We’re Adrian, Katon, and Jeremy from PointOne (https://pointone.ai). We’re building an app that automatically figures out what lawyers are doing and generates timesheets for them. Here’s a quick demo: https://youtu.be/yrL3e1hgaNc, and here’s an even quicker one: https://youtu.be/giIaAxZp2M0.
If you’ve ever hired a lawyer, you know most of them bill by the hour—or more specifically, by the 0.1 of an hour (hence our name!). What most clients don’t realize is how painful it is for lawyers to track all their work in 6-minute increments. Lawyers hate time tracking, and many say it’s the worst part of their jobs.
Adrian started out his career as a corporate lawyer at Fenwick & West. The first thing he was taught was how to track and bill his time. Between the 70-hour work weeks and billing to 10-15+ clients per day, staying on top of timesheets is surprisingly hard. To make things worse, law firms are extremely particular about how narratives (that is, descriptions of tasks performed) are crafted—down to the punctuation and diction required. So, Adrian became chronically delinquent in submitting his timesheets, and the firm threatened to take away his bonus multiple times as a result.
Attorney time tracking is not a new problem, and companies have been promising to solve it for years. But pre-LLM attempts at automatic timekeeping never worked as advertised. We were inspired by products like Rewind, and felt that a narrower vertical application could finally solve this problem for lawyers.
Our product is a desktop application that a lawyer turns on at the start of their work day. It runs passively in the background and captures logs from everywhere they work: the OS itself, Word, Excel, calendar, emails, web browser, Slack/Teams, etc. We then clean, pre-process, and interpret the logs. Modern LLMs enable a bunch of cool features. For example, we can pull subtle context from an attorney’s browser activity to associate that work with a client. And for each client and project, we use these models to generate a time entry with a narrative description that matches both the firm’s and the client’s style preferences.
Besides the fact that lawyers hate timekeeping, using PointOne lets them be sure that they’re not letting time slip through the cracks, and frees up hours per week they can spend on other things. It also helps firm leadership by getting more consistent narratives, and faster timesheet submission.
Given the sensitivity of the data captured, privacy and security are massively important. As such, we have customizable data retention periods, we do not use firm data to update models, and we encrypt all data (in addition to employing other standard practices for processing confidential data).
Since our app primarily works for legal workflows, it might not be super useful for most people here (maybe some though!). We would love it if you could check out our demo video, leave your thoughts in the comments, and introduce us to any lawyers you know.
We've come so far since the slave driver with the bullwhip.
If there’s anything wrong with this comment it’s that there should have been a link, so I’ll fill in the one and only gap: https://en.m.wikipedia.org/wiki/Time_and_motion_study
Appreciate the wiki link, as well.
Cheers!
Firstly the slave driver thing. I understand there is evidence in Egyptian pyramids and chinese canal workers where the whip was a rota based position. If you are a group of canal workers you can get paid well to rope to a boat and drag it along the canal to get to market quicker.
And as it takes say a dozen fit young men to drag the boat, any one of them can shirk and still get paid. The whip was used by the workers themselves to punish shirkers, and ensure everyone was “pulling their weight”. Importantly you would need to be a skilled “puller” to work out who was shirking and who not.
(Please note I am not saying chattel slavery did not exist, that brutal punishments were made and violence used to keep slaves and workers under control. Just that in some cases the whip was a communal decision)
Me I hate timesheets. And I even have gone so far as to use my git commit logs as my own timesheets. I just see this as an extension of that.
We leave digital footprints everywhere. I expect a computer can pick up those footprints and provide me with a clear accurate description of my whole day.
I don’t think that is a bad thing. I can even imagine it being useful epidemiologically
But there is case for misuse. And that’s a different class of problem. It’s not ok to have delivery drivers think they need to pee in a bottle else they won’t hit their targets. But the measurement of a delivery driver around town is not the problem.
It’s who holds the whip - that’s the problem
But are you sending data outside my machine for processing? Some asshole is going to argue I waived privilege. We all know that's bullshit but convincing my firm to pay for this without very specific words in the contract around your access rights to the data would make this a non-starter. And that seems to vary state to state.
Good luck! Great use of an LLM.
If you can not answer with a firm YES here, and have to lawyer wriggle out your answer, then this is a non starter.
Edit: formatting
- Lord, why did I have to die so young at 45, in the prime of my life?!
- Hmm, let me check our records, what's your name, John Smith? ... John Smith ... here it is! You say you are only 45? According to your billable hours you are past 112.
