Getting Started with Claude for Legal

Hello everybody. I'm gonna share my screen in a second, but I'm really, really excited to be here today because honestly, when Claude Flegal came out just two weeks ago, I was like, "Okay, I want to learn more about it." And I noticed that Anthropic actually did a webinar on it. And I attended that webinar. Honestly, it made me so overwhelmed. It wasn't very simple. It was confusing. And I thought, "I still don't really know where to start." So if I'm feeling like that, there must be a lot of people feeling like that. So I started to explore a bit more. And honestly, I didn't find so much information online. Um, I've got to admit, I guess because it's it's completely brand new. I went to the Claude Anthropic website, looked for information, couldn't find it. And that's why I'm here today. Basically, that's that's the reason why I'm here.

I am going to jump into sharing my screen. As I say, this is a, I would say, beginner guide to Claude for legal because it's just come out. It's just been introduced and as I say, I've I've looked and there's not so much information out there. So, let me just share my screen with you. I'm going to jump into it because there's a lot to go through today. It's completely jam-packed. I'm going to be testing out Claude. I'm going to be showing you how I have been playing around with it, been using it. I've basically been testing it out and I want to show you what I've learned. So, we're going through Claude for Legal mainly for in-house legal teams because they're the ones who are likely to use it most. Um, also independent consultants. I think it would be really useful for you guys as well and and law firms.

So, what we're going to be covering today is I'm going to show you exactly how to set up Claude for legal. If you're not sure how to do it, don't worry. I'm going to tell you in a few minutes. So, it's not going to be time-consuming if you already have it set it set up. But, it's really important for those who have not got it set up to know how to do it. I'm going to be also going through the privacy information because I think it's really important for you to know that information being, you know, lawyers or in the legal profession. It's always good to be aware of of privacy, uh, governance, these types of things. People are concerned about confidentiality, data protection, rightly so. And so I'll go I'll be going through that quickly today as well because it's if we're talking about Claude and AI, it's important to run through. I will be teaching you how to do a cold start. This is how Claude words it. It's not my wording. Cold start basically refers to starting off each plug-in from scratch, from from cold essentially. I will also be telling you which documents and the types of context you should give Claude to give it a better output when going through these different plugins. And I'll also be doing some live demonstrations as I mentioned and giving you some prompting tips as well because if you don't prompt well, if you don't prompt properly, then then you won't get the best output, will you?

So, yeah. Okay. So I'm really, really excited to get into this today with you. So if you don't know who I am, my name is Holly Cop. I used to be a lawyer. I worked in the legal profession for around seven years and now I am focusing on AI for legal teams because I think it's fascinating, fun, and I am very, very passionate about helping lawyers save time, be more efficient. I will explain in a little bit more detail exactly what I'm doing because this actual webinar, nobody knows this, but this is actually to celebrate the launch of my new business, which will actually be tomorrow. So, you are special guests to my new business launch training celebration. So, thank you for being here. I have a podcast called More Than a Lawyer and I have hosted various different trainings up until now on AI. I hosted a a global AI skills summit for lawyers and various other trainings including this one and one that I did a few weeks ago and others. And these are the types of legal teams that have actually registered and attended my training.

So as I mentioned tomorrow I am launching a new business. You are apart from my email list who may have seen a document where I've had this logo. You are the first to to have to set eyes on this logo. This is my new business. It's called Friday AI Advisory. I'm really, really excited to share it with you today. As I say, you are special, special guests this this uh launch training. And what I I am providing um I'm already working with teams is AI audits for legal teams where I basically audit um assess investigate your day-to-day workflows, identify where AI can save time and improve efficiency. Perhaps you're doing repetitive tasks or things that could be done with AI. And then I provide AI team training to train your team on how to use AI efficiently, specifically tailored to their workflows and day-to-day activities through trainings and workshops. So this is what I do. And since you are in this training today, if your legal team wants to implement AI safely or strategically, I open a number of spots per month to work with teams because obviously I can't work with everybody and I want to dedicate my time to teams. So if this is something that you might be interested in, your or your team might be interested in, then you can contact me either via email or on LinkedIn.

