Episode 63: Unveiling the Power of Document Automation with Catherine Bamford

This is Legal Design Podcast and we are hosts Nina Taiwanen. And Henna Taiwanen, today we went through one of the most classical examples of automating legal processes and talk about document automation. We are joined by the notorious Guru of this field, Katherine Bamford of Family Legal. Katherine explains why she is so passionate about document automation and what are the benefits of it. With over a decade of experience of transforming and digitalizing legal work, she shares her secrets on what to do and what to avoid in automation projects. Katherine also tells us why organizations should create and maintain contract labels. We also discussed the role of document automation and GPT tools. Katherine explains whether we still need human-made templates that are automated or should we concentrate on prompting and training GPT tools to provide us drafts that will get us started in our work, just like templates. Hi Katherine and welcome to Legal Design Podcast. How would you like to introduce yourself to our listeners? Hi Henna Hi Nina, thank you so much for having me on as a guest, really excited to be here. So yeah, my name is Katherine, I am the CEO and founder of Family Legal. We are a Legal Technology Consultancy and we help law firms and in-house legal departments to improve the way they deliver their legal services by using technology and what we focus on and kind of my area of expertise, if you like, is document automation and automating the contracting process. Welcome to the podcast. Hi one question, I think legal, Family Legal has existed for at least a big aid now. Yeah, I've been, I started off in legal tech kind of when it wasn't known, you couldn't just Google legal tech and have all these amazing resources come up, there was lots of people that were doing a lot of cool stuff around knowledge management and starting to look at how technology could be applied and used by lawyers to make their services more efficient but there definitely was, it just wasn't like a topic, an area if you like that everyone knew of and we actually, when I first left practice, so I was a practicing lawyer, I was a real estate lawyer when the first recession hit in 2004, 2005 and I was asked to go on an internal succumbent with my law firm to look at how we could use automation software to speed up the way we were doing things because that was the point where the whole more for less idea came around where clients were saying we want you to do more legal work for less money, because we're just less money around and everyone was having to like, you know, look at reducing the numbers of their stuff and smaller legal budgets and budgets and things, it was like, what can we do to improve? And so I ended up going on this succumbent, playing around with automation, loved it, thought don't really want to go back to being a practicing lawyer and I'd quite like to keep doing this as a job where I apply technology to what lawyers do and see how it can improve. But we didn't know what to call me, we had no job title, so one of the people I worked with at the far more lander, Kanata, he said, well, years ago, Richard Sothkin wrote that he predicted in the future, there would be something called, there would be a job title called a legal knowledge engineer, so we stole that and used that, so that was like the first kind of basis of a legal engineer, if you like, and then when I set up by a legal, I dropped the knowledge bit because it's like, well, it's not just the content, knowledge is just the content, it's also about the process and everything else on the actual tech. So it became legal engineer and now 10 years on, you know, it's, it's, so we made it up, basically it's a job title and now 10 years on, you know, I meet people at legal gig that have just entered the kind of market and they're like, oh, I'm a legal engineer and I'm like, I love that this has become an actual career path that people say is if this has been a job title for years and years because we kind of made it up, so yeah, I kind of like that that became a thing. So the legal tech world knows you as a legal tech guru, a legal engineer who is especially passionate about document automation, that's the first thing that comes up when we Google your name, so legal documents in general, however, they may not be something people are typically passionate about, the first thing that comes to mind about legal documents is not that fascinating necessarily. But what sparked your interest in document automation, Catherine? Probably exactly that, probably that, so I was a commercial real estate lawyer. So working for kind of large investors, investment companies, pension funds, that kind of thing, who had a portfolio of properties, so like office blocks, shopping centers, industrial estates, and we would do all, you know, when you say you're a commercial lawyer, you know, everyone's like, oh, so exciting, commercial real estate, you're working on these big developments. It sounds exciting, but the day to day as the junior was I was drafting leases, licenses, rent deposit deeds, notices again and again and again, inserting the same information into these documents over and over and over. So the fact that it was kind of boring was what sparked my interest because I was constantly thinking they're mostly a better way. And with document automation, your real estate is a fantastic place, an easy place to start because the information, you very rarely draft just one document for a client on any given deal, so like the idea of a lease, so you've got a landlord client who owns a shopping center and they've got an empty unit in the shopping center and a new tenant is moving in. And the tenant going in is more than likely, unless they've been around for years and years and years, going to parent deposit, and they're more than likely going to go in and say it's like a