LongCut logo

How Small Law Firms Get Ahead with Claude Cowork

By Masai Brown-Andrews

Summary

Topics Covered

  • AI Renames Messy Legal Files Automatically
  • Natural Language Finds Specific Legal Documents
  • AI Generates Invoices from Billing Data
  • Agents Reason-Act Loop Fixes Errors Dynamically

Full Transcript

Anthropic just released an AI tool that is probably the single most useful tool for legal professionals. That tool is called Clau. And today I'm going to show

called Clau. And today I'm going to show you what it is, what it does, and how you can use it for actual legal work in your small law firm. You can look at templates, generate documents from templates. It can find files for you,

templates. It can find files for you, rename files for you, a bunch of stuff.

In this video, we're going to go over three really practical use cases that you can apply Claude Co-work to in your small law firm today. If you don't know who I am, my name is Masar Bon Andrews and I help small law firms to figure out how to build automations and use AI in

their day-to-day practice. So, let's get right into this demo. So, today we're going to dig into Claude Co-work.

A lot of you guys will be familiar with Claude, the LLM that you can use on the web. If we go to claude.ai, then

web. If we go to claude.ai, then obviously this is a this is a very similar interface to what you get with ChatGpt. And it's just a normal LLM,

ChatGpt. And it's just a normal LLM, right? So you can say, you know, tell me

right? So you can say, you know, tell me about whales, for example, and it'll give you a response in the interface.

But what really makes LLM powerful is when we can give them access to different tools and actually give them access to our our file system so they can make changes to different files and different folders, right? So that's what

Cloud Co-work allows us to do. So in

order to get access to Claude Co-Work, you're going to need a subscription to the Claude AI plan. You're also going to need a Mac as well. And once if you have both of those things, what you can do is you can go to claude.aid

download.

And then this will pull up the download page for downloading Claude to your desktop. So you're going to want to go

desktop. So you're going to want to go ahead and download Cloud Desktop for your Mac. And then once you have that,

your Mac. And then once you have that, you can open up Cloud Desktop. And

you'll see something that looks like this. So there's three different ways

this. So there's three different ways you can use Cloud Desktop. You have the normal way, which is the chat. same

features as a chat that we just used on the web application. But then we also have cloud code which allows us to write software. But the thing that we're going

software. But the thing that we're going to talk about today is called co-work.

And co-work allows us to go into certain folders on our computer and actually manipulate those files on our computer.

So today what I'm going to show you is three specific use cases that are specific to law firms to small law firms just to show you how you can get practical usage out of a tool like claude code for. So, what I have here on

my desktop is this folder here. And

we're going to walk through three different use cases in this folder.

So, in this folder, we have a use case where we're going to retrieve documents.

We're going to have a use case where we're going to generate an invoice. And

we also have a use case where we're going to go ahead and clean up naming conventions. So, we're going to start

conventions. So, we're going to start with the naming conventions use case.

So, if you open up this folder, we have a bunch of really messy document names here, right? We have complaint three,

here, right? We have complaint three, demand new, new, two. I mean, nobody knows what these really mean, but what Claude Co-work is going to be able to do is to go into this folder, look at each of these documents, read the documents

very quickly, and then come up with a name for these files that follows our file naming system for us quickly, right? Obviously, this is useful with 12

right? Obviously, this is useful with 12 or eight or nine documents, whatever we have here. But you can imagine how

have here. But you can imagine how useful this gets when you have hundreds and hundreds of documents from months and months and years of of practicing of practicing law that are not really named

in a clean fashion, right? So, what I'm going to show you how to do is actually get Claw to go in and name these files for us. So, what we're going to do is

for us. So, what we're going to do is again, you're going to come into Claw desktop. You're going to make sure you

desktop. You're going to make sure you are in the co-work section.

And then from here, all we're going to do is first we're going to choose the folder that it has access to. So, you're

going to come down here to this folder icon here.

Choose a folder. And then all you're going to do is go in and select the folder that you want to give it access to. A quick caveat, if you're giving it

to. A quick caveat, if you're giving it access to important folders and files, I would strongly suggest that you have a backup for those files and folders, right? Either in the cloud somewhere or

right? Either in the cloud somewhere or somewhere else on your desktop. This is

just like software development basically where you want to have version what's called version control where you have um you basically save different states of your folder at different points in time.

