SOFIA 3.0 is Tirant lo Blanch's legal AI, presented in Costa Rica in May 2026. AI Maestro is not an alternative to SOFIA: it is the program that determines what share of your firm's work SOFIA covers, and what to do with the rest.
If your firm just contracted SOFIA, this page is not here to argue with that decision. SOFIA is a solid product from a legal publisher with decades of Spanish-language legal corpus behind it, and a Costa Rican firm that bought it after May's Legal AI Summit bought something real. What this page clarifies is different: SOFIA and AI Maestro do not occupy the same category, and evaluating them as competitors leaves a firm with half the work undone.
SOFIA 3.0 is Tirant lo Blanch's legal artificial intelligence. It is built on the publisher's own document corpus, takes natural-language questions, and is sold to lawyers, notaries, judges, in-house teams and legal advisors. It is available in Spain and several Latin American countries, adapted to each jurisdiction's regulation and case law, and positions itself as the most specialized alternative in the Spanish-speaking market. Tirant formally presented it to the Costa Rican market at the Legal AI Summit in San José on May 13, 2026, at the Hotel Parque del Lago.
Spanish-speaking firms that need legal research over local regulation and case law, backed by a publisher corpus adapted to the jurisdiction.
This is not a list of flaws. It is the category boundary: no software platform is designed to do what follows.
Buying the tool is an afternoon decision. Getting the team to change how it works is a months-long project, and it is the one almost nobody runs. The numbers below are third-party, not ours.
What happens after the purchase
84%
of companies have not redesigned a single role around AI
Deloitte, January 2026. Survey of 3,235 C-suite leaders across 24 countries.
78%
use AI, and more than 80% see no measurable impact on operating profit
McKinsey, April 2026.
37%
use it superficially. Only 34% are in deep transformation
Deloitte, January 2026.
In the 1890s factories swapped steam engines for electric motors and left the floor plan untouched, still built around the old central transmission shaft. Productivity barely moved for thirty years, until someone redesigned the factory around the flow of the work instead of the shaft. Most firms with an AI license today have swapped the motor and left the layout alone.
AI Maestro is a two-month discovery program, not a product. We map how your team actually works, train people in AI fluency, rank the opportunities by impact and risk, and close with a Go/No-Go gate that decides whether building anything is worth the investment. Sometimes the conclusion is that a task does not need AI at all, and that answer is a deliverable too.
SOFIA covers legal research. A law firm is considerably more than its legal research. AI Maestro works on that difference, and on the adoption of what you already bought.
The Process Reality Map shows how much of your team's time is legal research and how much is everything else. In most firms, research is not the largest part.
Role-by-role training anchored in real matters. A tool used by two people out of fifteen did not change the firm's productivity, it changed two people's.
The opportunity ranking evaluates every process by ethical risk, impact and complexity, and sometimes concludes that the right answer is not AI but a classic automation or a process redesign. That honesty is part of the deliverable.
If the context is legal, the version of the program applied to the work of lawyers and paralegals, with the confidentiality and privilege boundary drawn from Stage 1, lives on its own page.
See AI Maestro for law firmsAI governance
Throughout the program we define with your team which data can travel to the models and which cannot, which tools are used and under what rules, and in which decisions a human stays in the loop. The platform recommendation is neutral: we take no commission from any vendor.
No. SOFIA is a legal AI sold under license that answers questions of law over Tirant's corpus. AI Maestro is a two-month process discovery and training program that answers no legal questions and replaces nothing. A firm can have SOFIA and hire AI Maestro with no contradiction; that combination is in fact what gets the most out of the license.
Yes, and probably more than before. A firm that already paid for SOFIA has accepted that AI matters and put budget behind it. What is unresolved is whether the team will change how it works. Deloitte measured in January 2026 that 84% of companies have not redesigned a single role around AI, even though 60% of their people already have access. Having SOFIA and not redesigning the process is the default outcome, not the exception.
It was a legaltech gathering convened by Tirant lo Blanch in San José on May 13, 2026, at the Hotel Parque del Lago, bringing together legal, technology and business professionals to formally present SOFIA 3.0 to the Costa Rican market. If your firm received a legal AI pitch in recent months, it likely traces back there. Worth knowing before assuming the competition for your firm's attention is purely local.
No. AI Maestro sells no software and gains nothing if your firm switches tools. Discovery starts from an inventory of what you already have, SOFIA included, and works around it. Asking a firm to cancel something it just bought is not a recommendation, it is an argument.
Deloitte surveyed 3,235 C-suite leaders: 60% have AI access, but only 34% transform real processes. The bottleneck isn't technology anymore.
Raphaël Dabadie named the new model: software plus service. Traditional consulting runs sampling. AI transformation needs agents that map everything.
Kirkland & Ellis is putting $500M into AI and half of LinkedIn misread it: they aren't training a model, they're building the layer that sits on top of it.
Two 2026 court rulings confirmed it: what your team types into ChatGPT or Claude is discoverable evidence. The problem isn't AI. It's having no policy.
Anthropic analyzed 9,830 AI conversations. Iteration doubles output quality, but teams accept first responses uncritically. How to fix this at your company.
Alex Lieberman mapped 14 stages of AI adoption after 14 months with executives. Most companies skip the first ten and start by building. Then they stall.
Tell us about your project and we'll get back to you within 24 hours