closed shop, they're going to want to put in their own changing rooms and maybe a back warehouse to keep their stock, so they're going to do small durations, so you're never just drafting a lease, you're drafting a lease, a rent deposit deed, license alterations. The information that goes into all three documents, the data that goes into all three is exactly the same. So no lawyer should be sitting doing that over and over, just for anyone that hasn't heard of it, document automation and what we're talking about here is rather than a lawyer preparing a legal document themselves in Microsoft Word, typically is how they do it these days, how they used to do it was printed out, marked it up by hand and give it to a typist, nowadays it's kind of on Microsoft Word, they'll go through the entire document and fill in all the client specific information, what we're talking about now is rather than doing that, they answer a series of questions about the particular deal online and it then automatically puts the information, the answer to those questions into the document and drafts the document for them. And if they say, you know that there is a rent deposit, it will automatically put a clause in around the rent deposit and it will put definitions in that relate to that clause. So just, yeah, probably stinging step back, that's that's kind of what it is. But what sparks my interest was was just that I'm definitely one of those people, it's like there must be a better way, like I used to be a bar made, when I was in law school, I used to kind of work in a bar in the evenings and often during a really busy shift, one of the barrels of beer would run out and we'd have to run down to the seller and we'd have to change the taps. The most common beer was at the back of the seller behind other barrels and you had to kind of move them out of the way to get to it and then the tap was awkward. So one quiet day without asking anyone, I just went down and I rearranged all the barrels at the time to make it faster for everyone, you know, and that's just, so I'm just always being that kind of person. Yeah. So, Katherine, you mentioned that often those documents are conducted by two nearest lawyers, they're the ones drafting. Actually, there's been some discussion in Finland, like how do we teach the two nearest staff members if we use automated documents and they don't have to do the drafting and learn the law by doing any comments on that. Great question and comes up a lot and actually the research that we've done and the kind of, because we always go back to feedback and interview people that are now using document automation and that's how it's changed. The feedback is that it is a better and easier way to teach junior lawyers because the way of getting an example document, the old way of getting an example document that maybe has like the notes at the bottom saying include this clause if this, you can put that next to the question. So it's what's known as just in time knowledge. So as they're considering the question, you put that guidance next to it and you can, you can make certain questions compulsory. You can have links to external websites that'll open up in a new window to explain it. By having the question, so what I used to always do before document automation when I was drafting certain things was I would print off there was a checklist on practical at the time that I used to always print off to make sure that I'd done all of those things of those steps, your tomato question out. works as that mental checklist to make sure you've considered all the things you should be considering when drafting that document. Most of the software as well, so a nice other feature has a have a preview screen. So as you answer the questions, you can look at what's changing in the document and rather than having to turn every page, you can jump between the relevant sections and see what's changing. And then obviously it's only helping with the first stage, the drafting. You then do have to check it and then you have to do the bit that's in the most interesting bit with the negotiation with the other side. So you're constantly working with the documents and getting to know them. So yeah, I think the idea that it works as actually a mental checklist, you can put just in time knowledge next to every question and link to external guidance. And the fact you have a preview screen that shows you exactly what is changing in the entire document. So you can't miss anything. Actually helps the lawyers to skill up faster than take away their learning of drafting. That's a brilliant point of view. Thank you. So when we're talking about document automation, is there like some sort of like academic definition for that or how proly should we understand the concept, for example, do smart contracts or automated legal decision making fall into this category of document automation? I don't know if there is an accepted definition, but my take on it is that document automation is using software to automate the drafting of any legal document. So it doesn't have to be just contracts. A lot of people think straight away of contracts. I've automated litigation documents. I've automated a lot of reports. So due diligence reports, it can dramatically speed up the production of reports by multiple lawyers when you need the output to be consistent. And actually one of the first big projects I did was exactly that we had 50 lawyers looking at a huge portfolio of real estate properties that were securing loans. And it was when one of the banks, the Irish bank debts, it was again during the recession, and it was kind of they needed to turn around all of these reports in like a two week period just before Christmas to be able to get this deal done. And when we thought about it, the biggest bottleneck, the biggest time taken or delay in that process was a senior lawyer having to check all the junior lawyers reports. And often the way one