th request: why the impressive prototype is the easy 20%.
Most of the AI work is the pipeline no one wanted to build
Models are only as good as what feeds them. A field guide to the unglamorous data work underneath.
Building something with AI?
From data pipelines to agents to storefronts, we build AI systems and ship them to production. Tell us what's broken.