A few suggestions: can they install it on prem? (The LLM “server”) Can they install it in their Azure or AWS tenant? Can it gather data from popular DMS and PMS? On prem and cloud? What about geolocation? Data likely needs to reside in the country of the entity if that country has sophisticated data laws What about things like information barriers - if anything sensitive is going where your teams eyeballs are / permissions allow, how do you deal with that?
Finally as I’m not overly familiar with LLMs I wonder how reliable it is - lawyers will expect perfection and all I hear about LLMs is hallucinations
There is a beach head of smaller law firms who do not have to go through stringent data security practices. That would be a GTM I'll consider.
I'm always on the road, so very few of these 'track my screen for billing' tools appeal to me. Tenths, though, has been a good fit for my lifestyle. It's honestly helped me with my ADHD, too.
Good luck on your endeavors!
Tenths is only on iOS, the website doesn't have much other info: https://apps.apple.com/us/app/tenths-billing-time-keeper/id1...
Keys are labeled with the organizations logo or name. Hitting it once starts the clock, hitting it again stops the click. Time is added to an overall total for each button/org.
The keyboard comes with a simple software that allows programming the keyboard and logging time.
Keys change colors when on/off. Multiple keys can be on at once.
I must be missing some key domain knowledge to not understand why the above isn't used. It's very similar in essence to an old school hole puncher clockin out machine.
I've seen lab techs count different cells using a specialized counter where each button increments a different value while looking down a microscope.
The issue is when you're deep in work you might forget to toggle the switches. Maybe this is the biggest issue.
https://timeflip.io/
https://timeular.com/tracker/
I’m curious, though, how accurate is your llm for things like trying to “pull subtle context from an attorney’s browser activity to associate that work with a client”? I suspect this needs to be damn near bullet proof, otherwise attorneys will tear it apart. We stuck purely to file tracking, comparing diffs and syncing activity in real time on a an app with GitHub like heatmap that could also flag suspicious time entries. It worked great but we had a hard time getting adoption because people did not want to force their attorneys to use it. Interesting to see the times change dynamics, good luck!
I don't think I ever submitted my timesheet on time
Essentially it uses the accessibility features on MacOS to see what you are doing and generate time entries for you.
https://timingapp.com/
So now you’re over budget on this task. You can write down the actual amount of time the task took this time, if you want, because no one will be upset if you took extra time on only one task. But you also know that you’re likely to go over on several more tasks this week/month because you’re not only time-constrained, you’re also competing to be one of the better people at your firm in order to get noticed. (Or, alternatively, you’re just trying not to appear incompetent). So you decide to write down the “appropriate” amount of time for the task instead of how long it actually took.
So then the problem becomes that you not only have time budgets for individual tasks, you also have monthly/yearly minimum billed hours requirements to meet. So you’ve just worked 16 hours on a task you said you only worked on for 12 hours. You’ve lost 4 hours that you will have to make up somewhere by working extra hours unreported.
This becomes a pattern very quickly, especially for people who want to prove themselves by delivering a high quality work product. You’re squeezed from all directions.
In fact, my timesheet often has almost nothing to do with how I actually spent my week. Rather it’s just me attempting to fill it in in a way where no one is angry- clients, bosses, colleagues, etc. It’s not a software problem. It’s a systemic/incentives problem.
It's my go-to for hourly "clock your hours" work
Often, it also comes down to the idiosyncrasies of the client and their org structure! I wonder if part of the solution will include using the actual invoices sent to a client to train how future invoices to that client should be prepared.
Good luck to you!
Lawyers have to track their time in short increments and give detailed descriptions for what they did with that time because they charge obscenely high rates and many lawyers have a history of padding their bills. If their rates were lower, they could get away with generic descriptions and less granular timekeeping.
For example, accounting firms can, and frequently do, get away with stating "Preparation of [x] tax return" as the only work detail on an invoice, because they charge a fraction of what law firms charge. I recently approved a $75k invoice for "Preparation of FY tax provision" without any pushback because I know the firm spent hundreds of hours on it and probably swallowed a bit of billable time.
Other tie-in's include task switching and context saving. Remember Eclipse Mylyn would save the state of the IDE associated with a task, and even associate that with a bug so another developer could pick up the world of the task.
Indeed, rather than integrating with other applications by snooping them (raising confidentiality and security issues), you might consider making a legal workspace that can publish out, producing word docs or pushing url's to browse, etc. That way, people accept that everything in context is monitored, and you can automate the crap out of any use case.