junior lawyer would say the rent was payable was just different grammatically to how another lawyer would say it. So by automating those reports and asking the questions, you can make the output really consistent. And then it's massively speeds up the part of the senior lawyer checking the output. So that's something that can be useful. In terms of kind of decision making and workflow, I would say that's more just automation rather than documents. Document automation, let's focus on the word document. But there is also there is yeah, document automation is just one small piece of any legal transaction. There's obviously the taking the clients instructions, doing initial research and asking questions, then drafting of the documents, then negotiating the document, then generally some sort of money changing hands or consideration, someone selling something or they're settling something or decisions being made as two next steps. And then there's completion where generally documents are being signed. So there's things like electronic signatures can come in. Then the storing of the documents and information retrieval that I would say is more automation. Some people call it CLM, but I think that's not wide enough. That's automation of an entire legal document. And there's lots of different technologies you can do for that. There are document automation software that have workflow components. And when people are choosing which document automation software to buy, they really should think about the requirements very, very strongly and be really clear on that because every document automation software is slightly different. And if something like workflows, something you're going to focus on a lot, it does eliminate many and brings others to the top of the pile as to the ones you should go with. With smart contracts, people define those differently, but if what you're talking about with smart contracts, the idea of using blockchain for obligations management, then no, that's not document automation. But if what you're talking about is kind of automated contracts that go from beginning to end signed and into something without really human touch points, then yes, document automation is a part of that. Yeah, I like that. Very comprehensive answer. Thank you. Sorry, I talked too much. No, I know what document automation is now. I think I did some typical misunderstanding related to document automation. Something that people assume it is and then it is not. Yeah, I think things like the question you just had earlier about, you know, it'll stop juniors being able to learn. No, it doesn't mean you've got to switch your brain off as a learner, as they're still working to be done. I think things like it'll mean you do things faster, so you have to charge the client less or you're going to leave work on the table. Again, not the case. A lot of the time you're reducing the right off of hours that've been put down and you're enabling the same lawyer to do more added value work and just speeding up the heavy lifting. But I think the biggest misunderstanding with document automation, sadly, is that it's almost been oversold as how simple it needs to set up. And that's by the software vendors, I think. You don't just get a word document with some square brackets in and feed it in and magically it comes out as a perfect online questionnaire for your lawyers to use. There's quite a lot of work to be done in the setup of the automation, which is an area I'm doing a lot of work in at the moment, actually, like in terms of how do we speed up those parts, which I think is quite interesting. But yeah, the biggest misunderstanding is that it's just magically going to work out of the box. I think people need to understand that time and resource needs to be applied to getting your documents automations in the first place. Once they are, benefits are huge, obvious, lawyers will never want to go back to drafting in the old way. And we'll be really upset if they ever move firms and have to because the new firm doesn't have it. And that has happened to be friends of mine. And they're like, I have to do this the long way again. It's awful. But yeah, unfortunately, there is that setup time now. Hey, now that we're talking about legal tech and document automation, we, of course, have to talk about AI, which is getting so big in the legal industry. And AI and especially GPD tools have arrived in the legal industry and organizations are implementing, for example, Microsoft co-pilot and other tools like that. And I was just wondering what will be the relationship between document automation and GPD tools? Do we still need human-made templates that are automated or should we concentrate on prompting and make sure we train the GPD tools to provide us drafts that will get us started in our work just like the templates or do these two work together somehow or what should we expect? Great question. And something where we're really excited to be doing a lot of work on at the moment for various different legal firms and just experimenting in our in our own rights as well so that we can explain to people what works and what doesn't. For drafting of documents, a generative AI is is not a good tool to use for drafting because it will give an inconsistent, it'll give a different outcome every single time no matter what it's better than as it's training, you know, training information, train data. Unless you are looking for a completely obscure and bespoke clause that won in a million times that you don't need it, then you're much better to have your own consistent word templates. However, where it can be useful is in summarising extraction and summarising and that's where I think at the moment we're seeing a lot of law firms experiment and start to roll it out to their firms. I think in terms of its interrelationships with things like document automation, the bottleneck I spoke about earlier, so getting the original markup