Notes on AI that actually helps.

What we learn building AI tools for real businesses — written in plain language, with real examples. No hype, no names.

AI agents & automation

3 September 2026·5 min read

Claude Fable 5.1: cheaper AI agents and data that stays on your own cloud — what it means for your SMB

On September 1, 2026, Anthropic cut Claude Fable 5.1 cache-read pricing by 75% (up to 45% less for agent-heavy workloads) and announced Enterprise Frontier Safeguards, which lets activity logs stay in the customer’s own AWS, Azure or Google Cloud account. Two moves that speak directly to what SMBs worry about when deploying AI agents: cost and data governance.

Read →

AI agents & automation

26 July 2026·4 min read

Claude Opus 5: Anthropic halves frontier AI pricing — what it changes for your SME’s AI agents

On July 24, 2026, Anthropic released Claude Opus 5 at half the price of Fable 5, with a new “effort” dial to trade cost against quality task by task. Real savings are narrower than the headline — here is how to check yours.

Read →

AI agents & automation

21 July 2026·5 min read

Retrieval-augmented generation: the technical layer behind an agent that answers correctly

Ask an AI assistant a question and it answers from what it memorized during training — right up until it invents a plausible-sounding but wrong answer. Retrieval-augmented generation is the architecture built to fix that: search first, generate second. What it is, and what to check before an agent runs on it.

Read →

AI agents & automation

21 July 2026·5 min read

Why AI agents hallucinate — and how to catch it before it reaches a customer

A support agent that invents a refund policy, a compliance agent that cites a regulation that does not exist: hallucination is not a bug that disappears with a better model, it is a structural property of how language models generate text. What the research says, and what to build before an agent goes live.

Read →

AI agents & automation

21 July 2026·5 min read

Multi-agent systems: when splitting one AI agent into several actually helps

One agent that does everything, or several narrow agents coordinated by a supervisor? The multi-agent pitch is everywhere, but coordination has a real cost. What the research on multi-agent frameworks shows, and how to tell when splitting an agent genuinely pays off.

Read →

AI agents & automation

21 July 2026·5 min read

Autonomous web agents: what a benchmark says about the gap between the demo and reliable monitoring

An agent that browses a competitor’s site, checks stock or compares prices looks flawless in a demo. A realistic benchmark measured how often that actually works end to end — and the gap with human performance is the whole story.

Read →