Hacker News new | comments | show | ask | jobs | submit login
Startup wants to build tools for lawyers to speed up legal services (washingtonpost.com)
234 points by thelock85 123 days ago | hide | past | web | favorite | 101 comments

Headline is a little misleading. We started LTS, our legal tech company, with the goal of helping lawyers provide legal services to startups in a more speedy, transparent and price predicable way. To us, that means building tools that lawyers use, and not creating the Uber of lawyers or a master AI that replaces lawyers (I don't think the latter is possible right now).

Atrium, the law firm founded by two of my LTS cofounders, is putting that software to use to help startups now.

We are hiring engineers (http://atriumlts.com). Some reasons to join: learn from experienced founders who have done it before, and you can build on short iterations for a customer that's very thankful and you can just walk over and talk to.

I just had an incredibly similar thing happen with a slightly misleading headline. My legal tech company[1] was in the WSJ[2] on Monday, and the byline was "A nonprofit has developed an online program designed to help people prepare a chapter 7 filing without a lawyer"... which isn't quite what we are doing. We are actually licensing a product to legal aid clinics that do bankruptcy. The process is actually: legal aid clinic directs people to our software, we guide them through an interview, then we populate a PDF bankruptcy petition that we send directly to the clinic for a final review. So, while we have largely removed lawyer time from the process, we deemed that final review essential. The article did explain this by the end, but the headline got some communities pretty alarmed.

[1] https://upsolve.org

[2] https://www.wsj.com/articles/for-struggling-consumers-a-chea...

Uber for lawyers reminds me of the street walking lawyer skit on Seattle's old Almost Live TV show (with a pre-famous Bill Nye not being a science guy): https://m.youtube.com/watch?v=B7QtRCGQmrc .

> Headline is a little misleading

I guess you can't say "Your article hits a factual error before it even begins" to the Washington Post, huh.

I don't see why you can't. Most of their articles do that. Truth is often mundane, people wouldn't click through.

Sure, but my point was that justin probably doesn't want to antagonize a news outlet that's giving his project effectively free publicity.

Are you planning on building in house AI/ML or do you plan on licensing from existing players in the space?

I don't mean to pile on but I'd be interested as a lawyer dabbling in tech. In the last two years I co-founded startup with a patent pending legal pricing software called Statutory Litigation Automated Pricing, S.L.A.P. (acquired) and this past year I built a prototype "self-serve legal kiosk" and kiosk software Demoed at Walmart HQ for use in their stores with The Law Store (thelawstore.com)

I don't think that makes me an engineer but I am handy at applying existing technologies to the legal field to improve efficency.

Hi Will, I work at Atrium LTS. Please check out our careers page and let us know if there's something that might be a match!


How do you tend to compete with the likes of established players such as westlaw that have been around for so long? I can possibly imagine a startup which helps big company lawyers, as they tend to be more bullish on new tech, but when it comes to the legal industry I don't know.

It's not that I don't believe you guys have a fighting chance at having an impact in the legal technology space, the problem is that the legal industries is one of the slowest moving around, and silicon valley funding culture requires fast results.

>a master AI that replaces lawyers (I don't think the latter is possible right now)

We did this at Recommind (https://www.recommind.com/axcelerate-ediscovery-product-page...), but replacing low level lawyers for case management. A senior lawyer/case manager usually trained the system to identify relevant documents based on a few...

Ok, we reworded the title above using language from your comment.

I'm a software developer (mostly focused on operations and marketing) on a small, two-person team at a large, national plaintiff's firm.

Thanks for clarifying -- the headline was very clickbait and disingenuous -- so much so that I had a strong reaction to it.

I think your plan is a good one because you have a small group of clients with similar needs to focus on and they have certain expectations about the tech they interact with. Good luck!

It'd probably be worthwhile to mention what your Errors & Omissions coverage limits are / will be / can be / etc to the buying public. It'd probably be worthwhile to ask what your D&O coverage might be by comparison. I mean, just apples to different colored apples being offered for sale.

If you're ever looking for PM's this is right up my ally. I was the co-founder and head of product at Naritiv (Techstars/Disney Accel S14, acquired by Hootsuite). I also ran ops as well and dealt with our law firms (corp and entertainment) every week. Would love to talk! Email in profile.