from the subject matter expert of the precedent to then turn the precedent into the online questionnaire. And also one of the other big bottlenecks is once you've automated documents, then testing it to check that it's completely accurate and ready for the lawyers to use, those bottlenecks we can definitely use LLMs to help speed up and we are doing some work at the moment around that first one, though creating the test data for complicated ones is actually an easier problem to sort of. But around the first one, because we have automated it in the last ten years, so how do we do that? many documents that are common across all law firms. So we have an knowledge of the top kind of 2030 questions asked for any document in any department of, you know, top 100 law firms. By having those questions set up and then having, if someone gives us their precedent, we can, what we were previously able to do is say, well, here's a starter for 10. So we, we, we currently help clients get their pipeline more ready by doing a first cut of that mark markup for them by giving with this menu of top 20 questions and saying, look, even if you just apply those, here's your starter for 10. But how can we, we're thinking now about how can we use LLMs to take those questions and take the square brackets and apply them and then check it to speed it up dramatically. And the more we do these experiments, the more we see, actually there's definitely something here. So short answer, generative AI, not for legal drafting, more look at, if you're looking to explore how it can be helpful to you within your law firms, start with looking at summarizing and extraction. We're always happy to talk to anyone about kind of what their competitors are doing and where they should start and how they can start and understand all the different LLMs out there to help them and kind of that noise and scanning. In terms of how it will work with document automation, we'd love to speak to some people out there who would like to be the pioneers of these experiments that we're running to see how Gen AI can help speed up the document automation process. It's so, sorry, a bit of a blade answer out there, but if anyone would like to join with us to be the pioneers in this area, we'd really like to speak to you. Sounds good. Hey, now that we have a mention Gen AI and TBT tools, we can concentrate on other legal tech issues. Gathering from your point of view, what other things should we be concentrating on when we think of legal tech and what to do in law firms and legal departments? I think there has never been a time that it's been more important to actually have a look at your knowledge management, your content and what you have. So having those good word, they don't have to be in word, but having those good templates and precedents. The speed of technology, you know, technological advances mean that it's going to be very apparent very quickly if you're, we have this expression garden isn't in order. Your own back garden isn't in order, you know. So I think it's really important that people look at the foreign content. The area that I would focus on most if I was within law firm knowledge management leadership is something that a lot of people ignored for a lot of years. And that's around what I prepare to play books. Do you know what I mean by playbooks? Explain that concept a bit or explain because it might be unfamiliar for. Yeah. So when you are preparing any sort of document for any transaction for your clients, you sit off the other side or you'll receive a document from them to agree the contract or whatever's going on between the two parties. And often there will be things that you will either accept or reject as being reasonable or unreasonable or just because they're against your clients policies. And that is what I refer to as your playbook or your client's playbook, what you will accept or reject or compromise on and the different options if you're going to compromise. And often junior lawyers will go to a senior lawyer and say I'm not sure whether to accept or reject this, what should I do? This happens particularly in house as well where junior lawyers will go to the most senior council and ask the same questions or over and over and over again. How do we get that knowledge out of the senior lawyers' heads as to what to do? They've learnt this over years and years of experience, what to accept, what to reject and how to compromise and how to change the drafting to make it acceptable. We need to get that out of their heads and gets it documented. And a lot of people have really struggled to do that. The more we can get that kind of thing down on paper, then the more we can speed up the entire process using technology if technology is the next step, but even just getting it documented to start with so that the junior lawyers can self-serve by looking something up. What I mean is like an example of an insurance clause being the buyer must have insurance, the seller must have insurance of one million pounds to cover anything that happens in the next year once the buyer's bought it. And it could be that that particular client or there's a law that says if it's above a certain threshold, it's unacceptable or what we've seen as unreasonable at court. So it's just a playbook to go never accept more than a million. Or if it's more than a million, you know, do add in this bit about this such and such court act, you know. And just telling them what to do when certain things happen, it's getting those that down on paper. The more we can do that, then when we get more advanced with using LLMs and Denner Shivaii, then we will be able, or you know, GPT will be able to use these tools to automate the next stage, the negotiation stage and do a first pass of how would you, you know, when a document arrives on the other side, we can say how should this be amended to make it suitable for our client? Yeah. So let's start walking down