From Pilot to Practice: How Law Firms Are Building AI-Ready Workflows

Generative AI (Gen AI) has moved beyond the experimental stage. Across Europe, law firms and in-house teams are no longer asking whether to adopt AI but how to scale it responsibly and effectively. The shift from pilot to practice is defining 2025.

1. From Pilots to Purpose

Many firms have spent the past year testing Gen AI tools in controlled environments. BonelliErede’s Vittorio Pomarici cautions: “There is currently a gap between expectations – everybody thinks that AI can be a magic trick to do the work with no effort, but this is not true because the tools have not yet reached the level of reliability and quality of the results.” ¹:

His firm’s beLab unit screens all AI solutions before implementation to gauge suitability.

Such pilots are essential to separate genuine value from hype - a point echoed by Jeff Pfeifer, Chief Product Officer at LexisNexis:

“2024 was the year of experiments and pilots. As we look ahead, 2025 is shaping up to be the year where legal tech - and especially generative AI - goes mainstream.”

2. Integration at Scale

Once pilots succeed, integration becomes the real challenge. As Sébastien Bardou GM CEMEA International BU & VP Strategy CEMEA at LexisNexis, explains:

“The smaller law firms will be able to implement very quickly; for the larger ones, it’s better to implement in waves. You can’t just turn on a switch.” ¹

This staged rollout reflects the principle of responsible AI adoption - focusing on human oversight, transparent processes, and firm-wide alignment. It’s not about deploying a tool but about reshaping workflows and ensuring that legal professionals understand when and how to rely on AI.

Lexis+ AI’s  multi-model LLM architecture and retrieval-augmented generation framework mirror this need for reliability and control. Instead of relying on a single general model, Lexis+ AI combines several specialised LLMs trained on authoritative LexisNexis content, reducing “hallucination” risks and grounding every answer in traceable, cited sources.

3. Lessons from the Front Lines

Firms are increasingly pairing technology pilots with internal education and change-management programs. At Hengeler Mueller, Thomas Meurer notes: “You still need smart people – we are not afraid that AI will replace the human, because in our experience, it’s the user who makes the difference,”.¹ Meanwhile, at RocaJunyent, Beatriz Rodríguez Gómez underlines that AI training should reinforce lawyers’ ability to think critically and question results, not just learn how to use the tools.

Mathieu Balzarini, VP Product CEMEA at LexisNexis, frames it this way:

“Tech-savvy legal professionals are shaping how law is practised, not just how it’s delivered.”

That human-in-the-loop philosophy sits at the core of LexisNexis’ RELX Responsible AI Principles - embedding human oversight into every stage of AI development and deployment.

4. Focus on Workflow and Data

The Legal Tech Trends 2025 report underlines a new reality: successful AI adoption depends far more on organisational change than on technology itself.
The report describes 2025 as the era of the hybrid builder - firms combining foundational AI models with bespoke, firm-specific components and data frameworks.

Similarly, Lexis+ AI’s Vault feature lets firms securely test AI on their own documents, creating a bridge between institutional knowledge and AI-driven analysis.
This reflects a broader market trend highlighted in the report - the growing importance of capturing and scaling collective knowledge across the organisation.

5. Toward Sustainable Adoption

Beyond efficiency, Gen AI is prompting firms to rethink governance, security, and ethics. The Legal AI Buyer’s Guide  emphasises five pillars for responsible selection: privacy and security, model architecture, accuracy and quality, performance, and ethical AI principles. These pillars are now becoming benchmarks for procurement decisions across Europe and the rest of the globe.

Ultimately, the firms that move from pilot projects to long-term integration will share a common formula:

• structured workflows,
• responsible use,
• data-driven insights, and
• continuous training.

Those that combine these pillars with secure, built-for-law AI platforms, like Lexis+ AI powered by Protégé are not just experimenting with technology - they’re building a new foundation for the practice of law.

¹ Source:

Harnessing Gen AI in law. Lessons from the front lines in Europe. By Ben Edwards. Global Legal Post special report in association with LexisNexis.