Vendor-reported figures — source: legal.thomsonreuters.com
The Legal Aid Society of San Bernardino operates with a 45-member nonprofit team handling one of the most resource-constrained environments in public interest law. Their legal advice hotline receives nearly 27,000 calls annually — covering urgent matters including domestic violence, homelessness prevention, and guardianship — yet only two paralegals and two to three attorneys staff the line. Each call required a minimum of 30 minutes, often stretching to an hour, as attorneys needed to research applicable law on the spot across a wide range of practice areas. The result: roughly 16,000 calls went unanswered each year, leaving vulnerable community members without legal guidance they had nowhere else to turn.
Legal Aid Society of San Bernardino adopted CoCounsel, Thomson Reuters' AI legal assistant built on large language models and generative AI, to support attorneys and paralegals during live hotline calls. Unlike back-office drafting tools, CoCounsel was deployed directly into the call workflow — giving staff real-time access to legal research grounded in verified Westlaw and Practical Law content. This integration meant attorneys could surface accurate, jurisdiction-specific guidance on the spot rather than scheduling callbacks. Beyond the hotline, the team extended CoCounsel to draft responses to opposing counsel, summarize case documents, review evidence in domestic violence matters, and analyze medical records against DSM-5 diagnostic criteria — expanding its footprint across both intake and active representation work.
CoCounsel enabled the team to double the number of urgent hotline calls handled, cutting average per-call time from 30 minutes to 15. Additional outcomes included:
Critically, these gains were achieved without adding headcount — the same core team of attorneys and paralegals served a materially larger client base. Staff also reported being able to take on a broader range of case types and engage in longer-term advocacy work previously out of reach given time constraints.
Have a similar implementation?
Share your customer's AI results and link it to your vendor profile.
Submit a case study →