So, thank you for listening to that. Thank you for letting me introduce my new business. I really appreciate that. So, today I am not going to tell you whether Claude is good or bad for your for your legal work because obviously it comes down to you and your specific workflows. It's not for me to to say. Well, what I will be showing you today is how to use it and how to actually optimize it for the best results based on what I've learned and what I have been able to gather from all the uh experimenting and learnings that I have done. I think probably over time things will change and there'll be more efficient ways to use it and people will find better ways to use it. But for now, as I mentioned at the beginning, I I found it quite difficult to to find other trainings and information even that helps with this.

So, let's let's get into it then. So, how do you actually set up Claude for legal? So, I'm going to just jump into Claude. So, I've come into Claude and this is how you set it up. So, you go to customize. So, let me just go back. So on the sidebar here in co-work so you have to have the desktop clawed. So you go into co-work then you have this section here called personal plugins. You click the plus button, you click create plugin. Then you go to add marketplace. And then in here you need to add a GitHub link which is the link to the specific claude for legal area with all the different plugins. You put it in there and then you click sync and then you basically add the different legal plugins. As you can see here I have added added them to my personal plugins. If you do not have the link, I found it difficult to find online honestly. If you don't have it, don't worry because I have it and I have put it into a a free guide for you which is a guide basically explaining everything that I mentioned today. It took me a while to put together. It's a 27-page guide. It's very comprehensive. I go through exactly how to start Claude for legal. everything I mentioned today, what you need to get ready for the uh the interviews when when Claude interviews you for each specific legal plug-in, examples of how to use it for the different plugins, cross-plugin scenarios. So, let's say you need to use the you have a scenario where you need to apply employment law and then privacy law and then commercial law. So how they kind of can cross over, prompting tips, advanced prompting techniques as well to get better output and common mistakes that I find lawyers make and how to actually fix them. So this is all in a guide which I'm going to show you actually the guide here. I have the how to set it up and the link and then you've got loads of great information there. So, yeah, it's re it's super helpful for you. and it'll just explain everything so you don't need to take notes and things throughout this. It's much much easier.

Okay, so I am I'm going to show you inside the plugins if you've not seen it. Um, I'll do it I'll go through it very, very briefly because you can have a play around with it yourself. So, for example, if I go into the the employment legal plug-in. So, as you can see, you got independent ones. You've got privacy, regulatory, legal, clinic, corporate, legal, AI, governance. If I go into the employment law one, you've got skills, connectors, and agents. So, the skills are different specific uh tasks that it can do. As you can see, you know, it can do investigations. It can do policy drafting, termination review. These are all different skills and you can create your own customizable ones that are specific to your work and your jurisdiction and yeah, your your workflow or what you want it to do. You can create your own ones. As you can see here, there's a customize option. Then you've got connectors and you can connect different like external apps. I'm going to just go to one where there might be a connector. Okay. So, for example, the corporate one, you can connect to IM manage or uh definely or this box. I'm not sure what this is, but you can connect to like legal technology, legal tech apps. And then some of them have agents which is really cool. They have um their own agents which have been set up by Anthropic. And for example, this one's a leave tracker. So it's a weekly agent that monitors open employees, open employee leaves with hard legal deadlines. So I think that that's fascinating. There's so many things to go through. It's it's really, really cool. So you can kind of have a little exploration there.

Okay. So, just going to jump into the next section. Okay. Privacy and confidentiality. So, the privacy plans for the different anthropic/clude plans differ, but also they update and they change and it may be that anthropic doesn't let you know when it has changed. So, you need to be aware of them and always be up to date with the the latest kind of changes. So right now this is what the privacy plan the privacy policy the privacy rules say for anthropic but as I say it could change. So this is the the most recent one. So for example for the free and even the pro plan your conversations are stored by anthropic. They save the conversations. It's not as if they go and check them and check the conversations and kind of eavesdrop on them. It's just that if anything is is flagged as a a risk or a safety risk, they will go in there and double check it so they have access to that. Another thing to mention is that if you the the the model will actually train itself on your conversations. It's not automatically stopped, but if you want to actually stop it, you can go into the privacy settings and change that yourself. So by default it's not it's not set up. And also another thing is that obviously there's there's privacy risks for sensitive sensitive information for legal and any other sensitive sensitive businesses that that use claude the free version or the pro version. Now for the enterprise and team plans which are obviously they're more expensive, more high level, these have a little bit more protection and obviously it's down to your internal uh teams and your internal policies as to which ones you are more comfortable with. But with the enterprise and team version, your organization actually controls that data uh as opposed to anthropic having control of that information and being able to check what you have said in chats and these things. So you are the controller and anthropic acts as the the processor and the conversations are not automatically trained on AI by default. So it's not something that you need to go into the settings and change. it's already in place that they do not tr train the AI on the chats and things that you that you have in there. So you can control how long your data is also kept for. You can put custom retention periods. Uh so let's say you want conversations to automatically be deleted after a month or something, you can put that in place. So there's a little bit more protection with the with their uh enterprise and team plans. So that's just to be aware of. As I say, it does change and they often don't tell you or make they do tell you through the website, but they don't kind of make you aware of it so much. So yeah, it's worth just knowing.

