供资
Athena 融资 2.5M 与房地产交易律师对等 AI Agents
在 Google 上的首选来源中添加 Unite.AI
Athena拥有250万美元的种子前资金，由a16z Speedrun牵头，用于为商业房地产交易建设AI-本土法律服务。 宣布 November 3，融资支持一种模式，在这种模式中，专有的agents准备法律工作，在一个附属公司有经验的律师仍然负责向客户提供咨询和文件。
称公司在发布的两个月内签下了22个客户。 其附属律师事务所Athena Circle每笔交易收取固定费用，根据公告，通常比客户在最近的交易中支付的数额低30-50%。 Athena不是向一家外部律师事务所出售AI订阅书，而是在相互之间建立技术和服务提供。
资助AI-Native法律服务
商业房地产收购涉及一系列采购协议、尽职调查、融资文件和结算要求。 只有经过审查的律师能够确定问题，谈判条款，支持已完成的工作，快速的初稿才有用。
“大多数法人AI公司向传统律师事务所出售工具。 称 Cem Ucan是雅典娜的共同创始人兼首席执行官。 “从第一天起，我们的AI就与一家围绕它设计的法律事务所携手合作，由有经验的律师集中力量于实际推动交易的判断和谈判。”
固定收费结构使得操作模型成为技术故事的一部分。 当一项交易有商定的价格时，减少重复工作可以改善公司的经济状况，而不必要求客户购买额外的可开票时间。 所声称的节省是同客户以往交易的比较，而不是与每个竞争公司的普遍折扣。
Agents如何为律师准备工作
Athena 称其agents对照交易条款审查协议，找出不一致之处，编写红线和客户端通信草稿。 他们还利用较早的交易来反映客户的谈判偏好。 律师审查产出，作出法律决定，并在交付前批准所有工作。
反馈循环是方法的核心：获取律师更正和客户偏好，以减少以后事项所需的工作。 这描述了在累积上下文的情况下改进法律工作流程；公告没有披露基本的模型提供者、评价数据集或单个更正是否更新模型权重。
“我用律师审查AI agent工作的速度来衡量我们的进展，几分钟后起草的草稿仍然可以让律师工作几个小时。 我们的平台如何帮助我们提供更快的法律服务
早期成果侧重于审查和扭转
在一次早期交易中，雅典娜称的团队在不到48小时的时间内归还了购买协议。 释放后还报告说，律师审查时间在头六周内下降了70%以上。 这些是公司报告的早期结果，不是独立的基准测试，在所有交易类型上都规定了同样的改进。
起草和审查之间的区别很重要。 可以快速生成文件，同时仍然含有含糊不清之处或缺少谈判确定的例外。 减少律师检查和批准所需时间，比仅是生成速度在业务上更有用的措施。
客户经验旨在保持熟悉：向律师发送文件并接受已完成的法律工作。 自动化是这种关系的后盾，法律服务提供者保留了专业问责制。
从乐高拉到最终交易平台
乌坎和基南在勒戈拉相遇。 Ucan的背景包括Baker McKenzie和Legora的Big Law和法律创新。 基南学习法律，在较早的启动阶段建立了自动化法律推理系统，并成为勒高拉的创始评价工程师，致力于起草，研究和代理工作流程的评估。
雅典娜的想法是围绕一个基南的原型而发展出来的，该原型是将数百份文件变成一份尽职调查报告初稿。 何塞·布拉兹(José Braz)开发了勒戈拉的法律研究agent和更广泛的研究能力，他加入雅典娜成为创始工程师。 根据发行情况，基南和布拉兹建造了公司的端对端交易平台。
这种法律实践、评价经验和软件工程的结合与产品的执行有关。 交易工作流程需要的不仅仅是一个可信的答案：它需要一致处理交易事实、文件关系和律师必须行使判决的点。
雅典娜的公开平台新增
雅典娜的网站介绍合作伙伴主导的业务，由专有的内部架构、固定收费以及专注于勤勉、起草和关闭。 它还提出了涵盖合并和收购、风险资本融资和商业房地产等更广泛的做法。 供资公告将商业财产作为初步重点，并将进一步扩大范围描述为一个目标。
a16z Speedrun的方案将投资与客户介绍、招聘支持和旧金山居民相结合。 它的支持使得雅典娜在开发服务模式时可以进入一个启动建设网络，尽管公告并没有提供新资本的详细分配。
雅典娜的主张是，当周边企业的设计是为了从效率中获益时，AI可以改变如何交付法律工作。 下一个检验标准是，在更多的客户和更为复杂的事项中，其早期周转和审查工作是否继续取得进展，而律师继续拥有法律决定权。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Funding
Athena Raises $2.5M to Pair AI Agents With Attorneys for Real Estate Deals
Add Unite.AI to your preferred sources on Google
Athena has raised $2.5 million in pre-seed funding led by a16z Speedrun to build an AI-native legal service for commercial real estate transactions. Announced November 3, the financing backs a model in which proprietary agents prepare legal work and experienced attorneys at an affiliated firm remain responsible for the advice and documents clients receive.