So, I would advise you to make a a copy or a backup of your folder before you actually go through and do this. But

once you have the folder that you want it to work in, you're going to go ahead and open that folder in cloud code.

Allow it access to that folder. And then

all we're going to do is just in normal natural language, we're going to tell it what we want it to do. So I have a prompt actually set up for it to go through and follow a certain naming convention and rename all these files for us. So again, we have these really

for us. So again, we have these really messy files here. I'm going to grab my naming convention prompt, which you will be able to find in the description of this video. I'm going to go ahead and

this video. I'm going to go ahead and paste that prompt into cloud co-work and I'm going to let it go. And so what Claude coowork is going to do is it's going to do something called it's going to react and react is just short for

reasoning and acting. So it's going to think about what it should do. Then it's

going to actually go and take those actions. So if we go here, we can see

actions. So if we go here, we can see the actions that it's taking in our file system. And then we can in this white

system. And then we can in this white text here, we can see the actual thought process that it's that it's taking to decide what actions it's going to take.

Then over here on the right hand side, we can see the plan that it has. So when

we tell it whatever we want it to do, it's going to go and create a plan for how it's going to take that action. So

we can see that plan and we can see the progress that it's making on that plan as it's going through and doing that.

You can also click into the working folder whenever you want to. So if I go here, this will open up the working folder for me and I can see the the folder that it's taking action in. And

then we have connectors down here on the right which tells us some other connectors that we can connect cloud code into if we want to give it the ability to search the web or to deal with our Google Drive or notion or whatever else you're using. Right. And

the last thing we have as well is the ability to select a model. So right here we're using Opus 4.5 which is the most powerful model. But if you wanted to use

powerful model. But if you wanted to use a smaller model which we could probably use for a task like this for something that's really simple and easy you can select Sonnet or Haiku for that if you want as well. And so we see essentially

the thought process that it went through to decide how it's going to rename the files. So we have a a clean naming

files. So we have a a clean naming structure here. We have the client here.

structure here. We have the client here.

We have the type of document. We have

the law firm as well.

And so now it's telling us that it has finished. If we actually go into that

finished. If we actually go into that folder now, we can see all these documents have been named cleanly.

Right? So it went through, read all the documents and provided a really clean naming structure for those documents. So

that's one use case. Now we're going to go on to the next use case. So not only can it rename documents, but it can actually retrieve certain documents for us as well. So if I come into this folder here, this document retrieval

folder is just to illustrate the the ability of claw to go in and actually read the documents and find the documents that we wanted to find for us.

The the whole idea here is that it's whenever you want to find a file but you can't remember the exact name of the file or the exact content of the file or if you want to just find generally files that are that would help you prove a

certain point then you can use cloud co-work to go in and find those and search your file system and find the relevant files that you need based off of just you using natural language to query cla. So, I'll go in. I'll create a

query cla. So, I'll go in. I'll create a new task.

And then I'm going to choose the correct folder that we wanted to work in. Again,

allow Claude Co to work in that folder.

And then I'm just going to copy and paste some questions basically that would allow us to that would would have Claude co-worker go into my folder and find relevant files for us. So, we're

going to look for I'm going to tell it to find the documents with the no wet floor sign incident report. Right? So, it's going to go in,

report. Right? So, it's going to go in, read the documents, and find that document for us.

And so, we can see it's gone in, we can see the steps that it's taken to go through and read the different files.

And then it's found the incident report for us. And so, let's see if we can get

for us. And so, let's see if we can get it to So, it's giving us the key details from the report. And then what I'll do is ask

the report. And then what I'll do is ask it to, can you give me the text of the report?

in its entirety and it'll go grab the file and give us the full text of that file. Right? So

that's document retrieval. Obviously,

this gets exponentially more useful the larger your file system is, right?

Because going through 10 files is a lot easier than going through 100 files. And

so if you can have AI go through those files and find the necessary files for you, that's going to be extremely extremely useful.

And then the last use case we're going to go through is generating an invoice.

And this is where we have to really understand how AI works and how it's useful. Obviously, formatting is a

useful. Obviously, formatting is a really important thing in law firms, right? And in in the legal industry is

right? And in in the legal industry is is getting documents in certain formats.

