SimpleAI // AI-Powered Contracts in Action

So, uh, welcome everybody. Thank you for joining us today. We're going to be talking about AI powered contracts in action. We're going to be showing you how to, um, how to utilize AI for the review of contracts. We're going to be doing that by showing you how simple AI works, uh, which is the word adding that you can see here uh, in this screenshot. Um, so there's a lot of talk of AI. Um obviously lots of people are using it particularly in the legal space. Um and there's also a lot of talk around mistrust of AI and how people don't fully trust the outputs. Um, I've been doing a lot of thinking about this, of course, because this is what I do dayto-day. And that's why I'm really just so glad that we have now acquired Law Insider because Law Insider is just a wealth of data. And now by layering simple AI and the power of this AI tool over the Law Insider database, we're able to offer things outputs that are much more trustworthy and that are grounded in precedent and contracts that were actually signed in the past. So we're not feeding um our model anything that is aspirational like templates because templates when they're created that that's exactly what they are. They're aspirational. You hope that uh the other side is going to accept it. Now what you negotiate and get to sign at the end of it is a very different picture in many cases. And so to have access to that data of executed contracts that really demonstrate what's market, what's standard is um is something that's going to is is going to fundamentally change how people use AI in a way that's going to help them trust it much more. So most AI tools generate plausible text. It sounds really convincing when you read some text that comes from chat GPT. We're only human so we will have a bias. when we read something that's well written, it often feels like it's more authoritative and therefore reliable and actually that's not the case. And lots of lawyers have sort of fallen into the trap of using AI outputs that weren't entirely accurate um because they sounded accurate because when something says something with confidence, it comes across as more authoritative and trustworthy when in fact it may not entirely be that. Um, Simple AI is powered by the Law Insiders by Law Insiders database. Um, for anyone who doesn't know, Law Insider is a 10-year-old business. Um, it was founded by Preston Clark, who is also the founder of Simple Doss. It was co-founded with a different team. And, um, it was co-founded by Preston straight out of uh, law school when he realized that there was a gap when it came to accessible legal uh, information particularly in the realm of contracts. So he built out this database. We have 11,000 customers plus that are using this database and that subscribe to it to get access to that raw data. But of course with the rise of AI, there are better ways to utilize that data now. And so that's where the marriage of law insider and simple AI made perfect sense. Hence the acquisition. Um, of course without real world data, AI cannot know what's market standard. So it's very difficult for AI to just look at all the things on the internet and uh tell us what's market standard. Whereas if you give it a very um specific pool of data then the AI suggestion is much more grounded in real data rather than LLM guesses which you might see um when you ask a general purpose LLM about what's market standard. You may also see inconsistencies when uh you're using general purpose LLMs on that basis. And of course, legal teams lose trust and waste time checking outputs. It makes it inefficient. It makes it hard to adopt. Um, it makes it hard to to become as transformative as it can be. So, the result of what we've done is the creation of faster, more confident contracting because when you see something that's been generated by AI but then has been benchmarked against real data, then um that's what makes AI so powerful. So, precedent makes AI powerful and that's exactly what we aim to do. For anyone who doesn't know, I'm also um the co-founder of 1NDA, which is a market standard non-disclosure agreement that I created um a few years ago now uh when I was wearing a very different hat. But that came out of this pain around lots of lawyers negotiating documents that were very high volume but very low kind of reward if you like um without any standardization uh and and knowing full well that standardization works in other areas. I took the initiative to try and standardize one of the most ubiquitous forms uh globally and now 1NDA is um a standard non-disclosure agreement that has been adopted by thousands thousands of organizations across the world but of course um this this initiative needs to transcend the template that we had created and so again when law insider acquired 1 NDA it also acquired this this body of of uh templates that were crowdsourced and that were created by the legal community and so we have the opportunity to embed those market standards into our AI as well. The reason precedent matters so much in the age of AI both in terms of executed contracts but also templates that um that the community has said are good enough to become non-negotiable standards are the if you can embed that intelligence and that data insight into your AI it just builds confidence it drives alignment and of course very importantly it gives AI context. AI works really well when you give it enough context. Not too much, but you give it enough. And so it can start to understand patterns in a way that humans just can't. uh as well the four pillars of precedent-based AI or the way that we're approaching precedent-based AI is that we we really believe that if you can ground AI in in in in these pillars of data the output that you'll get will be much more reliable much more consistent much more aligned with the outcomes that you want and so this is