Let's get started then. Let's get into it. Let's get into the the actual nitty-gritty of using it. The best place to start is to run what's called the legal hub builder. So if I just jump into this into my personal plugins where I have downloaded the legal plugins, there's one called legal builder hub here and that is basically the the area I would say it's like the brain of the legal plugins for Claude. So, it's a place where it learns your practice, it learns how you work, it learns your role, and it kind of saves it there in that central location. Each legal plug-in will need its own context and references, and it will interview you for each individual plug-in, but the legal builder hub is like this central spot. And it will actually based on what you say to it or what information you give to it, it will advise you, okay, you need to download the AI governance plug-in because this is most relevant for you or you need to uh download the employment plug-in and start there first. So, it's quite useful actually. So this is where I would start and all you have to do is so let me just go back to right. Yeah. So I did this previously because it can take some time to actually like process and and load. So I'll just show you. I typed in run the legal builder hub start hub code start. That's all I typed. And then it proceeded to interview me. So you can go through the two-minute interview or the 15-minute interview. And it will basically ask you questions about your role, the jurisdiction you work in, the types of clients you have, your risk appetite, all these different things that that it needs context to be able to give you the best possible output. And then at the end it gave me some kind of feedback and it said okay, based on what you've described, I start with the AI governance legal and privacy legal and regulatory legal plugins. So that's where I started. So what I basically said to it was I am in-house legal counsel for Friday AI Advisory. I'm a sole practitioner within within the team and gave you information like that. So that's the example that I'm going to be giving.

What do you need to give to Claude? So it needs information and and context like anybody would. So it you need to approach it like you would a new hire, a new assistant. So, you've hired a new person and you wanted them to understand how you work, how you operate, and probably you'd give them your employee handbook or your policies about certain things within within the business. So, this is what you should give to Claude. And it's probably a good idea to take out any sensitive information, you know, redact client names or confidential identifiers and take them out of the of the policies of the documents before you hand them to to Claude. That's what I would do. But I guess with generic templates and things, the templates for, I don't know, employment documents or letters or an employee handbook, it's probably likely that it doesn't have confidential information in it because it doesn't refer to secretive things or or people's actual sensitive data. Just bear that in mind. And then also Claude needs to understand what is out of scope for it. So you may say to it, you know, I want you to generally draft like policy documents and risk assessments or whatever things like this, but I don't want you to give legal advice or anything like that. So it needs to know what's in and out of kind of the boundaries.

Okay. So I'm going to show you how to do a cold start, which is like starting from scratch, setting up a plug-in, and I'm going to do the privacy one. I I must admit I have not done this because I wanted to do it from scratch to show you today exactly how to run it. It could go terribly wrong. Let's hope it doesn't. But at least I will show you how it actually runs and how it works. So I'm actually interested to see how it goes. And let me know if you have actually tried this out yourself or what you have tried with with Claude. I'm going to be giving you an example in a second of how to do an actual legal task, but this is how to set up one of the plugins from scratch. So, I've not I've downloaded it. So, I've gone into Claude for legal. I've downloaded the plugins that I want and now I'm going to set up the privacy plug-in from scratch. So all you need to say is this this sentence here run the privacy legal cold start right, let's see, let's see what it does. So it's just starting up the session as I say, it takes some time to process and if again, if you want any kind of extra information or you you want to understand about the prompts I use or how how to set it up or extra tips. Also, some people are saying in the comments section that if I can go through it more slowly, you will have access to the recording. So, you can go back and as I say, everything that I mention in here is in writing in the guide.