a memory lane a bit of talking about GPT and other technological developments we have recently experienced in the legal sector. You Katherine have been practicing legal tech already for over the gate and you've been able to see the transformation of the legal work closely through your, your own work. So if we think of different technological developments or paradigm shifts in the tech sector during the past years, what do you think that have had the biggest impact on legal work? Unfortunately, I don't think there's been any paradigm shifts. It's before too slow, far too slow. And I think there has been regular incremental improvements, but because of problems away from the technology existing itself, a lot around the law firm structure, management and resource being put into innovation and improvement, I think it's been, it's been far too slow. Law firms who don't necessarily, the leadership don't necessarily want to disrupt themselves. So I think the biggest advances have not been actually around the technology. The technology is there and available and lawyers are using new things all the time. But I think the biggest changes have been around the resource and the new jobs that now exist. So like legal engineers, the fact that a lot of law firms now have innovation teams and groups of people that go out and speak both to clients and internally with the lawyers to see how to improve the way they deliver those services. But I think it's still under-resourced massively. When you think of other industries and you think of their research and development departments, you know, normally it can be like kind of 20 to 30% of an annual budget will be applied to R&D. In legal, it's less than 2%. Oh my god, you know. Very huge gap. Yeah, so I think there's a lot of work to do and I definitely wouldn't say there's been any paradigm shifts in the last 10 years. When I first started kind of 10 years or so ago, I was so excited and I really bought into the futurists that we're saying we're going to see more in the next 10 years than we've ever seen and, you know, end of lawyers. Excuse my language, but bullshit. Yeah, bullshit. And it seems that those people are still saying, oh no, the next 10 years, the next 10 years, you know, and it's still in the making. Yeah, and I don't think there's anything wrong. I do think, you know, it's exciting whenever we get potential disruptors. But I don't think there's anything wrong with marginal gains, continuous improvement, incremental changes. I think that's exciting, you know, formula one's exciting. That's what they do all the time. Look at any sport, you know, the British cycling team, what they managed to do with continuous improvement. So exciting. Look at rocket ships and engineering. They're constantly tweaking tiny things, implementing checklists and doing all sorts of little things to like, how do we shave off a second? How do we make it like 1% lighter? If we looked at legal and we said, how do we improve profitability on every transaction by 1% in the next year? That would be more than we're currently doing with technology at the moment because we're not applying the strategy from the top in that way. Yeah. Exactly. Are you agree? Hey, we already covered this a bit, but let's talk a bit more about the document automation and the status at where we are at the moment. Because I know that some of the organizations really struggle with document automation and the work, the automation work still might be super slow. And when it's slow, it usually irritates people and people often get bored really fast. And sometimes it's hard to understand the need of manual work to get things done. Do you have any tapes for ace-truckling organization with their document automation? Yeah, sure. and you're right, people can can lose the will to live a little bit when they when they have these mammoth tests of marking up every possible option ever in a document and then testing every possible option ever once it's been automated and projects that they thought was going to take six weeks suddenly it's six months and everyone's getting a little bit confused and frustrated and unhappy and trying to fit it around their day job as I said it goes to the bottom of the park because it's not enjoyable anymore my top tip would be to do a phased approach and what I mean by that is the subject matter expert to maybe just think about to start with the first five questions and just automate the first five questions and test those then once happy with those add in the next layer then add in the next layer by doing it like that and almost running it like a project where you have okay you mark up that question then I'll automate it then you'll test it then you'll mark up the next question then I'll automate it and I'll test it right if we do it that way we think there's going to be about 12 questions we're looking at about 12 week turnaround time then and if you actually show a visual timeline at the beginning of the matter you explain to people their responsibility when by which date or else the timeline will will move weekly project calls so not you go off and automate it I'll sit here and wait till it comes back comes back weeks later then I disappear for six weeks you don't hear from me at all then I give you something that's way too big for you to test by yourself because I've done everything in it um doing it in that that short way you see progress um it feels more bite size and achievable you're setting aside maybe an hour a week or two hour a week chunk down talk to people at the bottom like any project you talk to people at the beginning about a few any holidays planned okay well then those two weeks we need to move over there then when you're reporting to leadership and you're saying how long it's going to take you're being realistic and you're saying it's going to be 16 weeks because this and you show them the timeline you know yeah yeah so phased approach um and