Legora builds legal agents without reading client data
What happened
On November 3, Latent Space published an interview with Legora CTO Jacob Lauritzen on building and evaluating legal AI agents when client data stays confidential. Legora cannot read the execution traces its agents produce for lawyers, so three engineers and three lawyers built a test environment that mimics client work. One model update lifted citation accuracy from 70% to 99% with no change on Legora's side. Later reporting says the team tests due diligence in a mock data room of 10,000 files and sorts production issues into 10 to 20 known failure modes. The interview also covers model choice and product architecture.
Written by AI from the coverage · updated 2 hours ago
Coverage
Follow the reports to see the story from different sides.
- Latent SpaceBuilding AI for Lawyers - Jacob Lauritzen, CTO of Legora
Legora CTO Jacob Lauritzen 分享了法律 AI 智能体的构建经验，重点涉及客户数据保密条件下的评测、模型选择与产品架构。评测团队由3名工程师和3名律师组成，通过包含10,000份文件的模拟资料室测试尽职调查，并将生产问题归入10到20种已知失败模式。
- Latent SpaceBuilding Legal Agents When You Can't See the Data - Jacob Lauritzen, CTO of Legora
Legora 无法读取法律智能体为律师工作的执行轨迹，CTO Jacob Lauritzen 在访谈中介绍了利用模拟客户工作环境构建和测试智能体的方法。团队由 3 名工程师和 3 名律师搭建测试环境，访谈还涉及包含 10,000 个文件的数据室，以及一次模型更新让引用正确率从 70% 升至 99%、Legora 无需改动的案例。
Heat over time
Heat now 9·Comparable peak 10（Nov 3 08:00）·Comparable change over 24 hours –
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.