The Attorney-IDE is also how you can secure the environment, encrypting everything and even ensuring that information doesn't flow from one client to another. That could become a sales point for them to mention to their clients: that all client information is separately secured, with keys revocable at a moment's notice.
The key feature for usability of this information is seamless provenance for secure internal collaboration: all information should come with information about where it's been, who's seen it, etc. (Just don't say blockchain). If you want to add features for the managing partners to feel like gods, give them the ability to search all contexts without decrypting, so they can implement some policies while maintaining confidentiality for clients.
Business-wise, you can start with a few firms via high-touch custom integration with their policies and practices, and gradually productize through generalization and customization/assembly as you move to low-touch sales model. I suspect each use-case gets system/AI customization you can deploy out, so with luck you'll always have leading-edge high-touch feature development followed by broader deployment.
If you want the (monitored) attorneys themselves as stakeholders, layer in features for continuous tracking, to identify time-wasters (admin like time-tracking, new technology evaluation like research via Lexus vs. some new internal AI) and even negotiate or enforce ground rules (associates limited to 60 hours in exchange for lower salaries...)
Love all the detail here.
If you were shopping this around to firms and they loved it, you wouldn't need to be here - they'd be throwing money at you.
We do have a couple different approaches to filter out non-work. As far as firm management, we think there's a lot of useful information in time data that firms can leverage, but at the same time we don't want this to turn into spyware so we're mindful of only exposing aggregated/anonymized data
If you can make this work and get adopted, I know at least one person who would be very happy.
As to getting product-market fit, you will have a huge issue getting traction in big law firms unless you can run the app on-premises or on the local machine, or provide some serious (read: insane) security guarantees. There are simply too many information security issues that can be existential to firms and their clients to trust any startup with cloud solutions, whatever dept it is in, M&A, IP, Litigation, or even T&E. (I've tried to suggest solutions many times for many of their information mgt issues, and it is a hard NOT EVEN CONSIDER). Perhaps if you get bought by one of the major cloud firms, they might consider it; e.g., I know they are using MS products, but IDK if they are self-hosted or not). Small firms without those kinds of issues are your best bet right now, but even some of them have litigation issues with serious InfoSec issues. Perhaps (this is just speculation) dedicated and isolated server racks for each client firm and VPNs could work. In any case, put that up-front in your sales pitches.
The other thing you'll need is connections to the phone system and the lawyers' mobile phones to at least capture call logs and link them to clients. This should be easier as the firms install partitions with their own software on the phones.
The other thing that takes almost the most time, is composing the description for each time block for the client to read. These entries must essentially "sell" the time block to the client, informing them of the work that was done in a way that they agree that it delivered value. If this part isn't done well, the clients tend to challenge line items on the bill, consuming valuable time and getting write-offs. Presumably, you could train the LLM on all past bills to pre-generate good descriptions for each time slot, which would help as it's easier to edit than to write from scratch. Also, doing anti-training on past entries that were written off may be very helpful (at this point, they just have forbidden word lists).
You are definitely on to a hot market, and I hope you can get to market fit really soon — may the wind be at your backs!
We completely agree about supporting large firms, either with on-prem or deployment within their cloud environment.
I used to work a timesheet intensive job (consulting). The quiet rule was log the minimum 45h/w if you like your engagement; log your actual hours if you hate it.
When my lawyer bills me 1 hour, I know it doesn’t actually take him 1 hour. That’s just his rate. This has been my general experience with small shops. Do bigger firms pay closer attention to detail?
Even for firms/attorneys that are not super detail-oriented, the software would offer a log of their work as a basis for their bills.
I'm also concerned that automated time tracking is extremely close to automated surveillance and not only "productivity monitoring", but also "productivity judging". What do you think about this?
I'll be looking at the demo wistfully.
No wonder lawyers have such fancy suits and a reputation for overcharging :)
(It's also the basis for the book/movie "The Firm.")
Scary.
https://www.crunchbase.com/organization/digitory-legal
OK all jokes aside, honestly two things at play: (1) Impossible to keep tabs on the whole space all the time. No matter how hard you search, you always miss things. (2) Some amount of (proper) competitive research is really good (for obvious reasons), but at the end of the day--you gotta get on with doing your user research/building/testing testing.
Because I have a hard time believing they can actually extract time increments and higher-level tasks from log data without a ton of pre/post-processing. But then the problem is just as much work as it was 5 years ago when you might have been using plain old BERT.
Does this mean that historically, Lawyers have been leaving money on the table by billing less time than they spend? /s
I wonder how LLM can be used to help less fortunate people while still allowing for a breaking-even or even profitable business.