This bulletin was prepared by our Portuguese member firm Antas da Cunha ECIJA

 Google’s expansion of Gemini Enterprise into the legal sector deserves attention for reasons that go well beyond the features announced. I see this move as a sign of an important shift in the technology market: the growing focus on industry-specific AI solutions, integrated into organisations’ existing systems and business processes. 

This development is consistent with a trend we, at Antas da Cunha ECIJA, highlighted in After the Prompt, our report on digital trends and their legal implications, under the theme “From LLMs to Specialized Models”: the growing importance of specialisation, efficiency and contextual adaptation in enterprise AI. As we noted, the discussion is no longer simply about model size or computing power, but about choosing the right model for the right context. 

Google’s announcement reflects a related development at the enterprise level. The competitive advantage increasingly lies not only in the capabilities of individual AI models, but in how effectively they can be combined with industry-specific knowledge, existing technologies and operational processes. 

There is one aspect of Google’s announcement that I find particularly interesting. Google is not simply introducing new legal AI capabilities. It is building an environment that brings together different tools, platforms, knowledge sources and AI agents, including solutions from other providers. 

This raises a strategic question: who will control the platforms through which businesses access and use AI-enabled legal services? 

Until now, the legal technology market has largely developed around specialised solutions designed for specific activities. The emergence of platforms that integrate and orchestrate these solutions could change the balance between providers, how businesses select technology and even their relationships with legal service providers. 

For organisations, investment decisions are no longer simply about choosing the best tool for each task. They increasingly involve choices about technology architecture, systems integration, data usage, vendor dependency and the ability to retain control over processes that are becoming progressively automated. 

There is another dimension that I believe deserves more attention. 

When an organisation embeds its policies, risk criteria, contractual positions and decision-making rules into AI systems, it is also turning legal and business judgements into operational instructions. 

This has significant consequences. 

A contractual guideline that was previously interpreted and applied by a lawyer according to the circumstances may now determine how a system behaves across hundreds of transactions. A change in risk criteria may affect an entire chain of processes. An inappropriate configuration may no longer result in an isolated error, but instead systematically influence multiple decisions. 

The question is therefore no longer just about the quality of AI-generated outputs. It is also about the quality of the rules embedded in these systems, who has the authority to define them, who can change them and how their impact is assessed. 

This creates a management challenge that cannot be left exclusively to technology or legal teams. It requires joint decisions on processes, risk, investment and accountability. 

It also raises an economic question: as specialised AI systems increasingly perform legal activities on an ongoing basis, value may shift from executing individual tasks to designing, maintaining and controlling the systems that perform them. 

Google’s move is, in this sense, a sign of a transformation that could extend far beyond the legal sector. 

The most relevant discussion is no longer about which tasks AI can perform, but who defines the rules under which these systems operate, who controls their execution and who captures the economic value generated by these new models of service delivery. 

Joana Pinto 
Partner | Digital Contracting & Compliance 
Antas da Cunha ECIJA