And so, in order for us to be able to generate this invoice, we need a few things. We have this billing sheet in

things. We have this billing sheet in the folder. And what we want the the LLM

the folder. And what we want the the LLM to do is go in and tell us about the the matter number or and tell us about basically we want the LLM to go into

this file and to write an invoice for us based on the the billing data that we have in this in this sheet here. So it

should figure out which matter that we need the invoice for and then figure out how much should be built to that invoice. And then the really key part is

invoice. And then the really key part is we need to give it a template for it to we need to give it a template for it to reference.

And so that's what this invoice template is here, right?

Is we have this template that's just telling the LLM where to put each different variable and how to fill this template out for us. So what we'll do now

is we'll open up one more task.

We'll grab the generate invoice folder.

Make sure we're working in there. I'll

tell it to read the billing sheet and tell me what matters are open at the moment. It'll go off and do that for us.

moment. It'll go off and do that for us.

And here's another thing that's really valuable, right, is it can run code on the data that you have. It's much better for the LLM to run code to do the math basically for this for this situation than it would be for it to try to do it,

you know, on its own. So I can use the code and use the data to figure out the amount that's unbuilt. And now I can tell it please

draft an invoice for Clara Reynolds using the invoice template.

Then it'll go take this data. It'll use

the template that we have and it'll actually format that template for us and give that to us. And what you can see over here on the right that's really cool is you can see the skills that it's using. And skills are just a way I mean

using. And skills are just a way I mean it's like a skill for a human, right?

It's like it's a way for us to give our LLMs more capabilities. So this is a PDF skill that allows our LLM to work with PDFs, but you could basically build a skill for the LLM to do anything, right?

And this gets extremely extremely valuable in a legal context because you can build out extremely specific skills for specific legal tasks and you can have your LLM go into your file system

and execute those skills on the data that you have in your file system. And

then the really cool thing, the really powerful thing about agents is that they have all these different tools they can access and they can figure out how to solve problems in a more dynamic way.

Right? So we can see here the PDF, it sees right away that the PDF doesn't have any fillable fields. So, it decides to analyze the structure visually, right? I mean, that's really powerful.

right? I mean, that's really powerful.

It's something that you wouldn't be able to get in a more rigid automation or code script. You have the agent actually

code script. You have the agent actually thinking about what it should do, trying different things to actually figure it out, and then going through and trying to figure it out. And now, this is really cool, right? We can see it reviewing the invoice as well. So, we

have this verify step here on the right in the progress section, and we can see that it it saw that the invoice is mostly correct, but saw an overlap issue. And so, now it's going to go and

issue. And so, now it's going to go and fix that overlap issue for us.

And this is all without any input from me. It's going over and figuring that

me. It's going over and figuring that out. And it's it's also noticing that

out. And it's it's also noticing that the template has a a limit for how many line items it has. And it's going to go and create a better template from or it's going to go and now create a template where it can include more rows

in the actual template. And so we have it going through you. We can see it going through reasoning, taking actions, reasoning, taking actions. This is that

react loop that makes agent so powerful, right? Reason, act, reason, act, reason,

right? Reason, act, reason, act, reason, act. You can see it doing that over and

act. You can see it doing that over and over again until it goes through and actually generates that template for us.

So, if I open this preview, I can see the template that I created, the invoice here and did a pretty good job creating this invoice PDF. And then I should also be

invoice PDF. And then I should also be able to find this invoice in the folder as well, right? So, the the invoice also lives as well in this folder here. So,

those are just a few of basically an endless amount of use cases you can use Cloud Co-work for in a law firm. All of

the templates, all the prompts that you saw in this video will be available for free in the description. So you can go and grab those in the description if you want those. If you are a small law firm

want those. If you are a small law firm and you need help with AI implementation, there is a link in the description you can click on if you want to. This is just a really powerful use

to. This is just a really powerful use case, a really powerful and practical way for you to actually start using AI in your firm. So again, as always, thank you guys a ton for watching. Like the

video if you liked it. Leave a comment.

Let me know what other sort of types of videos you like to see or you want to see. Let me know if you want to see more

see. Let me know if you want to see more stuff around specific AI use cases. If

you want to see more automation stuff, if you want to go into different interface platforms like Air Table for example or Softer or uh any of those platforms, um just let me know what you guys want to see basically. And then

also subscribe so you don't miss any more videos like these. Things are only going to get better and more crazy as we go through the year. So make sure you're subscribed and you're staying posted for that. And that's pretty much it. Thank

that. And that's pretty much it. Thank

you guys a ton for watching and I hope you guys have a fantastic rest of your

Loading...

Loading video analysis...