I design AI for the people who have to run a business with it.
30+ years in digital, since 1995. I build and ship AI systems and agents that actually work, then turn ideas into digital businesses that earn without constant supervision.
Let's talk about your case →Mario Montanari is an Operational AI Consultant and Incubator. He builds and ships working AI systems and agents for companies and small businesses, then turns ideas into self-running digital businesses. He has worked in digital since 1995, across every wave of the web, and serves firms and SMEs in Turin, Bologna, and worldwide, in English and Italian.
What he does: AI strategy, custom agents, technical SEO and GEO, process efficiency, AI audits, and one-to-one mentoring.
Who it's for: firms and SMEs that want working AI, not slideware.
How to start: a five-to-ten-day AI audit, the lowest-risk way in. Contact is by form.
The difference: he builds the AI, he doesn't just advise on it.
Who I am.
I design AI systems for companies and small businesses, and rebuild the digital processes that cost time today instead of saving it. Custom agents are part of the work. And when someone searches for an answer in your field, I make sure the name that gets found and cited is yours.
I've worked in digital since 1995, through every phase of the web: from the first lines of HTML to today's models.
I didn't discover AI in 2022.
AI didn't show up with ChatGPT. It arrived piece by piece, inside everyday tools, for more than twenty years before any chatbot. Google's PageRank ranked the web by reading its links, starting in 1998. Amazon's collaborative filtering learned what you'd want from the choices of people like you, starting in 1997. Bayesian filters taught themselves to catch spam from 1998, instead of following rules written by hand.
I didn't read about it later, in a book. I lived through it as it happened, working in the web and in SEO. Today's models are the latest layer of a story decades long. Someone who has seen only the last layer knows the controls, but not the physics of flight. And after a while, it shows.
Primary sources
- Brin S., Page L., The Anatomy of a Large-Scale Hypertextual Web Search Engine, Stanford, 1998. (PageRank)
- Linden G., Smith B., York J., Amazon.com Recommendations: Item-to-Item Collaborative Filtering, IEEE Internet Computing, 2003.
- Sahami M., Dumais S., Heckerman D., Horvitz E., A Bayesian Approach to Filtering Junk E-Mail, AAAI, 1998.
What I do.
Deciding where to invest in AI
Before buying licenses or writing code, we work out where AI earns its keep. I map the processes, weigh the return, design the architecture. What comes out is a document you can decide with, not a promise.
Strategy · 1-3 weeksAutomating what eats your time
Custom agents, built on Claude, GPT, and Gemini, wired into your ERP, CRM, email, and documents. By the end of the quarter, someone works hours less and makes fewer mistakes. No "we use AI" banner on the homepage required.
Agents · per projectGetting you found by Google and AI
When someone searches in your field, on Google or inside ChatGPT and Perplexity, I make sure they find you and cite you. Technical SEO and GEO: the same method as always, updated to how people search now.
→ On this, I released an open methodVisibility · ongoing
Finding efficiency where you don't expect it
Often the right answer isn't AI at all. It's two systems that finally talk to each other. I take away the low-value work and put people back where judgment counts. When an automation costs more than what it replaces, I tell you before we build it.
Processes · per projectDiagnosing where you stand with AI
In five to ten days I look at what you already use, what works, where budget leaks, and where there's room nobody has touched. You get a report with priorities, timelines, and realistic numbers. It's the cheapest way to find out where you really stand.
Audit · 5-10 daysMentoring on your actual case
One hour a week on your real case: the quote that won't go out, the catalog stuck for months, the automation that needs a rethink. Real problems from your week, worked through one at a time.
Mentoring · weeklyAt a glance
| Service | Timeframe | Deliverable |
|---|---|---|
| Deciding where to invest in AI | 1-3 weeks | Decision document: priorities, expected return, architecture. |
| Automating what eats your time | Per project | Custom agents wired into your ERP, CRM, email, and documents. |
| Getting you found by Google and AI | Ongoing | Technical SEO and GEO: crawlability, structured data, entity signals. |
| Finding efficiency where you don't expect it | Per project | Reworked processes: fewer manual handoffs, clearer ownership. |
| Diagnosing where you stand with AI | 5-10 days | Audit report: prioritized actions, timelines, realistic numbers. |
| Mentoring on your actual case | Weekly | One hour a week on your real case; know-how that stays in-house. |
For a client in the European legal sector, AI citations (mentions inside tools like ChatGPT and Perplexity) and site visits grew 200% in three months. The agent runs on its own: every 24 hours it connects over SSH, runs audits and checks, and tracks keyword positions in Google Search Console. When results start to slip, it doesn't wait for me. It rebuilds what it needs, content, pages, markup, internal links, and brings the numbers back up.