Okay, so it's beginning to do this for me. Okay, set up the cold start. So, okay, this plugin is for running the privacy program. And it says, do you want a quick two-minute start or a 15-minute start? Who will be using it? Okay, so I'm just going to say I'm going to say a two-minute start. It's for a lawyer and it's for uh I'm going to put in-house. Okay. So, it's just going to ask me questions based on what's relevant and what it needs to know to be able to deliver better outcomes when I ask it to do tasks within this legal plug-in. So, as I as I mentioned, it can take some time to process because I think I guess it's a lot of information. It's a big tool that it's running here. So, we'll just give it some time. Okay. So, let's see what it says. So, it checks the in for integrations practice details. Okay. What does your company do? So it will it will ask more information. Okay. What does your company do? So, I work in-house for a company called Friday AI Advisory and we do AI training and AI audits for in-house legal teams. So, what does your company do? What do you sell? Okay, I'm just going to put that for now. Are you primarily a controller or a processor? I'm going to say controller. Well, actually, I'm going to say both. I'm going to say both. Going to say GDPR uh and UK GDPR. I'm going to put when something needs a more senior sign-off, an escalated DPA position, a regulatory inquiry, who does it go to? I'm going to put me. Okay. And then it will just ask questions. Okay. So, this is what happens. So it just in it simply just interviews you. It just asks you questions and you just answer it very, very simply. Obviously, the more context you give it, the better and it it just needs to know who you are, how you work, your jurisdictions. So it's creating its own tasks. It's running different commands. It's doing it itself. And it what I like about Claude is it keep it keeps you updated as to where it's at and how long it's um taking to do things. And I really, really do like this. I'm going to show you the cold start that for the for the AI governance that I did. Um, it asked me similar questions, but it also asked me which tools I use, which AI tools I use. Let me just go down and show you what it actually produced. Then it came up with really relevant laws and information based on the jurisdictions that I gave it. So I said, you know, I operate with clients in the US, the EU, the UK, for example. And it came up with, you know, US law, SRA compliance tips, which is for the UK, the EU AI act. So it's very in the know based on what I can see, and then it created me an AI policy based on the tools that I told it that I use. uh Notion, Claude, ChatGPT, and it gave me the laws and the relevant regulations around this, which was brilliant. I'm just trying to find the actual document it gave me. Let's see. Can't find the actual document it gave me, but it was I was quite impressed with the output that it gave me.

Okay. So just going back to the privacy one. So I've given it this information and it says done. Your practice profile is written. So it writes its own profile based on your information. So who you are, your role, your company, what you do, what information you might need. And it says to me, okay, the best thing to do now is to work through the privacy legal use case triage skill. Okay, I don't know what this does. I I don't know. If I had more time, then I would say, "Okay, let's go through that and and get guided through each section by by Claude." The five gaps worth filling when you have 15 minutes. So, it it's given me extra information that it wants me to give it to do things better. I'm just wondering if it's actually put anywhere the um the guide that it's created. No. Okay. So it's just giving me next steps basically. So that's how the the cold start kind of works. It's very simple. It just asks for information, then it tells you what to do next.

So the next thing I want you to be aware of is something called context mapping. So obviously when you go through these interview processes in the individual plugins, then it will have context about your business but and how you work and your job your job role as a lawyer, but it's important to for the at least for the the beginning tasks or tasks more generally as it's starting to learn about your practice and how you work. It's important to give it context. And if you do anything, the best thing to do is to give it context around your jurisdiction. So where you operate, where your clients operate, where your team operates, who are the parties involved? So with each with each task that you give it. So let's say you ask it to do a risk assessment for example. Who are the parties that are involved in this risk assessment? Is it that you want to risk assess a new client company for example? So these would be the parties involved. What outcome do you want from the from the AI from Claude? Do you want it just to review? Do you want it to create like an actual document, a policy, a summary? And who is the audience for that document or for that outcome? Who will see the outcome that AI gives it? Will it be you internally or will it be an external party like a client or an a partnering company, something like this? And what have you already decided about that task? So let's say you are asking claw to create a policy, but you already have a policy, you just want it updating, then make sure you give it the context around that. We already have a policy, but we want it written from scratch and fresh uh again, so it has an idea of of where what angle you're coming from. So these are all really con really important when you're giving Claude context.

