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.

Strategy, costs & ROI

23 July 2026·5 min read

Google AI Overviews is now available in France: official launch on July 22, 2026 — what your SMB should do

Yes, AI Overviews is available in France: Google officially switched it on with AI Mode on July 22, 2026, in the last major market to get them. Two recent studies show what an AI-generated answer actually does to a click — and what to check this week before it shows up in your numbers.

Read →

AI agents & automation

22 July 2026·5 min read

AI sales proposal generator: what a multi-agent pipeline gets right that a single-model tool doesn’t

“AI sales proposal generator” keeps coming up in sales-team searches: beyond simply filling a template, a pipeline of several specialized agents changes the nature of the work. What to know before choosing between a subscription and a custom build.

Read →

Strategy, costs & ROI

22 July 2026·5 min read

AI adoption in France: Insee’s new figures show how far SMEs are behind

Insee’s July 21, 2026 report puts AI use among French companies at 18%, up from 10% a year earlier — but only 9% of companies under 50 employees use it, against 33% of those over 250. What the gap means for an SME that hasn’t started yet.

Read →

Strategy, costs & ROI

21 July 2026·5 min read

AI agents: France’s competition authority flags an 84% lock-in risk — what it means for your SMB

On July 17, 2026, the French competition authority published its opinion on the AI agents sector: OpenAI, Google and Anthropic together control more than 84% of it, and the regulator warns of a “platformization” risk that could shrink the diversity of what is on offer. What this means for an SMB choosing an AI agent vendor — or trying to get found by one.

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 →