EDIT: HN title has been corrected.

So this needs a change from

"Startup wants to replace lawyers with robots"


"Startup wants to augment lawyers with automation"


There is a startup here in Austin that does seem to want to be the Uber of lawyers : csdisco

Both job links 404.

Headlines are almost guaranteed to be misleading. That's why I click on the comments before the article. Sensationalism should not be rewarded.

I'm finding that the top few comments here are often more informative and better supported than at least half of the actual posted articles. To be clear, I see that as an endorsement of HN more than anything else.

I've complained about some BBC headlines before. I was essentially told that there's a kind of headline exemption. So long as the article at some point contradicts the headline then it doesn't matter, apparently.

Justin – I reached out a couple times but had not heard back.

I'm an international lawyer and savvy developer (currently maintaining knockout.js) and have had a lean/quiet startup for several years now in the legal-augmentation tech space.

We're in the process of partnering with the world's largest law firms.

Feel free to drop me a line - brianmhunt at gmail.com

There's a big unanswered question in the article.

It says that the company is a law firm, that only two of the four co-founders are lawyers, and that it raised $10.5 million dollars, presumably on an equity basis.

However, law firms are not permitted to have non-lawyers as owners. It is considered unethical as it would mean decisions would be made by people that aren't obligated to follow the cannons of legal ethics or subject to bar discipline. The only U.S. jurisdiction I'm aware of that doesn't follow this rule is the District of Columbia, but Atrium is based in California.

We raised $10.5M for our legal technology company, Legal Technology Services (LTS). My cofounders of LTS who are lawyers, Augie Rakow and BeBe Chueh, started a firm called Atrium LLP. Atrium LLP is LTS' first customer, and allows us to reduce our development cycle on building that software. I am not a partner at Atrium LLP (and am not a lawyer, nor do I give legal advice).

It's a complex story that has been a little misinterpreted and simplified.

Thanks for the reply, that makes a lot more sense than how the story was portrayed in the Post.

Justin, with respect, there are many organizations working within the structure that you've put together. All of them get by on the basis that they argue the scope of what software does is not the practice of law.

My concern is that in order to secure regulator approval for this position most of these legal support contractors substantially misrepresent the scope and depth of the work they provide.

Do you have any plans in place to mitigate the risk of crossing the line, and how will that inform your product development plans?

AtriumLTS appears to be working on tools for lawyers, however; wouldn't you agree that's a much easier sell for regulators? As opposed to a LegalZoom or even Clerky model?

Now, if they're trying to gin up some "big data" tools, and using Atrium-client data as inputs for those tools, that raises really sticky questions. There would need to be a whole lot of education and disclosure about to clients about AtriumLTS's access to, and use of, the client's information, and the profits generated from that. You just can't monetize lawyer-client relationships like you do in the finance field.

In my experience, the regulators aren't really being sold in the first place. They don't have the resources or willingness to stir up a hornet's nest if they can get away with it.

Normally the key battle is in determining if the work in question falls within the definition of the practice of law. The more rote and mindless the work is (say, bates numbering a discovery compilation), the safer it is, but most of the low hanging stuff in this area is already done and most of the new entrants aren't trying to provide new tools in that area.

The big new wave is based on using ML and other tools to search, sort and classify documents for contract management, due diligence, discovery, etc. This is an area where support services already operate. As third parties, these support organizations allow firms to launder responsibility. They can claim that the primary firm is responsible for performing the legal work, which allows them to use non-lawyers to perform the work lawyers would, while simultaneously allowing the primary firm to claim that the searches they perform are defensible and therefore even if mistakes have been made that their efforts were sufficient to discharge their responsibilities.

Agreed that this is happening; laundering is pejorative, but AFAICT, the Bar has said it's all kosher so long a lawyer is ultimately responsible for the advice and making the effort reasonable.

So, for our friends at Atrium, whether or not their systems 'practice law,' as long as that product is 'given' by the client's attorney... doesn't that look like it meets the UPL hurdle? (At least at this oversimplified level, IANYL)

Depends on how the regulator interprets the bar on sharing billings with non-lawyers. Westlaw and Lexis have massive offerings that do a ton to make work easier, so there's obviously room to work with, but they aren't providing products that practice law. They provide tools to practice law with.

