What AI actually does for European small businesses.

Fundamentals for the non-technical reader, real implementation walkthroughs from our own stack, and commentary on what's actually changing. Written from Mallorca, for European small business owners.

• FundamentalsWhat is an AI agent? A guide for Mallorca business owners

  A plain-English guide to AI agents, what they actually are, and where a small business would start.

• ImplementationHow to install a speed-to-lead system in a Mallorca-based service business

  A walkthrough of the speed-to-lead system: what it does, what it needs, and how to install one.

• FundamentalsWhat the EU AI Act means for your business: a plain-English guide

  What the EU AI Act actually requires for a small business using AI, without the legal jargon.

• FundamentalsData privacy and AI: what a European business owner needs to know

  What GDPR actually requires when an AI tool touches customer data, and where consumer AI plans get businesses in trouble.

• ImplementationThe AI stack we run Viacala on, and why

  The actual tools behind Viacala's own systems, and why we chose each one over the alternatives.

• FundamentalsHow to choose an AI consultant: 10 questions to ask before hiring

  Ten specific questions that catch vendor lock-in, vague pricing, and unfinished systems before you sign anything.

• FundamentalsAI vs. automation vs. AI agents: what's actually different, in plain terms

  A plain-English breakdown of the three terms, why the difference matters for pricing, and how to tell which one your business actually needs.

• ImplementationHow a Palma-based clinic could implement follow-up automation

  A walkthrough of reminders, no-show recovery, and post-visit follow-up for a small clinic, and the stack behind each one.

• FundamentalsHow much does AI implementation actually cost for a small business?

  What actually drives the price up or down, what to budget for a first project, and the red flags to watch for in a quote.

• FundamentalsWhat "AI implementation" actually means (and doesn't mean)

  Why most of what gets sold as AI implementation is decoration, and the three things a real system actually needs.

• ImplementationHow we built our own AI audit system for Viacala, using n8n and Claude

  The architecture behind our own AI audit tool, and the failure mode that broke it first.

• CommentaryThe jagged frontier in 2026: why AI still fails at things that look easy

  New 2026 research says AI's unevenness is structural, still unmeasured, and changes shape with every model generation.

• FundamentalsDo you need custom AI, or will off-the-shelf tools work?

  Enterprises are quietly building AI in-house instead of buying it. That advice does not transfer to a small business, and here is the actual test for which one you need.

• ImplementationSpeed to lead: the three tools we recommend, and how they connect

  The three ordinary tools behind a speed-to-lead system, what each one does, and how they connect, from the stack we run on our own enquiries.

• ImplementationHow we set up client intake automation for Viacala

  How a new person becomes a record, a reply and a booking on our own site, including the people who start the check-up and stop halfway.

• FundamentalsWhat AI cannot do yet: a realistic guide for small businesses

  Where you can rely on AI today and where a person still needs to be in the chair, using 2026 evidence and the limits we build into our own systems.

• ImplementationHow to give an AI agent access to your calendar without breaking anything

  How a daily routine reads our calendar and prepares us for every external call, and the four decisions that keep it from doing anything we did not intend.

• CommentaryWhat the UK AI Security Institute's agent incident means for a small business

  A UK government lab found AI agents took 19 unsanctioned actions online. What it shows, what it does not, and four questions to ask any AI vendor.

• Fundamentalsn8n vs Zapier vs Make: which automation tool for a small business?

  What each automation tool costs as of October 2026, how they count usage differently, and how to choose the one your business can actually look after.

• ImplementationHow we design a lead reactivation sequence

  How to bring quiet leads back without spamming them: segment the list, reference where the conversation stopped, and set firm stop rules.

• ImplementationHow to build a simple lead qualification workflow

  A step by step design for rating incoming enquiries against your ideal client and routing each one, with the rules kept plain enough to inspect.

• FundamentalsChatGPT vs Claude vs Gemini: how to choose for your business

  Skip the ranking. Check the data terms, test each assistant on five of your own tasks, and build so you can switch models later.

• ImplementationHow we run a business with Claude: the folder, the rules and the retrieval table

  Why writing your business down once beats prompting every time, and the folder structure and retrieval table we use to run our own company with Claude.

• CommentaryAI and jobs: what the data says about European firms in 2026

  What Eurostat and the ECB actually report about AI use and hiring in European firms, what the figures leave out, and what it means for a small employer.

• FundamentalsDo you need to learn prompt engineering? Probably not

  For most business owners the answer is no. What you need is good briefing, four habits that do most of the work, and when prompting really matters.

• ImplementationHow we built the AI chat widget on our own website

  How our website assistant works, the limits we gave it, and what to know before building one: a narrow job, server-side keys, a person one click away.

• CommentaryThe EU AI Act now: what a small business must do after the 2026 changes

  The July 2026 Digital Omnibus moved the high-risk dates and softened the AI literacy duty. What changed, what did not, and five steps to take.

• FundamentalsHow AI chatbots work, and why most of them are bad

  What a chatbot really is, why so many disappoint customers, and the five differences between a bad one and a good one.

• ImplementationHow we build a client reporting dashboard on Supabase

  How we structure a reporting dashboard: sources into an automation, clean rows into a database, a chart layer on top, and why we removed Metabase.

• ImplementationHow we automated our own follow-up sequence

  The three workflows behind our own outreach follow-up: the AI drafts, a person sends, a reply stops everything, and a no is recorded.