Vendor-reported figures — source: www.legal.io
Allen & Overy, one of the world's largest law firms with over 3,500 lawyers spanning 43 offices globally, faced mounting pressure to deliver faster, higher-quality legal work across an extraordinarily broad practice footprint. Lawyers routinely handle document drafting, legal research, and client advisory work across 250+ practice areas and more than 50 languages — tasks that are time-intensive and difficult to standardize at scale. The manual effort required to produce consistent, jurisdiction-appropriate legal work across geographies created bottlenecks that constrained capacity and raised delivery costs, limiting how efficiently the firm could serve its international client base.
A&O integrated Harvey, a legal AI product built on OpenAI's GPT-3 large language model, across all 43 offices following a structured pilot launched in November 2022. Harvey was not deployed off-the-shelf — the platform was fine-tuned specifically for A&O's workflows, matter types, and practice areas, with Harvey's founders (formerly of DeepMind and O'Melveny & Myers) developing firm-specific models that could be further specialized for individual client matters or litigation. Lawyers interact with Harvey via natural language instructions to generate legal documents and conduct legal research, lowering the technical barrier to adoption. The phased approach — pilot first, firmwide rollout second — allowed the Markets Innovation Group to validate the model and build internal advocacy before broad deployment.
The trial period produced results significant enough to justify firmwide deployment. Approximately 3,500 lawyers submitted ~40,000 queries to Harvey across 250 practice areas and 50 languages — a volume that Harvey's team characterized as equivalent to nearly a quarter of the firm's daily workload and roughly 80% of its monthly workload.
Have a similar implementation?
Share your customer's AI results and link it to your vendor profile.
Submit a case study →