Charles Russell Speechlys shares practical framework for choosing the right legal AI tools

Charles Russell Speechlys has published new thought leadership setting out a practical, evidence-led approach to assessing and selecting generative AI tools for legal work.

The report, Knowing What Matters: A practical guide to choosing the right AI tools for your firm, draws on the Firm’s own accelerated assessment of five legal generative AI products, carried out over two intensive weeks with 46 practising lawyers and knowledge professionals. The process generated more than 7,000 structured feedback data points and was designed to identify what lawyers actually value when choosing a preferred generative AI platform.

The findings suggest that, once products meet baseline expectations around speed and accuracy, legal-specific capability becomes the key differentiator. In particular, the analysis found that lawyers place significant value on tools that demonstrate genuine understanding of legal complexity, use appropriate legal terminology, respond effectively to legal queries, and support higher-quality drafting, document review and reporting.

The report also highlights the importance of testing AI tools against real legal use cases, involving a diverse group of users, and building assessment frameworks around role, practice area and usage levels. It found that power users may score tools more generously than less active users, and that partners, associates, trainees, knowledge lawyers and different practice groups can have distinct priorities when assessing AI performance.

Lesley O'Leary, Chief Operating Officer at Charles Russell Speechlys, says:

Generative AI is now a core part of how law firms are thinking about service delivery, productivity and the future of legal work. But the decision about which tools to invest in cannot be driven by vendor claims or generic measures of efficiency alone. Our experience shows that firms need structured, comparable data that reflects how lawyers actually work and what they need from these tools in practice.

"This assessment gave us a much clearer understanding of what builds trust, what drives adoption and where AI can add the greatest value. For us, the most important lesson was that legal competence matters. Lawyers are not simply looking for tools that are fast. They are looking for tools that understand legal nuance, support judgement and help them deliver better outcomes for clients.

Tessa Bartley, Head of Legal Technology at Charles Russell Speechlys, says:

We designed this process to move beyond anecdotal feedback and create a rigorous, practical framework for comparing AI products in a meaningful way. By asking lawyers to test tools against real, non-confidential tasks across key legal use cases, we were able to generate clear evidence about where products performed well, where capability gaps remained, and what different groups of users valued most.

"One of the strongest findings was that drafting, legal query and document review capability were highly predictive of overall preference. That tells us that effective AI assessment needs to go deeper than ease of use or speed of response. Firms need to test how well tools handle the complexity, context and structure of legal work if they are going to make confident investment decisions.

The report sets out a practical assessment playbook for firms, including how to use existing usage data to identify priority use cases, structure RFP questions, brief participants, collect feedback, validate vendor claims and tailor implementation and training.

It concludes that structured assessment can help firms make better AI decisions faster, while also creating data that supports more effective adoption, training and long-term value measurement.

The full report can be found here: A practical guide to choosing the right AI tools for your firm.