For more than half a century, the billable hour has been the atomic unit of legal commerce. Lawyers track their days in six-minute increments — 0.1 of an hour — logging every phone call, every email review, every bathroom break that interrupts a research session. It is tedious, imprecise, and universally loathed. It is also a $300-billion-a-year billing mechanism that no one has managed to displace.
Until, perhaps, now.
PointOne, a San Francisco-based startup, just closed a $13.5 million Series A round to build AI that automatically tracks what lawyers do all day and generates billable-hour entries without the lawyers having to do it themselves, Business Insider reported. The round was led by Emergence Capital, with participation from Y Combinator and a roster of angel investors that includes partners at major law firms. The company’s pitch is deceptively simple: let the machine watch the work and write the time entry. But what PointOne is really doing is building an observability layer over the entire practice of law — and that has implications far beyond saving associates twenty minutes of diary-keeping at the end of each day.
The problem PointOne is attacking is ancient by tech standards but stubbornly persistent. Lawyers are supposed to contemporaneously record their time. Most don’t. Studies have consistently shown that attorneys who reconstruct their timesheets at the end of the day — or worse, at the end of the week — lose between 10% and 30% of their billable output to memory gaps and rounding errors. For a firm billing $1,000 an hour per partner, that leakage adds up fast. A 500-lawyer firm losing even 15% of capturable time is leaving tens of millions of dollars on the table annually.
The irony is sharp. Law firms are profit-maximizing enterprises that have spent decades optimizing associate leverage ratios, realization rates, and collections cycles. Yet the foundational data input — how much time was actually spent — remains largely a manual, honor-system process conducted by exhausted professionals at 11 p.m.
PointOne’s software sits in the background of a lawyer’s workstation, monitoring application usage, document access, email activity, calendar events, and communication patterns. It then uses large language models to classify that activity against client matters and generate narrative time entries that conform to each firm’s billing guidelines and client-specific requirements. The lawyer reviews and approves. That’s it.
“We’re not replacing the billable hour,” PointOne CEO Rena Gao told Business Insider. “We’re making it accurate.” It’s a deliberately modest framing for a product that could fundamentally alter the economics of legal practice if it works at scale.
And the timing matters. The legal industry’s relationship with technology has shifted dramatically in the last eighteen months. For years, Big Law treated AI as a curiosity — interesting for document review, maybe useful for contract analysis, but nothing that touched the core business model. That changed when generative AI tools started producing passable legal memoranda and first drafts of contracts. Suddenly, managing partners who had spent careers ignoring IT budgets were asking their CIOs about large language models.
The venture capital market has noticed. Legal tech funding, which had cooled during the 2022-2023 downturn, has rebounded sharply. Harvey, the legal AI company backed by Sequoia Capital, raised $100 million at a $1.5 billion valuation in 2024. EvenUp, which uses AI to generate demand letters for personal injury firms, has raised over $250 million. Casetext was acquired by Thomson Reuters for $650 million in 2023. The market’s thesis is clear: legal services represent one of the largest remaining sectors of the professional economy where AI can compress labor inputs dramatically.
But PointOne is doing something different from most legal AI startups. It isn’t trying to do the legal work. It’s trying to measure it.
That distinction matters enormously. Tools like Harvey and CoCounsel aim to augment or replace the substantive analytical work lawyers perform — researching case law, drafting motions, summarizing depositions. Those products face significant adoption headwinds because they threaten the labor model that generates law firm revenue. If an AI can draft a brief in two hours that previously took an associate twenty, the firm either bills less or faces uncomfortable questions about the value of the output. PointOne sidesteps that tension entirely. Its product doesn’t reduce billable hours. It captures more of them.
This is why the company has found early traction at major firms, according to people familiar with its client list. For firm management, PointOne is a pure revenue play: more accurate timekeeping means higher realization, which means more cash collected per hour worked. For associates, it eliminates the single most despised administrative task in their professional lives. For clients — and this is where it gets more complicated — the value proposition is less obvious.
Corporate legal departments have spent the last decade pushing back on billable-hour billing, demanding alternative fee arrangements, fixed fees, and success-based pricing. If PointOne makes the billable hour more efficient and more thoroughly captured, it could actually increase the total amount clients pay. A tool that helps lawyers remember to bill for the twelve-minute phone call they forgot about isn’t exactly a cost-saving measure for the company on the other end of the invoice.
The counterargument, which PointOne and its investors make, is that transparency benefits everyone. If time entries are generated by software observing actual work product rather than by human recollection, the data is inherently more auditable. Clients can see exactly what was done, when, and for how long. That granularity could reduce billing disputes and build trust — or it could give in-house counsel a far more detailed basis on which to challenge specific entries. Either way, the information asymmetry that has always characterized law firm billing gets compressed.
There’s a surveillance question here too, and it’s not trivial. PointOne’s software monitors everything a lawyer does on their machine during the workday. Every document opened, every email sent, every Slack message, every web search. The company says all data is encrypted, processed locally where possible, and never shared with firm management in raw form — only as proposed time entries that the individual attorney controls. But the architecture requires a level of workplace monitoring that would make most knowledge workers uncomfortable.
Law firms, however, are not most workplaces. Associates at major firms already operate under extensive surveillance — badge-in data, document management system logs, email monitoring for conflicts and confidentiality. The culture of Big Law has always involved a tacit surrender of privacy in exchange for compensation. Adding another monitoring layer may not provoke the resistance it would at a tech company or a bank.