So, I'm going to actually show you this in practice with the AI governance tool, the AI governance uh plugin. I'm going to start this again from scratch. And I'm going to go into a new task and I'm going to say I'm going to do the AI governance. It's not. I'm just going to put a a governance plugin and I'm going to say if we're if we're reviewing kind of jurisdiction and who and these context details, I'm asking it. So I I want us uh we want to implement Slack, the tool for our company. Uh, I'll put our company Friday AI Advisory and we are based in I put we're based in the UK. We have clients and t team members in the US, uh, EU, UK, and Asia. So global team and global clients. So that's the jurisdiction. And then who, so who does this involve? So I want to create a, let's create like a risk assessment, risk assessment document. Just an overview. I'll say just an overview and points about what I need to investigate further and things to be aware of in the risk assessment. I'll I'll give it context. We don't we don't yet have a an AI policy or a risk policy. Okay. So the outcome is I want, let's say I want a an overview document for who, for who, who is it for? The audience for our internal, oh, our internal processes and what do we already know? I don't think there's any additional information that I need to give here. I think this is probably, or I'll say we haven't ever used Slack before. I'll just say that. I basically send that and then it's going to start to think about it a little bit through the thought process and it should activate the AI governance plugin here. Additional sorts of questions that you can ask for this. So, will it train on our data? Uh, what happens to our data if we stop using the product? So you can add kind of questions into the prompt to understand a little bit more uh what you want to know. And the difference between literally saying to it, oh, can we use this AI tool? Can we use Slack? And something more contextual, the output is completely different. So you'll probably get a generic kind of output if you say to it, oh, can we use this AI tool, then compared to if you give it an actual context around it. So this is a great framework to use and apologies for the background noise there. You can actually refer to this um in the in the guide that I have. So in the guide I've got all this information in there. Um, so a prompt about more context. So, oh, flag the EU AI Act, for example, or data confidentiality risks. If you add it, add a bit more context, you're going to get a better output.

Okay, so it what it's tried to do is, and this is what I like about about Claude is you can see the thought process. So, it says it's tried to actually go into Slack and look at the different features and AI features and and these things. So while it is going through that, let me just go back into a document that I have used. So I asked previously and I gave it quite a generic prompt because basically I worked through the AI governance uh plug-in and it interviewed me and I answered the questions and then I asked it, okay, can I use Slack? Can you create some type of like policy document for Slack? and it started to assess the different regulations like the EU AI Act, the SRA regulations and these things and it started to give me different risk assessments and what it means for me specifically. So this was quite useful and then it actually took a like um a UK like law ruling. This is really, really helpful information for me. It said it said here are some flat slack AI features, data training and output. These are things that I need to be aware of when it comes to privacy and confidentiality. So it gave me a lot of information. Then it said about GDPR and UK GDPR and internal internal use uh risk things to be aware of, things about if training. So, it's really interesting, I would say. Okay, this is still running. Sometimes it can take time and sometimes it can be quicker. So, you really, really don't know, but this is kind of an example of of what it can do. And it gave me kind of a checklist of next steps, which was really, really interesting. So, please let me know where you're at. Um, is this something that you might use? What do you think of the output? I think it's really it's really great, actually. Obviously, I haven't looked at it in great great great detail. So, in terms of its accuracy, I would need to double check that. But any lawyer would need to double check that. Let's say a trainee did this or a junior did this. You would need to double check it anyway. But I think it's a it's a a good start. What I would probably do is just say to it, can you give me the links to the places that you refer to? Like for example, it mentioned this ruling. Can you give me the link to that? um the official government documents so I know what you're referring to so it makes it a bit easier for me to understand where it's actually coming from.

So a few prompting tip tips that I want to go through is that I've got two key tips here actually which are really, really, really helpful and when I am okay, it's just supposedly completed my my risk assessment. So let's see what it's actually come up with. Okay, this is interesting. Let's make it a little bit bigger so you can see. Okay. So, it's done like an executive summary and then it's gone through the different risk levels. This is interesting. So, we've got data protection, GDPR, Slack AI features, high, okay, high risk. Okay. Let's see the key notes here. Slack AI may use message content training, opt-out settings must be reviewed and configured. Okay. That's that's a good reference point. uh company context. So this is my company context, GDPR and privacy. So it's given me some information here. On the face of it, it looks like a good document, but obviously I would want to dig a little bit deeper as to what it says. Okay, invest. It's flagged something here to investigate further. Confirm whether Friday AI Advisory's EU client or team data processing triggers the need for an EU article 27 representative. That's interesting. I think what feedback I would give to it would be to actually give me a link to that as opposed to just telling me about it so I know that it's accurate. Key risk here, what Slack AI does, conversation summaries, that's really, really interesting. Again, I would want a link to like a privacy document to verify that this is accurate, but it's a great start, I would say. Key risk. Slack's terms have historically allowed Slack to use consumer data to improve its machine learning models. Okay, interesting.