With ML-driven autonomous determinations of relevance, however, you're literally doing associate legal work 101, with a system applying skill and judgement to the facts of the matter to make a determination of a legal nature. That's a difference in kind. This becomes doubly interesting, because most of these platforms are cloud hosted, so they're not 'given' by the attorney in the way that a land registry look-up would be after being vetted by a clerk. They're a stand-alone service that functionally makes choices for you.

I have no clue if Atrium is going down that route, but that's where the million dollar cost savings are. Maybe they're just making snazzier client-focused billing software. I've asked before, but no one will tell me what products they're actually working on.

Can you expand on the document classification etc - one would have thought that a large company would have all its contracts in one place / data room. For things like discovery etc I can see searching a gazillion emails is classification territory, but as you said that is mostly a done market?

I am just wondering where this next wave really is going going

> one room

At scale, almost no one has a "room." Think about a mortgage bank or cell phone provider, or a car insurance company. Even if they tried to use 'a form', they have to manage hundreds of little details about millions of clients -- based on whatever version of the form was filled out, where they lived, where the product is used, has a deed or title been properly recorded, did the Supreme Court of State X strike down part of your agreement, has anyone sent a written complaint sufficient to trigger the Fair Credit Reporting Act, etc. We've all criticized big lenders who can't come up with proof at a hearing -- but that's because managing that much data is hard. Compliance/contract management software is still growing rapidly.

Second, another wave of tech for lawyers is taking huge databases of publicly-filed contracts, and suggesting clauses in agreements similar to yours. If 80% of compensation contracts filed by publicly traded companies say X, and your deal says Y, you'd like to know that.

Isn't this the same sort of quasi-entity that the patent troll firm that is suing Cloudflare is using?

Not saying that what this gent is doing is even remotely as bad (or even really bad at all), but seems they are employing the same mechanism to skirt the regulations a bit.

Just do it Justin. Make it cheaper for the public to fight their cases and be massively successful. If Uber and Airbnb can get away with it, so can you. If people paid attention to the naysayers nothing would ever change in the world.

> most of these legal support contractors substantially misrepresent the scope and depth of the work they provide.


I work in the field, have money invested in the field and have spoken with both the organizations in question (as well as their compliance counsels) as well as their regulators.

In law, new business models often are in violation, compliance wise, until they make the case that they shouldn't be. If the model is successful & profitable, they fight to change the rules and sometimes end up with a complete or partial victory in that campaign - this is largely a political fight at the bar. If not, they vanish and the lack of regulatory oversight is irrelevant.

Can't give you more than that without going out of bounds.

Reviewing Atrium's legal careers section, they clearly understand the need to monitor the regulatory environment with respect to compliance - see their lead counsel post. That said, I'd counsel justin not to reply to my question if I was advising him.

It's not particularly complex compared to the structure of mutual funds: http://www.yalelawjournal.org/pdf/1228.Morley.1287_zmw59iqn....

Quick question: for jury selection, do trial lawyers used a statistically tested line of inquiry for prospective jurors, or is it just seat-of-the-pants, intuition based selection?

Somewhere in between. In many federal cases, there is little "picking" allowed. And where picking is allowed, it's mostly intuition.

In large cases, jury consultants have typically been hired to use their intuition, with a scientific-sounding gloss, so the lawyers and corporate managers have someone to blame if something goes wrong. They were famous for proposing questions that were supposed to tip you off to something important, like "Does anyone here regularly fill out a crossword puzzle?" Given the tiny samples, it was difficult to have scientific questions, or to prove that the questions were bunk. (I am less jaded about trial consultants, who deal with argumentation, compared to the 'jury consultant' industry -- which famously gave us Dr. Phil).

In today's mega-cases, the money is spent on rapid profiling of the potential jurors' actual opinions, using their public-facing social media. There's still lots of armchair psychology, but not as much based on bullcrap questions.

The crossword puzzle question as quoted could be a logical/grammatical trick-question, because most people who know they don't don't know though if any one of all of them does.