the whole premise upon which simple AI is built and this is a very different premise to some other tool s of course that are very much using the power of LLMs and and the speed that it gives them. We fully understand that without accuracy, speed is kind of useless because if it's not doing it right and you can't trust it, you're just going to spend a load of time checking the outputs and building quite an unhealthy relationship with it. So, we're really focused on ensuring that what we're building is fully trustworthy and something that will give you real speed because speed only comes with accuracy. So the four pillars that underpin our solution are firstly internal playbooks. This is really important. It is precedent and it is information that you provide the AI with that will align the AI with your preferred positions. And this probably sits at the top of the food chain because if you're able to tell the AI exactly how to behave, that's that's when you eliminate any hallucination or any unreliability or any random outputs that the AI can give. Um, and so if you can tell the AI how you want it to review an agreement for you in the form of a playbook, which is just a set of rules against which you review an agreement, then the AI will be able to execute. And the better you can structure your prompts and the better you can that the clearer you can be with the AI in terms of how you want it to uh to to perform, the better your outputs. Then we have internal contracts and templates. So we're offering a capability where you can layer simple AI over our repository where you can store your agreements so that you can glean data insights from your precedents. So this is something that um we've been working on quite intensely over the last period. And so we want to be able to give people the capability to see what they've signed before. So yes, you've got your playbook and you're telling it what your policy is, but that's a slightly different picture to what you've actually agreed to in the past because of course there are rules, but then there are deviations and exceptions to those rules that aren't always codified. So to be able to access that body of data to inform how you review an agreement or what deviations you can actually accept or what your risk appetite actually is is an important uh element here. Then we have market standards such as one NDA, one DPA, one SAS. These are templates that have been adopted in in the thousands. We estimate that one NDA gets signed around 10 million times a year now. So this is really penetrating the market and I and I do speak to a lot of people that use it and don't even know who created it. So it's gone beyond the small community and ecosystem that were early adopters and it's gained market share in a significant way. So starting from what good looks like and what the community the legal community says that what good looks like is is fundamental. Sorry that's the fourth pillar community standards and of course you have market standards. So this is where law insider um data comes through having the capability to amalgamate data insights across thousands of agreements and agreement types and to be able to um surface what's market standard for that space for that type of agreement for that value of an agreement is really important and we're not where we want to be yet. This is something that we're actively working on to be able to give you as much insight as possible to firstly help you draft and review contracts more efficiently, but also to give you negotiation leverage and tooling so that you can have these conversations in a much more constructive way. We all hate it when the other side goes this is market and that's their only comeback when you push back on a certain position. So to be able to back that with data I think is something that's fundamentally going to change the game. Um okay. Uh so I'm going to show you the tool today. We're then going to go through some options around um what you might want to choose if you think that this solution is going to help you. So I'm going to go straight into the demo. Now before I do that, does anyone have any questions for me? Electra, there's one question in the chat. Um, and we've sort of covered this, but how can we warrant that internal contracts or templates are not shared with other customers? >> They're not shared with other customers. So, um, everyone lives in their own instance of their uh, simple AI solution. So, anything that you give the uh the the the solution is just stored for that session only. If you then want to layer it over your AI repository, then of course that's all confidential and we maintain the highest levels of security. Uh nothing is is used to train the AI as such and nothing is used beyond that session um where you've where you've um opened a document. So if you're leveraging your data then that's great and that and that will stay within your environment only. but we don't use your data to train the AI or to um share insights from your specific contracts with other people. Um so that's not um something that you should be concerned about. Um any other questions? Okay. Thank you for for doing that. Um I'm sorry. Just one second. And I just need to share the right window here. Um, it worked a second ago, but now it's not, of course. Okay, I'm just going to go ahead and share my whole screen. [snorts] Right, it's my disgusting calendar. Okay, so this is a horrendous MSA that I created with the use of AI. it's deliberately imbalanced and one-sided so that I can show you as much um as much of the tool as possible when it comes to reviewing contracts that aren't great. Um as I said, this is this is simple AI and it lives in Microsoft Word. It it exists to predominantly review uh Microsoft Word documents. Um and so there are two sides to uh simple AI. The first is AI assistant and the second is document review. So the AI