The most useful legal AI news this week was a small round, not a large one. Athena, a startup founded by former Legora staff, raised $2.5 million to run an AI-native legal practice for commercial real estate deals, charging a fixed price per transaction instead of selling software to lawyers. I think this is where the lasting money in legal AI will sit: with whoever sells the finished outcome and keeps the efficiency gain, not with whoever rents a tool to a firm that still bills by the hour.
SiliconANGLE reported on November 3 that Athena Circle Corp. closed a pre-seed round led by a16z Speedrun, the Andreessen Horowitz accelerator, with no other investors named. The co-founders are Cem Ucan and Kieran John Keenan, the chief technology officer, and the founding engineer is José Braz; all three previously worked at Legora. Athena's AI agents handle the document work in commercial property transactions, while the legal service itself is delivered through an affiliated law firm.
The numbers are early but specific. Athena charges a fixed fee per deal that is typically 30% to 50% below standard rates. It signed 22 clients within two months of launching, and says its agents have cut review times by an average of 70%. One client, Mox Living chief executive L. Miguel Arce, received a complete purchase agreement in under 48 hours. Keenan also points to a supply of work: properties financed with short-term loans in 2021 and 2022 are now reaching refinancing deadlines.
The arithmetic shows why the model is attractive. If review time falls 70%, a lawyer spends 30% of the hours a deal used to need. If the fee falls 30%, the firm collects 70% of the old fee for 30% of the hours: 0.7 divided by 0.3 is about 2.3 times the revenue per hour of lawyer time. Even at the deepest 50% discount, 0.5 divided by 0.3 is about 1.7 times. The client pays far less and the firm still earns more per lawyer hour. Under the billable hour, the same 70% saving would simply remove 70% of the firm's revenue on that work, which is why so many firms adopt AI quietly and a fixed-fee practice can price it openly.
Set this beside Legora, which I wrote about in October when it reached an $8.8 billion valuation. Legora sells to law firms. Athena's founders left to sell to the firms' clients. I do not see that as a judgment against tool vendors; they will do well. But a buyer of legal work cares about the price and speed of a closed deal, and the company that controls that price is the one that decides who keeps the savings. Athena's $2.5 million is roughly one 3,500th of Legora's valuation, so this is a bet on structure, not scale.
We made a similar choice at iQuash on the consumer side. iQuash does not sell software to lawyers. It checks a person's ticket, fine or notice for defects and, when there is something worth contesting, connects them with a licensed lawyer. The person sees an answer and a next step, not a subscription, and the efficiency shows up as a lower cost to them. That is the same logic Athena applies to a purchase agreement, at a very different ticket size.
There are real risks. An affiliated law firm carries regulatory duties and professional liability that a software company never does, and 22 clients is a small base on which to judge whether a 70% saving holds once deals get messier. Arce put the standard well: speed only matters if it's quality work. As an investor through ELP Ventures, I would rather back founders who price the result and own that standard than founders who sell hours saved to someone else's billing model.