I don't just advise. I build.
As an incubator, I take an idea, or a technology that's just emerging, and carry it all the way to a self-running digital business: from the first sketch to a live service that earns. That's the difference between explaining AI and knowing how to build it. I work hands-on, from strategy to production.
How I work.
I listen first. Not out of politeness, out of method: until I understand the real problem, every solution is wrong from the start.
I organize, I don't scatter. No ninety-minute meetings when twenty will do, no thirty-page plans when one sheet says more.
I work well alone, and I coordinate a team when it's needed. With respect for people and no motivational exercises: since 1995 I've watched them all fail.
When a problem holds out, I don't let go until it gives. If someone pays me to close an issue, that issue becomes mine.
I'm creative in the literal sense of making things. I put together pieces nobody had combined this way before. Not from talent, but because I've seen enough solutions collapse to recognize the ones that will hold.
The experience that comes before AI.
Designed well, it frees hours and cuts costs. But only if whoever designs it has already seen enough systems fail to catch the signals in time.
I've seen those systems fail. Several of them I made fail myself, in the years I was learning. That's the one kind of competence you can't download from a course.
Questions I get a lot.
Where do you start when bringing AI into a company?
How much does an AI consulting project cost?
Will AI replace my employees?
Why a single consultant instead of an agency?
What stays in the company at the end?
Where does my data go? Does the model train on it?
How soon do I see results?
People who've worked with me.
These are a few of the clients I can name. Others stay unnamed, under NDA or for privacy. They don't talk about AI in the abstract. They talk about the method, and the results.
A great background across the whole of web marketing. Precise, reliable, always willing to find the right solutions.
A digital professional with boundless knowledge of marketing, web, UX, and branding. The strategy he put in place changed our lives.
He combines technical skill, empathy, clarity, creativity, focus, and speed: traits that are hard to find in one specialist.
He's fast, he's independent. In a world that runs, Mario doesn't just walk: he runs, fast and far. The ideal person to have on board.
Great professionalism, rare sensitivity and empathy, deep AI competence. He inspired real trust in me right away.
He listened to our ideas and our story, and built something that works and reflects us completely.
One of the few with a real vision for his work: design, usability, focus on the goal. Utterly reliable.
He redesigned the website for my US-market business. A real professional and AI expert, and above all a great listener: he didn't just do the work well, he thought it through with me and steered it toward the goal I wanted. Highly recommended.
He connects skills from different disciplines into a detailed, coherent systems view. Thoughtful and concrete, never superficial.
Experience that few people have. He got his hands dirty on the internet before anyone could even see the business opportunities in it.
He reads a need expressed in a few words: deep experience, with an attention to detail that makes the difference.
Not just a piece of work, but in-depth consulting that guided me through strategic, targeted choices. On schedule.
We work out the strategy, build the tools, then monitor and analyze the results. Expert, with real human qualities.
Well beyond every expectation. The creative spirit made the difference: solutions out of the ordinary, delivered fast.
Mario was the turning point for my ecommerce business. From a lifeless site to a real AI-based ecosystem that runs everything, even the products. Fantastic.
A true digital pioneer. Much more than a partner: a colleague with deep empathy to think things through with.
On every level, from communication to executing the tasks, he has been on top of everything. Excellent comments from anyone who saw my work.
I've never met such an expert AI professional. Critical thinking, a strategic approach, and natural empathy: they gave me authority in my field.
Let's talk about your case.
If you have a concrete problem and want to know whether I can help, write me a couple of lines. You'll hear back from me, personally. If your case isn't for me, I'll tell you, and point you to who to talk to instead.