assistant is perfect for very pointed questions, strategic questions around how the contract performs and also very surgical changes uh in in the form of red lines or drafts or comparison documents. So I'm just going to show you how this works. So um the first thing that I'm going to do is I'm going to ask the tool to summarize uh commercial in this agreement. Now, often if you're inhouse particularly, you might get a contract from your business stakeholder. You don't really know what it's about. So, you can just ask it to summarize what this is. Now, the second stage to that is to get it to give you an overview of key commercial risks in the agreement. Now this opinion is based on data right that we have provided to our model in order for it to give you information that is more aligned with what's market standard rather than just a very generic LLM. So it's going through the agreement provision by provision and it's giving you an opinion and a reasoning so it can act as a strategic partner. You can also ask it really simple things like what's the applicable law? Some uh teams um are only looking for very specific things and if they're okay then that's fine and they can move on and so you can use it as a buddy in that sense. You might also ask it to highlight any uncapped liabilities if that's a concern that you have or unusual indemnities. Again, this question is sort of opinionated or it's asking for an opinion opinionated response in terms of what's unusual. So it will tell you that it's giving that it has an a general liability cap. Um and so no there is no uncapped liability but it also has some unusual indemnities here. And again it's giving you a summary at the end. And of course this is an informed opinion rather than just a general um un non-ontextual opinion. Um so you this is one way that you can use this tool. The other is to ask it to make, as I said earlier, very surgical changes. So, for example, you might say, revise the indemnity clause to make it mutual. Now, when you're looking at the output, which you're going to see in a second, it's um it it's going to be using language to revise the agreement that is that it thinks is fair and balanced, but also just from an efficiency perspective, it's going to be using um language that aligns with the terminology of the agreement. So if you were in a in a world where you're not using AI, this is just I'm just talking about efficiency now rather than uh access to data that um that informs the output just from an efficiency aspect. The fact that it's using capitalized terms for for for definitions um is something that's going to save you time because uh if you're using precedent or if you're using a template that you've got or an agreement that you negotiated the other day that you liked that that provision in um you're going to have to retrofit that language in order to suit the language in this agreement. often you have to revise the tone of voice so you know that whatever the AI is doing here that the AI is giving you is tailored to your use case here. So all you need to do is just apply that and it will turn a one-sided indemnity into a mutual indemnity. Um you might go cap the liability to fees paid. Again, I I've talked about this a lot. Um, but you know, this is the future of contract lawyers. It's not drafting from scratch as much as prompting the AI. And of course, the prompts that I'm using here are quite simple um so that they're easily understandable. But if you wanted to go into more nuance and complexity, then of course you could. So um yeah, the future of uh of contracting is is prompting. Uh this is just taking a second to cap the liability to fees paid. In the meantime, I'm seeing some questions pop up. So um is there anything that I need to address story whilst we wait for the >> Sure. Can does the tool cover formatting changes? I know formatting is a a fun one. >> It's a fun one. It doesn't it doesn't do it as well as we'd like, but we're spending a lot of time and resource on trying to solve that problem because it's a huge bug bear of every single lawyer. I know firsthand that you might revise an agreement, review an agreement, and it will take you longer to reformat it than it did to do the legal substance. So, I totally understand the pain, which is why we're really trying very hard to fix that problem. But a word document might look quite simple. If you kind of look behind the hood at the code that um that sort of renders a word document, it is extremely complicated and it's quite a hard problem to solve that the whole I know the whole industry is is struggling with at the moment. So yes, we are trying our best but at the moment it's not amazing at formatting. >> Great. And then tied into um repository and also you know internal precedent um does the tool only work on one document at a time or can you ask it to check where we have agreed to termination for convenience before? >> So yes so that's that's something that you can do with the AI repository. So the AI repository, I'm not demoing that today, but if you do want a demo, Dory, I don't know if you could drop a link so that we can offer people demos of the >> Yes, I will add that in. >> But you can query the AI repository. So you can say, give me um a list of all the contracts where we've agreed this or you can ask it, have we agreed this before? Yes, here, here, here. So yes, you can chat to the body of contracts that you have. Um, okay. So, this has done it, which is great. Thank you. Uh, okay. Now, um, the other thing that you can do is that you can draft language from scratch. So, if you're like, I really wish I had a non-solicit here, draft a nonsolicitation. You can be more specific with your ask, but just for the purposes of this demo, this is where you can really see evidence of the use of law insider uh and the database that it is. So what's happening here is that the AI is generating