They hire jury consultants, who do those things.

And the jury consultants, in turn, consult statistical tables. For example, one dimension they look at is authoritarianism. A person is deemed "authoritarian" if they are likely to side with the State in a criminal trial.

I helped my buddy, who is an owner of a jury consulting firm, port one of the statistical tables to a ruby model. Here's the code, if you're curious.


We wanted to build software to democratize the jury selection process (and bring justice to all, not just those who could afford expensive jury consultants) but the marketing process was kind of challenging and the lawyers we reached out to were unresponsive. Also, jury consultants want to keep this knowledge a trade secret in order to justify their expensive fees despite these statistical tables being available for purchase by anyone who knows where to look.

Seems like there is enough leeway in ABA rule 5.4 that they are probably covered, but I'll bet some people will contest it.

Augie is a great lawyer so I would be confident that he's thought through all of that.


The ABA rules are model rules - they provide a framework for state bars to build upon, but they aren't directly enforceable unless a state bar decides the ABA model code is their code.

The state bar of california has their own set of rules. See: http://www.calbar.ca.gov/Attorneys/Conduct-Discipline/Rules/...

Where do you see any leeway in this?

(d) A lawyer shall not practice with or in the form of a professional corporation or association authorized to practice law for a profit, if:

(1) a nonlawyer owns any interest therein, except that a fiduciary representative of the estate of a lawyer may hold the stock or interest of the lawyer for a reasonable time during administration;

(2) a nonlawyer is a corporate director or officer thereof or occupies the position of similar responsibility in any form of association other than a corporation ; or

(3) a nonlawyer has the right to direct or control the professional judgment of a lawyer.

I think the organization disclosed above fits into that exception so long as there is sufficient separation between the tech company and the law firm. Structuring the law firm as a customer of the tech company probably qualifies.

That said, I wouldn’t want to be putting my law license on the line for uncharted territory. Ethics enforcement can be really arbitrary.

There are a lot of ways to structure things that would fall into one of those exemptions.


Rules 3.157 (A) A shareholder of a law corporation must be licensed and entitled to practice law.24

(B) The shares of a law corporation must be owned only by that corporation or a shareholder.25

Yeah, California prohibits this.

Just because something is considered "unethical" doesn't necessarily mean it's illegal. My guess, is that by building software no one is really going to question the ethics... In one sense the legal system should be devoid of anything but facts. So as along as a machine doesn't intentionally falsify information (which idk how it could), then it's not unethical.

I was using a term of art without being clear that's what I was doing.

"Unethical" as I was using it means in violation of the state bar Rules of Professional Conduct. Such violations aren't just a matter of being discourteous or something, they subject lawyers to professional sanctions up to and including being disbarred. As for non-lawyers, practicing law without a license is a crime in most states.

This is much more an issue of illegal than unethical in the colloquial sense of the word. My apologies for the confusion.

> Just because something is considered "unethical" doesn't necessarily mean it's illegal.

In the legal profession it sort of does. Lawyers are regularly fined or disbarred for violating ethics rules, and the ethics rules are typically enforced by the courts.

Please don't assert something as fact when you don't know what you're talking about.

A law firm is typically an LLP structure. LLPs are common even outside law firms and there is nothing I'm aware of that says only accredited/bar member lawyers can be the owners of these LLPs. Where are you getting that info from?

The restriction referenced is the bar association rule adopted in nearly all U.S. Jurisdictions that prohibits lawyers from sharing legal fees with non-lawyers.

Some non-U.S. Jurisdictions have similar rules but many don't. For example, there are publicly traded law firms in the UK.

> As anyone who has racked up legal fees due to lawyers’ mounting billable hours can attest, law firms remain almost surprisingly low-tech operations.

That’s like saying programming remains surprisingly low tech because you walk into a tech company and everyone is hunched over an emacs window.

Even at the fanciest Wall Street law firms, 10-20% of billed time is written off. The largest clients who get sued a lot are demanding things like flat retainers (e.g. we pay a flat $5 million per month and you hanldle all our products liability work). There is a lot of pressure to get things done within certain budget targets, but time is routinely written off because those targets cannot be met. Plus, competition is high. Even at the top of the market there are dozens of firms in direct competition. If there was a magic way to do more with less firms would do it.


