Working notes
Blog
Notes and articles about AI, automation and tools.
- RAG: how an assistant finds the answer in your documents An assistant that answers with your data does not memorize it: it looks it up on every question. Pick a query and watch which fragments reach the model.
- Tokens: how a model reads (and why it bills you that way) Models read neither letters nor words: they read chunks. Type something and see how it gets cut, how much it weighs and what it costs.
- Temperature: the knob that decides whether AI makes things up A language model does not pick words: it rolls loaded dice. Turn the knob and watch the odds change.
- What not automating costs you (do the math) Repetitive tasks never show up on an invoice, but you pay for them every month. Move the numbers and see how much yours takes.
- Generative AI for SMBs: a survival guide You do not need an engineering team to use generative AI in your business. I show you how to start with little budget and real results.
- From no-code to code: when to make the jump No-code platforms are amazing for starting fast, but they reach a limit. I tell you how to identify when you need to move to something more custom.
- 5 mistakes when integrating AI into your business (and how to avoid them) The most common mistakes I see when businesses try to implement AI: from expecting magic results to underestimating data cleanup.
- Chatbots that actually sell: from "hi" to a real customer Most WhatsApp bots frustrate the customer within two messages. What it takes for a bot to turn inquiries into sales.
- Prompt engineering: a practical guide to professional results What separates a prompt that returns generic text from one that produces a usable result on the first try. With real examples.
- How AI can automate your business in 7 days A realistic day-by-day plan to go from "I have no time" to tasks that solve themselves with artificial intelligence.