language for you. Again, the terminology is consistent with the terminology in the agreement, so you don't have to worry about retrofitting it. But what it's also doing is it's taking this AI generated language and it's benchmarking it against the law insider database, the law insider library. And what it's going to do in a second, and apologies, my internet is quite slow today. Um, what it's going to do in a second is it's going to give you a score. And that is what we call the law insider index. And this score effectively tells you how similar this language is to other language that lives in that database in similar types of contracts is um and whether you can trust that this is standard or not. So the lower the score, the less frequently the substance of this clause appears in agreements of similar type within the database. So if you've asked it for something obscure, you can expect a lower score. Um, and so you can you can still accept the language and introduce it into your agreement, but just know that you might cause some negotiation friction on the basis that it's not standard. The higher the the the score, the more confidence you can have that that language is has been used before. Not in its exact form. It's not checking word for word. And I'm sorry, I don't know why this is taking so long. It doesn't usually. Um, but the higher the score, the more substantially similar the language is to what's in the database. And so firstly, that's hallucination proof because you know that it's it's pegged against other contracts with similar language, but also you know that you're probably not going to have any problems getting it through the other side because it is market standard. Um okay let me come back to this uh because it is taking a little bit longer than I would like. Uh okay the other thing that you can do is that you can compare documents. So, if you had a term sheet um and that term sheet had a load of issues that you discussed or points that you'd considered, you can say compare the attached term sheet to the document and outline a list of differences. So, that's happened within seconds, which is great. Um, so here it's giving you the term sheet versus the MSA termination rights, fees and payment. So it's being quite very detailed. Um, and you can see what the differences are between uh the two documents. And if you wanted to redline it, you can also just say redline to align the MSA with um with the term sheet. Um, and now it's giving you a summary table. So yes, this is not necessarily pegged to the law insider database or the uh or the market insights that we can provide, but it does give you a sense of how many effic efficiency gains you can get with this feature. Okay, I'm just going to hop over to uh playbooks quickly. Um so when you're receiving a contract for review um often it might be quite a different contract to any contract that you've reviewed before. So in that instance you have the capability to use general review. Now, general review leverages again all the data that we fed it and the LLM that underpins it, open AAI, in order to create a playbook for you on the fly. As I said, a playbook is a set of rules that tell the AI what to look at in an agreement and to then offer you a remedy if there's an inconsistency between the substance of the agreement and your rule. So what general review does is it looks at every provision in the agreement. It extracts the principle and then it searches its data figures out what's market and creates a rule that is market against which it will then compare that provision again. So do you see that what I mean? It does this this double this double action. So it takes it looks at the provision, takes out the principle, checks how that principle fares within the market, creates a rule, and then it will tell you whether the substance of the agreement aligns with what's with what's market and balanced or not. So here, what it's asking you for is a bit of context. So here you're telling it that you're the customer. If you told it that you were the provider, it would take a different stance. Um, you tell it what your concerns are. You usually go with highest quality unless you're in a hurry like me in which case you go fast. So if you click this then what happens is it first begins by giving you a strategy summary. So here it's telling you how it's going to approach the uh the creation of the playbook. And the reason it's doing this is to give you an opportunity to repprioritize some of the elements that it's raised. So if you think that, you know, actually I do have some more concerns around this area. Um this is a sensitive MSA where we're actually sharing a lot of personal data. So you want to tell it to um to go harder on personal data, you can totally do that. Otherwise, if you're happy with what it's raised as an overall strategy, you go start review. And so now what the tool is doing is it's creating those rules and it's it's benchmarking the contract against the law insider index. It's benchmarking the principles within each provision against the index and it's looking at how these fare against market and then it's creating rules that tell you what's market and then it will provide you with a suggested red line or um a red flag so that you can go ahead and either align it with what's market or um or just ignore the suggestion if you like. Um, so here it's telling us that termination for convenience is customer only and that's actually fine in its opinion. Um, non-solicitation clause isn't there. You told it, I told it that I don't want one. So, it's the first thing that it's phrasing because I told it what to do. Um, but here it's telling me um that there is insufficient customer protection when it comes to the indemnity um because it's the market standard is that it should be mutual and balanced. Now you might go actually no I don't want it to be mutual and