Whenever someone on HN talks about bringing tech to law firms I inevitably hear they are opposed to improving their process. It would lead to less billable hours for the firm.

Is this true? Not specifically talking about the top firms but firms in general.

Clients are wise enough these days that they just care about the bottom line. They want to know: how much will this case that’s in discovery cost me this quarter (and can you help me justify it to my boss)? Often you give a realistic number and the client pushes back, which is how you end up with 10-20% of time being written off. This is true at every level of firm.

Within this framework, lawyers have an incentive to be efficient. If you cut your hours so you write off 10% instead of 20%, you’ve freed up hours you could spend on new billable work. If you enter into a flat fee arrangement (which in various permutations are very common), the only way you come out ahead if by harnessing efficiencies. I think most law firm managers are numerate enough to realize these things by now.

The challenges to adoption are risk. Law firms can’t push back the release date because the software isn’t ready. That makes them very conservative about changing their processes. Cost-benefit is also an issue. Even the largest law firms are relatively small companies. All major business decisions are made by lawyers, and not even the managing partner works more than part time on administration. They don’t have the bandwidth to implement software X that majorly improves some specific process Y. For the effort to be worth it technology has to deliver cross-cutting improvements.

This was the same way with managed IT services 10 years ago. Most companies now are not billing based on time, but based on the value of the services they provide. Changes started with smaller companies first.

Smaller law firms will handle some things on flat fee now, bankruptcy, DUI, and other transactional issues. It is going to continue to increase when appropriate.

even aside from the billable hours issue, there is the less obvious problem that in many cases, it is in the interest of the client that there be less transparency and more obfuscation of material that might be unfavorable. improving processes and making it easier to analyze data is not necessarily aligned with the incentives of large company who has been named as a defendant in a class action suit, for example.

I wouldn't conflate fancy with sophisticated. Law firms are structurally insulated from market demands and are highly inefficient. They will often just throw people at problems better solved by technology and then pass those costs on to clients. There are a number of reasons for the lack of adoption:

- Almost all firms are LLPs where revenue is passed through to the partners. Investing in technology would take a direct cut out of that, and so even if it would make things more efficient, the partners would rather just make more money.

- Most lawyers bill by the hour. There is a disincentive to invest in technology at any area in the firm which would reduce the number of hours billed. Sure you can cut a fixed price deal, but that is a rarity. The places where law firms have invested in tech (e.g. search tools for e-discovery) are because it would be totally inexcusable to spend that much time on a problem or the clients demand it. For example, I have heard of law firms refusing to work with Dropbox or other cloud storage services, but that is very slowly starting to change.

- Law has been an industry that has been historically slow to adopt any technology, whether it be word processing in 80s, email in the 90s, or now automation. There are a number of reasons for this, from lawyers being in a very hierarchial practice where the partner determines the process and may not be tech saavy to lawyers in general being quite risk averse in order to avoid ethical issues. That industry regulation warps competitive forces just makes this more of a problem. Many law firms are silo'd under partners where the process and the tools used can vary widely.

That sounds to me like a description of what being a lawyer was like in the 1980s. Today, you don’t get cases by calling up a golfing buddy. You go pitch cases to sophisticated clients who compare your proposal with 4-5 other firms’. Alternative fee arrangements account for almost 20% of revenue and growing, and that doesn’t include soft-caps and write-offs that lead to realization being around 85%.

I don’t buy that there is all this great technology out there that law firms aren’t using. There are differences driven by practical requirements for sure. For example, I don’t approve of the cloud, because who owns the hard drive your data sits on is a fact with legal significance. Most people can afford to ignore the difference; lawyers cannot. But it’s not like they eschew remote access to data. Things like RDS and VPN let you access data remotely while controlling ownership of the hard drive where it resides.

How about paying a one-time fee with a little extra to be regarded as the legal owner of a set of mirrored HDDs installed at a facility that provides predominantly cloud-based infrastructure?

Would that solve the ownership problem that would prevent lawyers from considering cloud storage?

