Mactores
13 Case Studies
A Mactores Case Study
Sterne Kessler, a law firm, faced a high-volume trademark classification process that was largely manual. Their attorneys spent one to two hours per application across 45 Nice Classes, which was inefficient and resulted in high USPTO rejection rates. Mactores was engaged to design an AI-powered solution to accelerate this work while preserving attorney oversight.
Mactores implemented a production-grade system on Amazon Bedrock that uses hybrid retrieval and multi-agent orchestration with Claude 3.5 Sonnet. This solution reduced classification time from two hours to under ten minutes while achieving over 85% accuracy. The results included a projected $60,000 in monthly savings and a 5-7x return on investment in the first year.