balanced in which case you ignore it. Um but if you're happy to take its suggestion which is quite robust you would hit apply and as you can see it will amend the agreement uh in a second. Um and we are one of the only tools that actually redines red lines. So um again we spent a lot of time trying to get that to to work. So this is your AI generated playbook as benchmarked against what's market. Um, but if you wanted to take a more pointed approach, then of course you could. So, this is our playbook feature here. You've got the capability to either choose from one of the 21 playbooks that we have. There are 30 more coming. Uh, we've had lots of requests from people that use these all the time for more types of um of playbooks. So, these are coming soon. There were only 12 of these last month, so we're building very quickly and we're prioritizing this. It's something that people use quite a lot. Uh and each of these playbooks is is created by using the data from law insider. And in some instances, we're also replicating the principles that underpin some of our standards. So like one NDA, um one DPA, one SAS, the the rules that are here are [clears throat] are are sort of on par with the principles that underpin those standards. So you can use if you're an adopter of one NDA or one SAS one DPA um you can rest assured that if you're using one of our playbooks you are aligning any third party agreement against those standards and soon we'll be releasing um a revised view of this because the more playbooks we create the more usability we need to also create when it comes to choosing the right playbook for your needs. So, we are revising this and we're also going to make very clear which uh of our playbooks are based on a community standard versus the law insider uh market insights. Um so, at the top here you can see your custom playbooks. These are playbooks that you can create and I'm going to show you some tricks in a minute on how to do that really quickly. But here you've got your um you've got your playbooks. I'm going to use the buy side MSA because that's my role today. And here I'm going to again run the playbook and it will do what it did previously um but based on the rules that we've already introduced. Ah here's the law insider index. I don't know why it didn't do it earlier but this is what it looks like. So when the AI generates language so when you're using a playbook and one of the rules is that the uh contract should include language around audit rights but there are no audit rights in the agreement. the AI will pick it up and it will also generate language that you can introduce into the agreement in order to uh include that provision. And here's how um we demonstrate the the the similarity against um other other contracts in our database. So here it's telling us that this generated language is 80% similar to clauses in Law Insiders database. Higher percentages indicate closer matches. So you can sort of glean from that that um there's a lot of uh a lot a lot a lot of agreements that contain language that is substantially similar to um the language here. Um, so I'm just going to kind of lift the hood here, um, of a playbook so that I can show you how rules are structured. And I'm going to go into the anatomy of a rule just because I think it's interesting to understand how we speak to the AI. Um, and also to say that you have the capability to make changes to the rules. If there's a a standard playbook that we have that's kind of ready to go and easy for you to pick up, but if there are certain rules that don't necessarily align with your preferred position, it's very easy for you to change what the rule says in order to align it. And you can duplicate all of our playbooks. So, if you were to duplicate this playbook and call it MSA ABC limited, um this would go and live in your custom playbooks and would then fully align with how you prefer to review your agreements. Um, I'm going to pause there. I'm seeing a few questions come through. Dory, is there anything that I I need to address? >> Yes. Just firstly, since I didn't know this one off the top of my head, but when we're talking about um the AI repository and the tool being trained on user specific data as well as generic data, sometimes the data can actually contradict market standards. So, can the user switch off general data so the outcomes are grounded really just on the user's data? >> Yes. So, so this these features are still being built out. This is a quite a new product and so all of these nuances are still being being built out. So, yes, the idea is that you'll be able to either glean insights from the body of data that you hold within your contracts or you just want to go by your policy. So, you don't want to um contaminate your policy, which is your playbook, and your preferred positions with stuff you've done in the past, which you may or may not be proud of. So, yes. >> Great. And then, um, when you do a review and you ask the AI to perhaps summarize and, um, draft an email, for example, of changes, is that also going to include open comments that are still within the document? Um, you can actually ask it to review comments in the document. So, if you wanted to ask the tool to give you an overview of all the comments that are in the document, yes, it can do that. Um, you can also ask it to give you a summary of all the changes that you've made and a rationale as to why you made them. Uh, and it will give you a table or whatever format you ask it for. So, I'm going to I was going to show that at the end after I'd reviewed, >> sorry, jumping ahead there. No, [laughter] no. Good, good, good to know that what I'm showing is actually of interest. >> Yes. >> Cool. Okay. So, I'm going to go into the anatomy of a rule. So, I don't love this this uh this example. Let me show you something that is really straightforward. Governing law. So, here this