78.4% of Spanish lawyers already use AI, but the debate shifts to rigour and sources, says the Tirant Prime 2026 study

AI is no longer a promise for Spanish legal practice: 78.4% of practising lawyers already use some form of it in their daily work. That is the headline finding of the Tirant Prime 2026 positioning study of the professional legal market, produced by Coto Consulting together with the Universitat de València for Grupo Tirant, based on a survey of 760 practising lawyers across Spain.

The figure confirms a consolidation the profession had been sensing, but the study's value lies in the nuances: mass adoption coexists with rising demands for precision, privacy and verifiable sources, and with significant differences by tool type and by the lawyer's age.

General tools lead usage, specialised tools lead ratings

General-purpose AI applications concentrate the largest share of adoption: 67.9% of respondents use them, ahead of solutions embedded in legal databases (26.3%) and standalone legal AI tools (13.7%). The question allowed multiple answers, so one professional may combine several types.

That usage lead, however, does not carry over into ratings. The report notes that legal-specialised tools score better on the parameters that matter most to practice: technical reliability, source accuracy, regulatory updates and data protection. For SOFIA, Grupo Tirant's own AI integrated into Tirant Prime, the highest marks among its regular users come in privacy (8.28 out of 10), content updates (8.05) and technical rigour (8.03), according to the study.

There are two practical takeaways for firms. First, choosing AI tools is no longer about fashion but about professional risk, something we covered in our checklist for evaluating legal AI software before buying. Second, source traceability becomes a minimum requirement, in line with what the CCBE guide on generative AI and the legal profession demands.

The generation gap in numbers

Age once again makes a difference. Among lawyers under 35, only 9.5% say they do not use AI tools; among those aged 55 and over, that figure rises to 36%. The type of tool also differs: AI embedded in legal platforms has deeper penetration among seniors (32.8% of those over 54) than among juniors (20.9% of those under 35).

For training and internal rollout decisions, the data suggests the challenge is no longer convincing staff that AI is useful, but harmonising usage and review standards across generations working with different tools.

AI transforms research, it does not replace it

The study also measures AI's effect on traditional databases. Among lawyers accustomed to AI, 43.3% report no change in their volume of searches on specialised platforms, 29.7% have reduced it moderately and 16.9% acknowledge a sharper decline. Only 10.1% say AI fully replaces their usual database on specific occasions.

In other words, AI changes the entry point of research work, but the verified documentary corpus remains the destination. A conclusion consistent with the professional standard that requires checking any generative output, as we noted when reviewing the risks of AI-fabricated case law and court sanctions.

As for the selection criteria declared by Tirant Prime users, ease of use (37.2%) and price competitiveness (33.9%) top the list; asked for a single decisive reason, price comes first (21.6%) followed by simplicity (19.7%). The breadth and updating of the documentary corpus earns an average score of 8.01 out of 10.

What this means for daily practice

With adoption now in the majority, the professional debate moves from "whether" to "how": which tool, with what controls, under what internal policy and with what level of human oversight. Firms that have not yet formalised that framework have regulatory and deontological references available, from the Spanish legal profession's AI decalogue to the AEPD policy on generative AI in law firms.

The full study is covered by Economist & Jurist, and reading it together with the generational usage data offers a useful snapshot for planning the firm's technology training into 2027.