if you can get a good project manager to help you um it's a very light touch project for a good project manager so you could have the legal engineer's subject matter expert and then it's a very small project for project manager but their input can dramatically improve the delivery right points yeah hey um so document automation it's the obvious example of automating legal work what other flow hanging frauds there are when it comes to automations in legal think electronic signatures is a really easy one that's not used enough yeah um reporting you can automate the reports like I spoke about earlier consistent reports I think there's definitely something that we don't do in legal that every other service industry does and that's just having quite clear like um you would refer to it in any other industries like a ticketing system so you know just knowing kind of like what work you have on when like decent task management would help everyone massively if they could see that kind of thing because then the lawyers could see where all of their deals are up to you would have a better view of capacity um of particular lawyers you'd more easily be able to report to the partners about where every deal is um and to the client about where things are up to so I think some sort of just simple ticketing um it's good um talked earlier about playbooks I think there's a lot that can be done just to get the knowledge out of the senior lawyers head and get their document some people think well how will I do that when I talk to people about playbooks are like oh it's so much work because they see it as sitting down with a blank sheet of paper yeah and having to come up with everything this is not a blank sheet of paper say for example I wanted to playbook on a commercial um supply agreement I would take an example contract and I would take the headings of each of the clauses and I would do a two-hour workshop with the most senior best commercial lawyer in your firm and I would say on the liability clause for example what are the possible things that someone might try and change around liability and what are the possible options of what you might do two hours together just asking those questions okay this topic common things changed come things up you've suddenly got a really really good you know it's not a starter for tennis a starter for 90 playbook yeah and then you just give it to a few people and go we've been thinking about this would you mind catching your eye over it's ready to add or any comments you know it doesn't have to be perfect from the beginning as lawyers were terrible at needing a perfect output before we do anything or or just starting things we're quite bad at starting things if we're not like you know so anything we can do to help people start and a lot of that is is often just around asking the right questions because they do have all the knowledge they just with a blank sheet of paper have sort of like perfection paralysis and don't know where to start yeah it sounds like every contractor should have a playbook accompanied with exactly oh it's time for our last question gathering um you work in the legal sector and your ambition and the energy the positive energy that you give into the changing the the way of legal work is done is very inspiring um but what inspires you to do this work at the moment or in general gathering i like the fact that working in legal technology and working in improvement of legal services and the delivery of legal services you're always at the forefront you're always learning new things are coming up new ideas can be applied you're constantly experimenting um and that keeps me excited and interested it's a wonderful community i love everyone i work with and come across because everyone in this space is trying to improve and change and are they facing the same obstacles i like hard problems i get bored easily so doing anything repetitively isn't isn't really good for me i just not built that way um so yes i think it's it's a combination of of those things but for me it's always thinking about what's next and what else is possible um i'm being being someone that likes to imagine maybe i daydream too much i don't know um but daydreaming is good yeah right it is good to i'm i'm very bad at um just complaining and thinking oh it's a problem it's a problem it's a problem i get really annoyed with with people that do that i've been to groups before for example i went to women in law groups that it was all about how there wasn't enough females in senior positions in law firms but sometimes the groups were just complaining and not offering any solutions and i would get a bit annoyed because i'll be like how can you blame the leadership if you're not saying look here's 10 things you could do to help you know don't just complain come up with solutions so um yeah i think like rather than just stay being the unhappy lawyer that i was it was kind of there must be a better way how can we change things and uh yeah hopefully the inspiring other people to do it has a bit of a a catalyst butterfly effected often i have workshops with people or if it's conversations with people and then people say they go oh do you remember you mentioned that well i went and i did that where i tried that and i'm like yes you know this is brilliant because the more people we have all thinking about how to improve the situation the faster we're all going to move exactly absolutely thank you gasrin it was a joy to have you it was so nice to to speak i found it really cathartic when people ask as questions it makes me kind of when i'm not in the weeds of my day job and you sit back and you go oh actually yeah just go stuff so thank you for that i've really enjoyed it thank you for listening to the legal design podcast episode with gasrin pamphored for more info about the episode please check our website legaldesignpodcast.com and our social media channels on instagram x and link it in