is a generic playbook and it's meant for use by broad audience. So, of course, we can't determine the governing law, which is why we have marked it as a suggestion. And what we're saying here is check the governing law and that it aligns with the customer's home forum. This is just an assumption that that's what most people would want in the first instance as their preferred position. But of course, you would have a preferred um governing law. So, what we're going to do here is we're going to change this to personalize it. So, governing law is the rule name. It's important that you keep this nice and short and sweet. So when you're reviewing a contract, you're not overwhelming the reviewer with really long rule names. You want people to be able to go through the playbook as it appears on the right hand side of your page. Uh and to be able to quickly kind of process what the rule is and and move on to the next bit. So it's important to be user centric when you're drafting your review rule name. Here we have the instruction to the AI. Now we're using our standard format which is identify, check, act. And we like to always start our prompts with the word insure. There are other ways to prompt as well, but we found that this is the most reliable way to prompt the AI to get the output that you want. I have a prompt engineering master class that we recorded. Um Dory, maybe if you can drop the link at some point during this webinar if anyone's interested in going deeper into how we how you can prompt. Um but here, this is the prompt that does the thing that tells the AI what action to take. It's telling the AI to ensure that the governing law and jurisdiction are those of customers home forum and if not suggest revising. This is marked as a suggestion. A suggestion is just it's a red flag. The AI goes excuse me look at this. It doesn't actually tell you what to do to remediate anything. But if you did want to remediate it, you choose red line. And here you can say England and Wales. if not red line accordingly. So you've just personalized this so that it align with your preferred position, but you might also have a fullback where you're actually quite happy with New York. So you give it the the negotiation um logic and uh reality if you like. Sometimes playbooks are too static or um they're really difficult to follow because the fallbacks are kind of there's lots of interdependencies. Here the way we've structured it is so that you have to tell the AI what your positions are and they can't have any interdependencies and they they they're standalone. So here you're like England and Wales is fine, New York's fine. Also France is fine. Um and then when you review the agreement, it will tell you um whether that's actually fine or not. Um again, this is a red line. So this is the action that you're asking the AI to take. And this is a guidance note. This appears only to the reviewer and your internal team. So you might say, um we prefer to use New York when we are contracting with our US entity. it's up to them to kind of make a decision there. Um, and then you might also want to enable comments. So here you might say our preferred uh jurisdiction is England and Wales. And so that's where we're based. We do not yet have dynamic comments, but they're coming. So in the future, you'll be able to change the comment based on the fallback that you've chosen. Um, so in this instance, we suggest that you create a comment based on your preferred position on the basis that that's the position that you're going to take most often or that you're going to push back on most often. So if you save your changes here, you can then do that for every single rule. But as you've seen, it is quite easy to to do it. So um it's easy to speak to the AI. We've ensured that the way that we structure the prompts is understandable to both humans and AI. The way we train our customers is on the basis that they can be autonomous after we've deployed their playbook for them and they can go ahead and update their rules because of course playbooks and contracts are not static. Um and so that's why we we just to talk a little bit about the process. If you were to um buy our solution and you wanted to create bespoke playbooks, we will either help you tailor a standard playbook or create a playbook for you from scratch based on an agreement that you have or um [clears throat] even for the use of your review of your own templates with markups. So um you get an implementation team with uh the simple AI enterprise product and so you will get full support to create your playbooks but also to iterate on them if you need to and get as many training sessions as you need to help you roll it out and get it adopted within your organization. Um okay I just want to go to governing law for a sec to show you what that fallback structure looks like. Where is it? Here we go. So here it's giving you the pri it's telling you that the primary position is not met. The fallback New York is met um and your second fallback is not met. So if you click on the language here it takes you to the relevant language in the agreement which you can then read. Um you might say do you know what my second my second preferred my second best position is met. So I'm not going to make any changes. Um here's the guidance note for the reviewer. Here's the suggested red line if you did want to uh stick to England and Wales. And this is the comment that will appear in the margins if you did did give effect to it. So I am going to insist on England and Wales here even though New York is fine because I want to be a pedant. Um and here I see the uh I see the comment that appears in the margins. So this is how you can encode your negotiation logic into this agreement. And of course take comfort in the fact that where language is missing even if you haven't told the AI exactly what language to use in the absence of