The company says it signed 22 clients within two months of launching. Its affiliated law firm, Athena Circle, charges fixed fees per transaction, typically 30–50% below what clients paid on recent deals, according to the announcement. Rather than selling an AI subscription to an outside law firm, Athena is building the technology and service delivery around each other.
Funding an AI-Native Legal Service
Commercial real estate acquisitions involve a chain of purchase agreements, due diligence, financing documents and closing requirements. A fast first draft is useful only if the lawyers reviewing it can identify the issues, negotiate the terms and stand behind the completed work.
“Most legal AI companies sell tools to traditional law firms. We took a different approach,” said Cem Ucan, Athena’s co-founder and CEO. “Our AI works hand-in-hand with a law firm designed around it from day one, with experienced attorneys focused on the judgment and negotiation that actually move deals forward.”
The fixed-fee structure makes the operating model part of the technology story. When a transaction has an agreed price, reducing repetitive work can improve the firm’s economics without requiring the client to buy additional billable hours. The claimed savings are comparisons with clients’ previous transactions, rather than a universal discount against every competing firm.
How Agents Prepare Work for Attorneys
Athena says its agents review agreements against deal terms, identify inconsistencies, prepare redlines and draft client communications. They also draw on earlier transactions to reflect a client’s negotiating preferences. Attorneys review the output, make legal decisions and approve all work before delivery.
The feedback loop is central to the approach: attorney corrections and client preferences are captured to reduce the work required on later matters. That describes improving a legal workflow with accumulated context; the announcement does not disclose the underlying model providers, evaluation dataset or whether individual corrections update model weights.
“I measure our progress by how quickly an attorney can review an AI agent’s work and confidently stand behind it. A draft produced in minutes can still leave a lawyer with hours of work. Making that review more efficient is how our platform helps us deliver faster legal services with the quality clients expect from Big Law,” said Kieran John Keenan, Athena’s co-founder and CTO.
The Early Results Focus on Review and Turnaround
On an early transaction, Athena says its team returned a purchase agreement in under 48 hours. The release also reports that attorney review time fell by more than 70% during its first six weeks. These are company-reported early results, not independent benchmarks establishing the same improvement across all transaction types.
The distinction between drafting and review is important. A document can be generated quickly while still containing ambiguities or missing a negotiated exception. Reducing the time needed for an attorney to check and approve it is a more operationally useful measure than generation speed alone.
The client experience is intended to remain familiar: send documents to an attorney and receive completed legal work. The automation sits behind that relationship, with professional accountability retained by the legal service provider.
From Legora to an End-to-End Deal Platform
Ucan and Keenan met at Legora. Ucan’s background includes Big Law and legal innovation at Baker McKenzie and Legora. Keenan studied law, built an automated legal reasoning system at an earlier startup and became Legora’s founding evaluations engineer, working on assessments of drafting, research and agentic workflows.
The idea for Athena developed around a prototype Keenan built to turn hundreds of documents into a first-draft due diligence report. José Braz, who developed Legora’s legal research agent and broader research capabilities, joined Athena as founding engineer. According to the release, Keenan and Braz built the company’s end-to-end deal platform.
That combination of legal practice, evaluation experience and software engineering is relevant to the product’s execution. A transactional workflow requires more than a plausible answer: it needs consistent treatment of deal facts, document relationships and the points at which a lawyer must exercise judgment.
What Athena’s Public Platform Adds
Athena’s website describes partner-led engagements supported by proprietary internal architecture, fixed fees and a focus on diligence, drafting and closing. It also presents a broader practice spanning mergers and acquisitions, venture-capital financing and commercial real estate. The funding announcement centers on commercial property as the initial focus and describes further expansion as an ambition.
a16z Speedrun’s program combines investment with customer introductions, recruiting support and a San Francisco residency. Its backing gives Athena access to a startup-building network as it develops the service model, although the announcement does not provide a detailed allocation of the new capital.
Athena’s proposition is that AI can change how legal work is delivered when the surrounding business is designed to benefit from efficiency. The next test is whether its early turnaround and review gains persist across more clients and more complex matters, with attorneys continuing to own the legal decisions.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.