Almost nobody in Mexico has actually started with AI. That's your opening.

Key takeaways
- Only 8% of Mexican companies with 10+ employees use AI, against a 20% OECD average (OECD and INEGI, 2025). Adoption is early, and the room is still mostly empty.
- 56% of firms cannot even identify AI's commercial value yet (KPMG, October 2025). The advantage goes to whoever gets specific first, not whoever waits for the perfect plan.
- You do not need a strategy. You need to write down the method that lives in your head and hand its mechanical parts to AI.
- The difference between starting and stalling is almost never budget. It is whether you turned one clever attempt into a repeatable process.
If you are an independent expert in Mexico, a doctor, a consultant, a coach, an educator, or an agency of one, the AI conversation has probably started to sound like noise. Every panel, every feed, every well-meaning post, talk and more talk, and almost no one you know has actually changed how they work. It is easy to read that as a reason to wait. It is the opposite. All that talk is cover. Underneath it, most people in your field are frozen in the "I should do something" phase, and the few who quietly started are already getting hours back every week. Noise is what a head start sounds like from the outside.
The numbers almost no one shows you
Only 8% of Mexican companies with 10 or more employees use AI systems, against a 20% OECD average (OECD and INEGI, 2025). Held up on its own, that picture reads like a country running behind. But a low number can mean two very different things, a race you have already lost, or a room almost no one has walked into yet. For you as one expert, it is the second. Your peers are not quietly outworking you with AI. Most of them have not started either, which means the ground is open in a way it will not be in a couple of years. Turn the number over and it stops being a warning and becomes an invitation.
Why so many can't even name the value
Here is the part that should stop you: 56% of Mexican firms cannot identify AI's commercial value at all, and 40% have not implemented a single project (KPMG, October 2025). Read that as a description of the field around you, and then read it again as a description of an opening. The reason so few can name the value is not that the value is hidden. It is that naming it requires getting specific about your own work, which is exactly the step most people skip. They ask what AI can do in the abstract. You can ask a sharper question: what is the one thing I do every week that drains me and follows a pattern I could describe out loud. Answer that honestly and you have already passed most of your field, because you named a real place for AI to land instead of waiting for a plan to name it for you.
How to start this week
Start by writing down one thing you already know how to do well. This is the whole method, and it is quieter than it sounds. The knowledge that makes your work good lives in your head, in the way you structure a report, the way you frame a first reply, the way you prepare for a session. It repeats, and right now it repeats by hand, every time, powered only by you. Pick one of those recurring pieces and write down how you do it well, in plain steps. Hand the mechanical parts of those steps to AI, the gathering, the structuring, the first pass. Keep for yourself the judgment, what it means and what to keep. Run it once this week, then run it again next week without rebuilding it. That is the move the 8% made, scaled down to one person. It was never a budget or a boardroom. It was writing the method down so it could run again.
One warning
As you start, decide which tools you trust with client information and which you do not. Mexico has no national AI law in force yet, but data-protection rules already govern the personal and client data you feed into any tool. Caring about that from day one is not bureaucracy. It is what keeps a good week from turning into a bad one because you saved a step with data that was never yours to paste.
FAQ
How many companies use artificial intelligence in Mexico?
About 8% of companies with 10 or more employees use AI systems, against a 20% OECD average (OECD and INEGI, 2025). Adoption is early, which for an individual expert reads as open ground rather than a race already lost.
How do I start using AI in my own work in Mexico, step by step?
Pick one recurring piece of work with friction, write down how you do it well in plain steps, hand the mechanical parts to AI while you keep the judgment, run it this week, and run it again next week without rebuilding it. One repeatable process before any strategy.
What parts of my work can I hand to AI?
The mechanical, repeatable parts with a clear definition of correct: gathering material, structuring notes, first drafts, summaries, tidying data. Keep the judgment, the diagnosis, and anything that carries your name for yourself. AI drafts around them; you make the final read.
What AI regulation applies in Mexico?
There is no national AI law in force yet, though several bills were presented in 2025. In the meantime, data-protection rules (LFPDPPP) govern the client and personal data you feed into AI tools, which is one more reason to decide from the start which tools you trust with client information.
- México se queda atrás en inteligencia artificial: solo 8% de empresas la usaOCDE / INEGI, vía Líder Empresarial
- From Potential to Productivity: Latin America in the Intelligent AgeMcKinsey & World Economic Forum
- Empresas en México y Centroamérica apuestan por la IAKPMG México
- Índice Latinoamericano de Inteligencia Artificial (ILIA) 2025CEPAL

Nina Brenes
AI partner for thoughtful experts · Founder of The Harmony Labs
Nina helps thoughtful experts turn what's in their head into AI-powered work that still sounds like them. She reads the research, tests the tools, and sits in the conferences so the people she works with get the short version: what is real, what to try this week, and what to ignore. Her focus is practical judgment, not hype.
Connect on LinkedIn