certain language that we are leveraging a very reliable body of data to create language for you that you can use in your agreements. Okay, any questions? We actually just got one that I didn't have a chance to take a look at yet, but I do think that we have answered just about every other question. If anybody else has any other questions, please feel free to put them into the chat or the Q&A. We'll make sure we get to them. >> Dory's keyboard's on fire. Um, so question on playbooks. Yes, they take time. It's an ownorous process, particularly if you are creating a playbook for the review of your own agreements. I'm going to go into a bit of a nerdy spiel for a second just to explain a nuance. Um, there are two types of playbooks that you can create. The first playbook is a principalbased playbook. Principal-based playbooks are playbooks that are less concerned with the language that's used to express provisions. So, you might have your own terms that you send over to the other side. Say the other side is mega pedantic and is making small changes to commas and formatting and structure and of language but actually the substance remains the same. Your playbook will pick up nothing because you've structured it in the way that it's so you've structured a principle-based playbook. So it just looks at the substance and if it aligns with your principles then it doesn't kind of go into revision of language to align it with your preferred expression of language. The other is a prescriptive playbook. So this is usually um created for the review of your own templates and markups to your own templates. It's usually used by organizations that are regulated where your terms may have gone through lots of rounds of governance or even reviewed by a regulator and approved. And so you are very precious about the language that's used to express certain provisions. And you're also very precious about changes that the other side makes to your contracts. And although in principle certain changes might be okay, you want to stay in control of the drafting of those principles. So they are prescriptive playbooks. They are the most ownorous to create. They take a couple of weeks and you need to work with us to get that playbook up and running. When we get it up and running, it's an amazingly efficient tool because you can review any markup to your agreement and the reviewer knows exactly what's acceptable, what's not, what fallback positions they can accept, what language is acceptable. So something that's probably quite um method very very methodological at the moment and very processorientated, you can encode and the AI will really help you streamline those types of agreements. But if you have um a principalbased playbook that you want to create and you've got a great template contract that you like, say you love your NDA um and you want any NDA that you review to align with the principles in that NDA, then what you can do is use the AI to create a playbook for you in seconds. So what you do is you'd open your say your your perfect MSA here. You'd go to new playbook from current document. You tell it what the name is. You tell it that you're the You wouldn't say anything here. It's optional because it's an NDA and it's mutual. Um any additional context, you'd choose highest quality uh because you'd get better out output. And now it's going to go again and ask you give you its playbook strategy summary. Um, and then once you're happy with that summary, if you think that it's understood the gist, which it usually has, then you could just ask it to create your playbook for you. And you hit a button and it will create all of those rules. And those rules will reflect the content and substance of your agreement. So, this is a really quick way of creating third party playbooks. Um, and yeah, it's one of our most used features uh when people are building out their playbook library. Any questions on that before I pop back to the slides? Okay, I think I've demoed absolutely everything which is amazing. Um because I never get the chance to do that. Uh, okay. So, apologies. Okay. Questions to fill this gap while I awkwardly look for stuff for my slides. Um I I think we've covered it you know the the different things that we can do with the platform besides just contract reviews. So you know summarizing comparing documents um any others that you can think of Electra that perhaps we haven't covered? >> Um I don't think so. I think that yes the use case of I don't know why this happened. the use case of um of of summarizing I didn't show that sorry I just want to make sure but you can ask the AI afterwards if you go back into the um chat feature you can ask it to summarize all your red lines for you and also create emails draft an email to my internal stakeholders explaining what changes we made and why or draft an email to my counterparty explaining the key um changes that we've made and why we've made them. So, um you can absolutely test the tool if you're interested. Dory, could we drop a link for anyone who wants to request a free trial um in the in the >> Yep. We've actually got a poll here. So, if you'd be interested, >> you can go ahead and just make a selection here and we can make sure that we follow up as desired. [snorts] >> Thank you. Uh, quick question, Elector, that did come up. I'm not sure if I missed it, but can the AI generate push back comments? >> Yes, it can. So, you can ask the AI to do that. So, the AI, you can just chat to it in a very natural way. And again, it will glean insights that we've given it in order to give you um that output. So, yeah, you can ask it to to give you push back and and help you with your negotiation strategy. Great. Okay. Um, so Ben, yes. Could you book a demo and we can show you um [clears throat] all the features of Simple Docs? Um, yeah, we do have more products than this, but this is the one that's really in high