Still, the ethical dimensions are real. Bar associations in multiple states have issued guidance on AI use in legal practice, generally requiring lawyers to maintain supervisory responsibility over AI-generated work product. Time entries are a form of work product — they’re representations to clients about services rendered. If an AI generates a time entry that overstates or mischaracterizes the work performed, and the lawyer approves it without careful review, that’s potentially an ethical violation. The “lawyer reviews and approves” step in PointOne’s workflow isn’t just a UX feature. It’s a liability firewall.
The broader market context for PointOne’s raise is instructive. Emergence Capital, which led the round, has a portfolio heavily weighted toward vertical SaaS companies that target specific industries with tailored software. Their bet on PointOne reflects a thesis that legal practice management is ripe for a new generation of tools built on LLM infrastructure, not just traditional workflow automation. The firm’s previous investments include Salesforce (very early), Zoom, and several enterprise software companies that defined their categories.
Y Combinator’s involvement is notable for a different reason. YC has historically favored companies with massive total addressable markets and consumer-scale ambitions. Legal timekeeping is a niche — but it’s a niche that touches every practicing lawyer in the country and connects directly to the revenue line of a $350-billion domestic industry. If PointOne can establish itself as the default timekeeping layer, it would have access to the most granular dataset ever assembled on how legal work actually gets done. That data — anonymized, aggregated, and analyzed — could be extraordinarily valuable for benchmarking, pricing, and workforce planning across the industry.
So the real question isn’t whether AI-assisted timekeeping will become standard. It almost certainly will. The real question is what happens next.
Consider the second-order effects. If every minute of legal work is accurately captured and categorized, firms will have unprecedented visibility into how their lawyers actually spend their time. Which associates are efficient? Which partners are padding? Which practice groups generate the most revenue per hour of actual effort? This data exists today in fragmentary form, buried in billing systems and partnership reports. PointOne-style tools would make it comprehensive and continuous.
That has profound implications for law firm management. Partnership decisions, compensation structures, lateral hiring evaluations — all of these currently rely on self-reported billable hours as a primary metric. If the data becomes machine-generated rather than self-reported, the politics of Big Law compensation shift. An associate who bills 2,200 hours a year by diligently recording every task looks very different from one who bills 2,200 hours because the AI captured work that would otherwise have gone unrecorded. And a partner whose self-reported hours don’t match the machine’s observations has a problem.
Clients, meanwhile, will eventually demand access to this data — or at least to the analytics derived from it. If a corporation is paying a firm $50 million a year in legal fees, and the firm is using AI to track exactly how that money is being spent, the client will want to see the dashboard. That creates a new kind of transparency that could accelerate the shift toward alternative billing models, even as PointOne’s immediate effect is to strengthen the billable hour.
The competitive dynamics are also worth watching. PointOne isn’t the only company in this space. Intapp, the publicly traded legal technology firm, has been building time-capture features into its practice management platform. Time by Ping, another startup, offers similar ambient timekeeping capabilities. And the major legal practice management vendors — Aderant, Elite (part of Thomson Reuters), and Clio for smaller firms — are all integrating AI-assisted time entry into their products.
But PointOne has a first-mover advantage in positioning itself as a pure-play, AI-native timekeeping solution for large firms. Its architecture is built from the ground up on modern LLM infrastructure, rather than bolted onto legacy billing systems. That matters because the quality of the generated time entries — their accuracy, their narrative specificity, their conformance to client billing guidelines — is the entire product. A tool that generates sloppy or generic entries will be rejected by lawyers who are already skeptical of AI touching their billing.
The company reportedly has contracts with several Am Law 100 firms, though it hasn’t disclosed names. In conversations with partners at firms that have piloted the software, the feedback has been cautiously positive — with emphasis on the caution. “It gets about 80% of entries right on the first pass,” one litigation partner at a major New York firm said, speaking on condition of anonymity because their firm hasn’t authorized public discussion of the pilot. “The other 20% need real editing. But even 80% saves me thirty minutes a day, and that’s thirty minutes I can bill.”
Thirty minutes a day, at $1,500 an hour, is $750. Multiply that across 500 timekeepers and 250 working days. The math is hard to argue with.
And that’s just the direct revenue capture. The indirect benefits — better data for pricing, improved realization rates, reduced write-offs from inadequate time narratives — compound the value. Law firm CFOs, a constituency not known for technological enthusiasm, are reportedly among PointOne’s strongest internal champions at its client firms.
The legal profession’s resistance to change is legendary, and for partially good reasons. The stakes of legal work are high, the ethical obligations are real, and the consequences of error can be severe. But timekeeping isn’t legal judgment. It’s administrative overhead. And the case for automating administrative overhead with AI is about as straightforward as it gets in any industry.
PointOne’s $13.5 million Series A is modest by current AI funding standards — a rounding error compared to the billions flowing into foundation model companies. But the specificity of its target market and the directness of its value proposition make it one of the more interesting bets in applied AI right now. The company isn’t trying to build artificial general intelligence or replace human professionals. It’s trying to solve a $30-billion-a-year data entry problem that the legal industry has tolerated for fifty years because no one had a better option.
Now someone might. The six-minute increment isn’t going away tomorrow. But the human suffering involved in tracking it just might.
WebProNews is an iEntry Publication