> (e.g. we pay a flat $5 million per month and you hanldle all our products liability work)

Question. Is it standard when a client does that to then demand or ask to review time records? The purpose would be to be able to then use that to bid out the work to other firms.

How would the law firm handle the request? Does this happen?

If so I would think it would be an advantage to structure billing and time tracking in a way that wouldn't allow it to be used so that a lower bid could be place since a competitor would know the time requirements.

Yes, it always happens -- which is why large corporations are successfully reducing fees where litigation is routine. Refusing to give any account of the time spent on the matter would probably get the lawyer disbarred.

There's a legal duty to give the client an accounting of your fee, and the basis of that fee, so as to show the fee was "reasonable." And a key metric for reasonableness is the time spent. In failing to account for the time, you would be unable to prove the fee is reasonable -- and the lawyer has no right to keep a fee unless it is reasonable.

You might try to get the client to agree in advance that $5m is reasonable, and to waive any right to you supplying an accounting of your time spent. But I suspect the courts would refuse to enforce that agreement, and the relevant Bar would sanction the lawyer and/or his firm.

If this is an example of scratching one's own itch, Justin Kan needs new hobbies. Consider how much you have to interact with lawyers in order to want to automate them.

"I’ve raised money. I’ve done a merger. I’ve been sued. And yet every time, bills would pile up and I had no clear idea what I was paying for. In Silicon Valley, we want everything to be transparent."

From the sound of the story, their competition is not the Bar, but DoNotPay (mentioned in article), which tells me they're not automating lawyering so much as mechanizing Nolo Press.

What is Nolo Press?

Nolo Press is a publisher who specializes in books that help you do things without using a lawyer, such as writing your own will.

My spouse is in-house counsel (lawyer) for a software company.

Her objective, and the reason she's on staff, is to get stuff done efficiently and cost-effectively.

When she writes a sales contract she tries to write it so everybody can say, "that's fair" and sign it, rather than send it off to some slow and expensive review. Sales people love that.

Same for smaller stuff like NDAs.

She did that stuff for some of my entrepreneurial ventures. Big win.

Litigation: she looks beyond the white-shoe gold-tie-clip downtown firms to handle that kind of stuff.

Escrow for code: If some source code needs to be held in escrow, she says, "look, I'm an officer of the court, answerable to more than just my boss. Suppose you put the thumb drive or DVD-R in an sealed envelope. I'll put it in another sealed envelope and hold it in my locked file cabinet." Many parties say "good idea!" because escrow services aren't cheap.

She has a rubber stamp that says, in some kind of legalese, "We don't pay unitemized bills. Please provide more detail and send this bill again." That can save a bundle of money. The concealed "bonuses" for partners sometimes evaporate from the bills. The "stamp" works well; the billing firm has no idea whether it's legal or A/P refusing the bill, so they just provide more detail.

Financings: not much leverage there. They still have to hire outside counsel with expensive lawyers and paralegals who can bench-press massive tomes of dense legalese.

I suspect a software service that inspected and tracked legal bills payable by tech companies might pay for itself pretty quickly. (Ingest pdfs? Look for keywords and numbers?)

The legal industry is ripe for innovation. Just look at the amount of value Clerky has managed to provide with just incorporation and formation.

Legitimately curious about this:

"Throughout the day, the lawyers field questions from a roster of start-up clients looking to execute routine legal tasks, like fundraising from venture capitalists and issuing stock options to employees. Engineers watch the dealings closely, extracting bits of information from the conversations and the documents exchanged.

So what happens to client-lawyer privilege, if engineers are watching / listening in / reading? Isn't that automatically waived, and isn't this a problem?

(Edit: formatting)

Truthfully as a lawyer , well I am pretty close to be an ex-lawyer cause I now code for living we really dont need software so much. In my office we had an old database running on DOS when I became a lawyer in 2008. I got rid of it and replaced it with Dropbox. Used extremely simple folder structure and that was pretty much it.

I was even considering making specialised software for my father's law firm but I never saw the use for it.

99% of lawyer's work is not arragning meetings or organising documents. Even I who I was very new at the time I did a ton more photocopying documents and answering calls from clients than any form of organisation.