So, I'm just going to hop back into the the prompting tips, which is actually quite relevant for this. Whenever you do a task to make the output a little bit better and go that deeper level and and just just push it a little bit further. You could actually say to it, and I'm going to do that now for this task. What am I missing? So it makes the AI think a little bit more like, what have I not asked? Think like a Okay, let's let's just try this one. Okay. So, think like an expert AI governance lawyer. Don't give me legal advice. Just tell me what I am missing here. Anything I am not seeing or considering? Anything I have missed? And then let's see what it says. So, while it's doing that, let me just jump back into the prompting tips. So, this is a good prompt actually. What would outside counsel flag that I haven't mentioned? That's a good one. Okay, so let's see. It's thinking. It's thinking. But let's just go through this document again. It's gone through quite thoroughly actually, different things based on the jurisdiction that I've given it. you know, EU, US, it's got security and access controls, third-party integrations. Yeah, as I say, it seems to be pretty good and it's flagged certain parts, but as I say, I would want the links. I would want the reference points and it doesn't give me them here. So, that would be my feedback for this. And then I would actually update this skill to say, I want you to give me feedback actually.

Okay, let's see what it says. Good question and worth being direct about. Here are some gaps. Okay. The biggest miss, Slack Connect. The document doesn't mention Slack Connect at all. This is a feature that lets external organizations, clients join a shared channel inside your Slack workspace. Okay, I don't know about this uh because I don't have Slack. Okay, that's interesting. The EU AI Act dimension is completely absent. Okay, given the EU AI Act, compliance is almost certainly what you're advising clients on. Not having addressed it in your own Slack assessment is a gap. I'm pretty sure we had some reference to that, but maybe it's not gone any deeper in that. Also, you're not processing, you're not just processing personal data, you're processing commercially sensitive intelligence. Okay. The Slack AI training issue is treated as a current slash future matter. The subprocessor chain breach response is missing entirely. Okay, so it's I like how it's starting to review and critique its own work without me having to do anything. So this is how you get an even better output from it, which is pretty helpful.

Another thing that you can do to get a a better output and explanation would be to say do these things step by step. So stage one, just tell me what the risks are. So as opposed to writing a whole report here which I now have to go through, review and verify that it's correct. Firstly, just give me the points of the risk analysis. Don't draft anything yet. And then after it's created that stage two, okay, now can you draft points one and two so I can check it? And by doing it step by step and prompting it step by step, you're able to actually assess it better for yourself, review it better, and you get better outputs from the from the AI as well because it's not having to think of a million different things like for example with this one. It's having to think of all this content. It's having to review all the content that it has access to online. It's having to create the document, format it, and do everything. Yeah, this is a lot. This is a lot. So, if you ask it to do it step by step, then it's much easier for you, and it's also better for the AI because it just has to think about one thing at a time. This has been me kind of playing around with it and showing you what it kind of can do, what are the limitations, what it what it can't do. It's also worth mentioning that if you want to activate a skill within a plug-in, you just click the forward slash button and then it will come up with um the specific skill. So for example, if I hover over this one, this policy monitor. So this is the under the AI governance legal skill and that's been activated when I ask for this to do this task for example. So that's worth just worth mentioning.

That's all I'm going to to go through today. It's a lot to go through. Let me just skip to this. There's so much to cover. So I've gone through the setup, how to actually set up Claude for legal, which is important. You can't do anything without doing that. I've gone through how to run through the the cold start properly. Gone through the privacy information and how to prompt and also have gone through a few different tasks and how to do it in practice. There's there's so much to go through. It's just it's massive. It's huge. Tomorrow I'm super excited that I'm going to be launching my new Friday AI business. As I mentioned earlier, this will be including AI audits and AI team training, including Claude for legal um if this is something that your team uses. So, I just want to say thank you so much for coming today. I really, really appreciate your support. There's been so many people register and if you have any questions, then please let me know. So, thank you so much for coming today. It has been an absolute pleasure and so great to see you all.