demand at the moment. Um, Amy, if you would like to get some personalized services from our team to get stuff set up, uh, you can totally do that with the enterprise solution. So, if you're on our enterprise package, you get professional support alongside it. Okay. So, um, we have two types of plans that you can look at if you're interested in what I showed you today. So, we have an individual plan that can be sold directly through Law Insider. You can just go in with your credit card and buy it today and you will get access to um the law to to the law insider version of um of the um of the of the Simple AI solution. Uh, and with that, you get unlimited contract reviews and red lines. Um, it will support multi- language drafting and review. It's integrated with the clause database. So if you were to draft language, you would still get the score and you can get access to the database to look deeper. And of course it's sock 2, type 2 and GDPR compliant. Um if you did want to leverage playbooks, uh then you would have to go onto our enterprise solution which is available through simple docs. Uh and to to buy that solution, you need to book a demo with us um and ask or or start with a free trial. And here you can get you get everything that's included in the individual plan um of Law Insider, but you also get your playbook builders, unlimited custom playbooks. You can build as many playbooks as you like. There's no limit on that. I know that some other tools do have some limits on that. There's no limits on how many playbooks you can create. You have the ability to share playbooks across team members. So this is a really important feature because if you're part of a team and you want to scale the knowledge that you have encoded into a playbook, you can do that through um through this um and you get dedicated account manager. So as I mentioned earlier, you get support, you get professional services basically baked in um to the to the product. So if you go into the enterprise solution, we really believe that to give you quick access, quick value, um you do need professional support from us because there's a learning curve. And so in the first instance, we'll get your playbooks up and running. We'll do a discovery piece with you. We'll understand um what your needs are, what your preferred positions are. We'll help you extract those and then we will embed them into the AI tool. We'll do a testing phase with you and then once you're all happy, we'll help you deploy it within your organization. and I lead that legal engineering team and I'm having a lot of fun. So yes, I would love to work with you if you did want to go onto that solution. Um, as I mentioned earlier, we are already working with thousands of teams, over 12,000 teams that have adopted our solutions. These are some of the brands that we work with and who are um are reporting some great gains. Uh we work with regulated organizations, large organizations, um public organizations and so we are not afraid of your due diligence process. We are data privacy and security uh kosher. Um we do not train our models on any of your data or input. So as I said previously, every session that you um undertake within the word addin remains in that session. We don't then store the data or train our models. That's all yours. Um so you don't have to worry about that. uh there is no data retention beyond what's necessary to process your request. If you did want to find out more about our security, then feel free to go to the link that's here. Um Dory, if you could drop that link in the in the chat, then anyone who's interested can go there or perhaps it's better if we email everyone afterwards with some more information around how you can ask for a free trial, book a demo or read more about our how secure we are. Okay. Any last questions? We have eight minutes left. I'm just gonna go into the chat. >> I think I see any additional questions. I think we've managed to handle them all this time. >> Amazing. Um, so there is a link that Dory's just dropped. So if you wanted to book a demo, please feel free to do so. Um, again, if you want to ask for a free trial, please do. um is the playbook creation on the basis of existing document NDA available with law insider AI. So there are two ways that you can create a playbook. You can either um build it yourself [snorts] or you can use your document. So if you opened up your template NDA, you can use the AI to build a playbook based on the principles that underpin your NDA. Alternatively, you can just take one of our standard um playbooks. So, our mutual NDA playbook, we're going to brand it clearly soon so that you know that it's one NDA, but it's based on one NDA. So, you could just pick up our standard NDA playbook and then either tweak it or leave it as is and use it uh yourselves for your own needs. Um, does the system work with SharePoint? We do offer integrations. Um Aisha, so please do book a demo so that we can talk you through that and our sales team can explain how integrations work. Okay. Well, if there are no other questions, uh of course Adele, that's great to hear. Um if there are no other questions, I can give you seven minutes minutes back. I'm going to hang around in case there are any last minute questions for the next couple of minutes. Otherwise, thank you all so much for joining. It's been a pleasure. And uh if you would like to contact me directly about anything, I'm just going to drop my email here. So you can totally do that. If there's anything that I covered today that you wanted a bit more insight, the recording will be sent directly after this. So please uh bear with us an hour or two or maybe a bit more Dory. Um and then we will send you the recording. Um Afia, yes, we integrate with anything that you need us to more or less. So again, please book a demo and we can uh explain that more. Thank you. Yes, Dory. Thank you.