This is why law orientated software fails to become popular for lawyers. What we need as lawyers , what would be really appealing is an AI that can look through case law and help us formulate lega arguments. Unfortunately the technology simply is not at that level.

What makes it worse is that even though coding can be an extremely repetetive process law practices is not, you literraly dealing with diffirent problem each hour and you have to filter and combine various amount of information. You may think that organisation is the key here but because the main focus are legal arguments not so much evidence there is very little to do in that area other than having AI and feeding you ideas.

Problem is that lawyers are using a lot of thinking power and because their is not technical in some cases it can be called even philosophical , it would be very difficult for an AI to be actually useful in this area, unless there is a large leap in technology I am not aware of.

I am not a US lawyer , I am a Greek lawyer and yes we have plenty of legal software in my country as well, that almost none is using. DOS and a dual core at 2.0 GHZ is more than enough.

Couple things regarding law firms and lawyers... based on my experience working at PageVault, a startup in Chicago that creates and runs an online evidence/webpage scraping and gathering service for case discovery for law firms:

1. Lawyers actually don't want tools that will necessarily reduce their billable time. This does not mean they won't ever utilize or accept new technology but it has to be built up over a fairly long period of time before a mass of lawyers all decides they need to adopt whatever the "latest" tool is. Even fax machines took quite awhile to get rooted and they've taken longer to get unrooted (yes, there are still plenty of lawyers who love send and receive faxes, amazingly).

2. For any tech that a law firm is expected to pay for, they want easy means of finding creative ways to bill it back to their client's accounts. This is why lawyers have the reputation they do for charging for things like 10 cents/page for making copies at the copy machine.

Hey justin, good stuff, two questions for you.

A- How is your approach different than Axiom Law [1]

B- Why did you decide to target lawyers vs. directly to startups like Clerky [2]

[1] https://www.axiomlaw.com/

[2] https://www.clerky.com/

> B- Why did you decide to target lawyers vs. directly to startups like Clerky [2]

They didn't. Atrium's customers are (currently) startups needing legal services, not other law firms.

While a lawyer-replacing AI may not be possible yet, it has to be possible to use some ML techniques to aid in the search of relevant legal records, laws, regulations, and so on. I think the world may not be ready for "Uber for Lawyers" but it certainly needs "Google for Lawyers".

Talking to the title for a second - (what they are actually doing seems quite different.)

This is funny, I think we might be underestimating the recalcitrance of human beings, and negentropic characteristics of life itself. (to say nothing of lawyers). After all there are still people operating the barriers at parking structures where I park in the morning, and they certainly cost more than a robot...

It seems that behind the recurrent question of automation and the anticipate crisis of salaried employment is a question of entropy and negentropy, not of autonomy and automaticity.

Law is distinctly negentropic and so is not the best candidate. Avoiding legal recourse would be a different matter ...

> Law is distinctly negentropic

Meaning? In a closed system of your choice, law decreases entropy, i.e. increases complexity? Your comment is quite cryptic to me.

However, negentropy seems to be in violation of the laws of thermodynamics, so you are really saying that the law can't be described as a closed system. Is that right?

I hope that automating a lot of routine legal tasks will feed back into writing laws in way that makes them more consistent and easier to automate. Most people's legal needs are very similar and should be possible to automate.

There is a very big difference between a good lawyer and a bad lawyer. I hope Atrium has the good lawyers on their side while the machines are initially getting trained.

In the meantime..... https://arxiv.org/abs/1707.07328

I hope this leads to a reduction in legal costs, especially for start ups that have been sued for frivolous patent law suits, by patent trolls.

I hope it goes beyond just reduced costs. Increase supply enough so that it outweighs the demand. Until that happens, the legal system will remain mostly a place where the deeper pockets win.

I'm wondering what kind of legal tasks they cover for poor startups. Incorporation and whatnot is mostly automated , trivial and inexpensive but if they can help out with lawsuits, patent, trademarks, and general legal inquiries in an innovative way, it'd be awesome, as those cost a fortune.

I thoughts sales people would be the first to go. I mean, they are are often colloquially/affectiohally known as sales droids in the Valley :)

Kissing aside, if you got laws and said laws can be interpreted to the letter I.e deterministically then maybe this is bound to happen sooner rather than later.

