Blog
What you need to know before putting AI into your company.
Enterprise AI security, real costs, GDPR and corporate knowledge, told without the hype. There's no brochure theory here: there are decisions already in production and the reasoning behind each one.
How much does it cost to implement AI in a company? The August 2026 numbers
Real market ranges for a custom build, what companies in Spain actually pay, where the budget really goes (spoiler: not the model) and the five questions for reading a quote without surprises in month six.
Read article →10 questions to ask someone who is leaving, before they leave
A key person is leaving and you have one hour to rescue what they never wrote down. The questions that work and in what order, what to do with the recording and the three mistakes that ruin the session. Printable, to take into the meeting.
Read article →Tacit knowledge: what it is, why it gets lost and how to capture it
The know-how your team uses every day and has never written down. How it differs from explicit knowledge, why documenting more doesn't fix it, four methods that do work and which part of this AI can genuinely do.
Read article →The context layer: what your AI sees, where from and with what permissions
What context engineering is, why it isn't the same as RAG or a chatbot's memory, and what all of that means at a forty-person company with no data team and no plans for one.
Read article →AI knowledge management: what really changes compared to a wiki
Documenting is a separate job that competes with the real work and always loses. What changes when the system captures instead of asking, and the four things to demand of it so it isn't another graveyard of documents.
Read article →AI and GDPR: where your company's data ends up when you use an assistant
Who is really accountable, what paperwork has to be signed before handing over a single piece of data, and what happens the day someone exercises their right to erasure over information that is already vectorised.
Read article →Questions you should ask any AI vendor for your company before signing
Five uncomfortable questions that separate a serious product from a future problem: where permissions are applied, what happens to sensitive data and what happens to your data if you leave.
Read article →Why most enterprise AI assistants can leak what they shouldn't
Semantic search finds what is most relevant, not what each person has permission to see. That's how the most common leak in corporate assistants happens, and how it's prevented by design.
Read article →What it costs to build an AI assistant on your company's knowledge (and what the alternatives are)
Real market ranges in 2026, the maintenance nobody budgets for and why a subscription spreads across many a cost that, built bespoke, you would pay on your own.
Read article →What gets lost when a senior employee leaves (and why nobody measures it)
The exit checklist recovers the laptop, but not the judgement or the context. What exactly is lost with each departure and how to capture it without asking anyone to fill in a wiki.
Read article →