I like the fee structure!

However, I'm curious how they measure and ensure quality in their work. It seems like the fee structure (fixed bids) will incent them to reduce the time spent on work, which could be good or could be bad, depending on the answer to that question.

The cool law startups are not all based in the valley guys and girls, we have https://www.clocktimizer.com/ from the Netherlands, for example.

In a world where AI actually replaces or augments white collar professions, how does liability play out? Will the ability to sue for or defend against malfeasance depend upon an opaque box of algorithmic litigation?

This is a really important question. I wonder how insurance companies would price this as well. If the machines have access to all written legal data then are they allowed to pick and choose which information they include in their legal good or legal service? Will machines ultimately be depended on to develop laws which could then benefit the machines over the humans?

Good luck to you Justin, and to your team. I left law and became an engineer because I didn't see how the legal industry could keep up its bloated and inefficient structure. Shake it up!

Hello - I work at Atrium LTS and we noticed your comment here. If you have any interest in applying, we currently have two engineering roles we're looking to fill: https://www.atriumlts.com/.

First they came for the bodegas, and I did not speak out.

Turre Legal in Finland is doing something similar. Albeit, not sure if their aim is to help other lawyers, just their own clients.

Build an application better than iManage or NetDocuments and you're a shoe in.

It’s interesting because the article correctly points out the existence of barriers to tech adoption in the law — but, as justin and others point out, the coverage of real innovation usually misses the devil in the details.

A number of possible explanations, but the most prevalent I’ve encountered:

• Differences between the variety of ways a given attorney goes about practicing create inherent barriers to a one size fits all B2B software solution. Attorney personas vary wildly based on factors such as practice area, firm size, region, jurisdiction, client demographics, the list goes on and on. This makes it very difficult to build out products that scale rapidly industry-wide. While this sort of end user diversity isn’t inherently unique to the legal industry, it’s probably where its most entrenched, dictated by regulation and long-standing custom, etc etc — all things very startup/entrepreneurial unfriendly.

• Even accounting for legitimate industry-based reasons, there’s a second very real, risk-averse layer: both legal education and practice are very hierarchical, such that aspiring attorneys are being ranked and learning to guard their interests well before they ever begin practice. Once they do, protecting partnerships, colleagues (intra-firm / former pals), profits, rolodexes, business processes, etc is well entrenched in a given attorney’s MO. News writeups usually include a bit about how the Great Recession started to prompt innovation, which it did, but it also created a margins race that further drives attorneys to protect their book of business and profits, further entrenching mindsets which reject disruption, consciously or not.

Regardless of whether you buy these reasons or have your own, the nuance quickly exceeds what captures readers in the mainstream or tech press. Yet the rise of automation in other industries does justify coverage of how those trends would be affecting the law, and the not-uncommon “Let’s kill all the lawyers!” Shakespearean mentality fuels the fire. The result? Every new startup aiming to disrupt the law gets some pretty gloating coverage…and then they need to start acquiring users and scaling. In fairness, that’s what real startups have to do, so not all sanity goes out the window.

Post-2008, the real driver for innovation has been client demand, so Atrium’s approach makes sense there. Also, shoutouts to A2J nonprofits, they’re doing really meaningful work for clients that need it most (albeit with their own set of complexities that aren’t inherently tuned to startup culture)

Obligatory 1: Attorney, but IANYL. Work in the industry, currently building software for firm/legal services ops and biz processes. Previously involved in A2J efforts. Views above are my own, not those of my employer or previous initiatives.

Obligatory 2: If you’re the type with a respect for the industry, but a drive to use tech, product, and entrepreneurial methodologies to break through these types of barriers — we’re hiring! Engineers, product, QA, project, or attorneys that have some meaningful experience in both worlds: https://news.ycombinator.com/item?id=15148885

Startup wants to replace lawyers

Startup wants to replace doctors

Startup wants to replace gym instructors etc etc

I think any kind of job which involves referring to established practices/information can be automated in the near future.

How will you tell the difference?

I would have been much more interested if it were the other way around.

I was wondering the same

Sadly, they can't obey the Threw Laws...

What are Threw Laws?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact