3025: Unlocking the Future of Legal Tech with AI-Driven Document Management

[00:00:03] [SPEAKER_00]: How is AI reshaping the legal industry?

[00:00:08] [SPEAKER_00]: Particularly in the realm of document review and analysis.

[00:00:13] [SPEAKER_00]: Well today I'm excited to welcome Lyson Casey onto the podcast.

[00:00:18] [SPEAKER_00]: He's the innovative founder of Discrepancy AI, a company that's pioneering the use of

[00:00:24] [SPEAKER_00]: artificial intelligence to revolutionise how legal documents are processed and analysed.

[00:00:31] [SPEAKER_00]: And with a background in AI engineering and software development, Lyson has created a platform

[00:00:36] [SPEAKER_00]: that goes beyond traditional OCR technology, addressing its limitations and offering a smarter,

[00:00:44] [SPEAKER_00]: more accurate way to handle everything from contracts to tax documents.

[00:00:49] [SPEAKER_00]: So today I want to learn more about how Discrepancy AI is transforming unstructured documents

[00:00:55] [SPEAKER_00]: and turning them into structured, searchable data, essentially making it easier for legal

[00:01:01] [SPEAKER_00]: professionals to sort, filter and gain insights from their files.

[00:01:06] [SPEAKER_00]: And we'll also explore the advanced AI capabilities that allow the platform to

[00:01:11] [SPEAKER_00]: handle complex data types in everything from charts to tables,

[00:01:16] [SPEAKER_00]: and how its pixel analysis AI is also setting new standards in detecting document tampering

[00:01:22] [SPEAKER_00]: and fraud. And finally, Lyson will also share his insights on the growing interest

[00:01:29] [SPEAKER_00]: in AI solutions within the legal industry, where some firms are even striving to become

[00:01:35] [SPEAKER_00]: AI law firms. And I find this particularly interesting because I've been doing this

[00:01:39] [SPEAKER_00]: podcast since 2015. Yep, we're approaching 10 years of doing this. And when I first started,

[00:01:45] [SPEAKER_00]: law firms have always been accused of being slow to adapt to technological change.

[00:01:49] [SPEAKER_00]: I've seen this change and evolve over the years, but to see it to go from where it was

[00:01:54] [SPEAKER_00]: when I started this podcast to law firms striving to become AI law firms, something

[00:01:58] [SPEAKER_00]: that interests me greatly. So what role will AI play in the future of legal practices?

[00:02:05] [SPEAKER_00]: And how can it enhance both accuracy and efficiency in document processing?

[00:02:10] [SPEAKER_00]: Delivering daily content to 140,000 of you wonderful monthly listeners across the globe

[00:02:16] [SPEAKER_00]: is no small feat. I don't want to take all the credit here because it wouldn't be possible

[00:02:20] [SPEAKER_00]: without the backing of our dedicated sponsors and partners. And I also want to shine a light

[00:02:25] [SPEAKER_00]: on the fact that legacy managed file transfer tools are looking dated. They often lack the

[00:02:29] [SPEAKER_00]: security that today's remote workforce demands. And companies that continue relying

[00:02:33] [SPEAKER_00]: on that outdated tech, they're in danger of putting their sensitive data at risk.

[00:02:37] [SPEAKER_00]: Attention IT professionals, are you tired of juggling these multiple servers for secure

[00:02:43] [SPEAKER_00]: file sharing, integrated shared folders and email plus a comprehensive rest API?

[00:02:48] [SPEAKER_00]: Kiteworks simplifies your workflow. For administrators, you can experience unmatched

[00:02:53] [SPEAKER_00]: functionality and integration that traditional MFT servers just can't touch. Step into the

[00:02:58] [SPEAKER_00]: future of secure managed file transfer with Kiteworks by going to kiteworks.com to get

[00:03:07] [SPEAKER_00]: also FedRAMP moderate authorised. Thank you for your patience today. This is the moment

[00:03:12] [SPEAKER_00]: you've been waiting for. It's time to welcome my guest onto the show.

[00:03:17] [SPEAKER_00]: So a massive warm welcome to the show. Can you tell everyone listening a little

[00:03:21] [SPEAKER_01]: about who you are and what you do? Hi, my name is Lyson. I am the founder

[00:03:26] [SPEAKER_01]: of Discrepancy AI. We're a platform that reviews and analyzes documents.

[00:03:34] [SPEAKER_01]: Specifically, our tool is able to convert any document into structured JSON and insights

[00:03:40] [SPEAKER_01]: and is able to standardise all of your unstructured data. Whether you have

[00:03:46] [SPEAKER_01]: thousands of documents that you don't know what they're about, or you have to review

[00:03:50] [SPEAKER_01]: documents on behalf of your clients, we make it easy and automate the entire process

[00:03:55] [SPEAKER_01]: of extracting information from these documents, analysing it and providing our customers with

[00:04:00] [SPEAKER_01]: only the important information that's within those documents.

[00:04:04] [SPEAKER_00]: Well, a question I've got to ask here is what was your origin story? It's got to be something

[00:04:08] [SPEAKER_00]: that inspired you to create Discrepancy AI. How do you see revolutionising the legal document

[00:04:15] [SPEAKER_00]: review process? Were you involved in that industry in a former life? What's the story there?

[00:04:21] [SPEAKER_01]: Yeah, so I was actually... My background is as an engineer, as a software developer,

[00:04:26] [SPEAKER_01]: I was working at a company called WiseDocs, helping them build their GenAI capabilities.

[00:04:34] [SPEAKER_01]: I kept seeing the same problems over and over again with what they do specifically in medical

[00:04:39] [SPEAKER_01]: documents. I kept seeing the same problem over and over again where the actual... It's called

[00:04:44] [SPEAKER_01]: OCR, where the actual first step in extracting the information from these documents where

[00:04:49] [SPEAKER_01]: you have an image or a PDF with no text in it and you have to extract the text that's

[00:04:55] [SPEAKER_01]: document. Right now, the standard that everybody uses like 99% of document processing goes through

[00:05:01] [SPEAKER_01]: a process called OCR. This OCR extracts the text from a document and puts it in a position

[00:05:08] [SPEAKER_01]: in that document where it thinks it is. Believe it or not, this system was created

[00:05:12] [SPEAKER_01]: in 1974 and it's still being used today with all of the most cutting edge LLMs and all of

[00:05:19] [SPEAKER_01]: other AI tools to analyze these documents. Our first run through these OCR models,

[00:05:26] [SPEAKER_01]: they were created in 1974. So I kept seeing the same problems over and over again where

[00:05:31] [SPEAKER_01]: the extracted data was incorrect. It was missing a bunch of context and it just wasn't

[00:05:35] [SPEAKER_01]: working well and it was having a really hard time building the GenAI models because

[00:05:41] [SPEAKER_01]: you've probably heard it before but garbage in, garbage out. So after that position there,

[00:05:48] [SPEAKER_01]: that's what I wanted to fix. I knew that was a problem with every company that deals with

[00:05:57] [SPEAKER_01]: documents, not just the medical space, not just the legal space but every single company that

[00:06:02] [SPEAKER_01]: deals with unstructured documents has to first OCR them and everything they do after that

[00:06:08] [SPEAKER_01]: carries the problems and issues that the OCR has. So that's why when I started Discrepancy AI,

[00:06:14] [SPEAKER_01]: I knew that that's the problem that I had to solve first and that's why I built our Optic AI

[00:06:19] [SPEAKER_01]: model that can convert any file directly into the structured JSON, that's the technical term,

[00:06:25] [SPEAKER_01]: but into the structured data that you need. So whether it's an image, whether it's a PDF,

[00:06:30] [SPEAKER_01]: a screenshot, whatever that is, we extract the standard information from that document.

[00:06:36] [SPEAKER_00]: Wow, that's incredibly cool. I was getting flashbacks of the garbage in, garbage out

[00:06:41] [SPEAKER_00]: problem. I think that would have resonated with so many people listening and the question

[00:06:45] [SPEAKER_00]: I've got to ask as well is how do you at Discrepancy AI, how does your technology

[00:06:50] [SPEAKER_00]: convert these unstructured documents which can be anything from images to unformatted PDFs

[00:06:56] [SPEAKER_00]: and put them into searchable, filterable structured data? Can you talk me through

[00:07:02] [SPEAKER_01]: the process there, how you do that? Yeah, well we've created kind of a unique system

[00:07:06] [SPEAKER_01]: that we're taking advantage of something called AI tool chaining. So we have dozens of specific

[00:07:14] [SPEAKER_01]: and fine-tuned and trained AI models that all have independent individual small tasks

[00:07:20] [SPEAKER_01]: and they go out and work, some of them synchronously, some of them asynchronously

[00:07:25] [SPEAKER_01]: and they work together, they pass data from one to the other in a chain and building these

[00:07:33] [SPEAKER_01]: AI tool chains in such a specific way that we extract the information and convert it into

[00:07:40] [SPEAKER_01]: the JSON structure. So it is not just like one model that we built, it's multiple models that

[00:07:46] [SPEAKER_01]: all work together, they all have their little part and they pass on what they generate to the

[00:07:52] [SPEAKER_01]: next AI models and by chaining these AI models together we can get results that are way better

[00:08:00] [SPEAKER_00]: than the sum of their parts. And I suspect that anyone listening who works in a law firm will say

[00:08:05] [SPEAKER_00]: hey yeah that sounds great, but legal documents often contain complex data and complex formats

[00:08:11] [SPEAKER_00]: and everything from charts to financial records. So how does your AI handle some of these

[00:08:17] [SPEAKER_00]: complexities and what makes it stand out in analyzing so many diverse types of documents

[00:08:22] [SPEAKER_00]: that that typical law firm will encounter? Yeah, so you hit the nail on the head that

[00:08:28] [SPEAKER_01]: is always the most difficult part, these non-standard document types like charts, graphs, tables.

[00:08:35] [SPEAKER_01]: So the revelation that we had is that converting this data to JSON and a lot of the listeners

[00:08:42] [SPEAKER_01]: might not know what JSON is but it's just a type of structured data that the entire internet

[00:08:50] [SPEAKER_01]: runs on. So the entire language of every website, the data that comes in from like a server or

[00:08:59] [SPEAKER_01]: database and populates like a front-end website for example with that data, that data is always

[00:09:04] [SPEAKER_01]: JSON and this JSON data is structured in its nature. So the revelation that we had is that we

[00:09:11] [SPEAKER_01]: can actually convert these tables, these charts into this JSON structure and that actually

[00:09:17] [SPEAKER_01]: preserves all of the information within where we can actually even reconstruct this table

[00:09:23] [SPEAKER_01]: that was in the PDF from the JSON data that we extracted. So that is the big revelation here

[00:09:30] [SPEAKER_01]: is that every other platform is going to JSON using some custom code. They have an AI that

[00:09:38] [SPEAKER_01]: returns text or they return some data that's not standardized and then they have developers

[00:09:43] [SPEAKER_01]: that are through either logic like doing like if statements or building specific custom code

[00:09:51] [SPEAKER_01]: to convert that text data into the JSON data that their website expects because every website,

[00:09:58] [SPEAKER_01]: every app at the end of the day works via this JSON data and what we've done that's

[00:10:03] [SPEAKER_01]: different is we skip all those steps and go directly into the structured data that

[00:10:09] [SPEAKER_01]: companies need. So our model is specifically designed for these difficult data types,

[00:10:17] [SPEAKER_01]: for these charts, for these tables, for information that's hidden within kind of like

[00:10:23] [SPEAKER_00]: a complex structure. Before you came on the podcast as well I should say I was doing a

[00:10:29] [SPEAKER_00]: little research on you and I was reading that your pixel analysis AI can also detect signs of

[00:10:35] [SPEAKER_00]: tampering or fraud within documents which feels incredibly cool and a huge importance

[00:10:41] [SPEAKER_00]: in that industry. So can you just tell me how that feature works and also for people

[00:10:44] [SPEAKER_00]: listening outside of legal firms the significance of this in legal practices?

[00:10:50] [SPEAKER_01]: Yeah so our pixel AI analysis is like an image analysis model. It looks at the document

[00:10:59] [SPEAKER_01]: on the pixel level and looks for signs of tampering fraud or editing. So it's looking for

[00:11:05] [SPEAKER_01]: as a human would if they were looking at a document, it's looking at hey does this font

[00:11:10] [SPEAKER_01]: match throughout the document? Is the line spacing even throughout the document? If somebody

[00:11:16] [SPEAKER_01]: has edited and not placed the edited thing in the exact place it's going to be slightly

[00:11:21] [SPEAKER_01]: askew. So it's looking at things like that whether hey is there a slight border?

[00:11:26] [SPEAKER_01]: Is this white box different than another white? A lot of times people will just put a box over

[00:11:34] [SPEAKER_01]: information they're trying to hide and that will leave kind of like a light border or a

[00:11:39] [SPEAKER_01]: light shadow. So it's looking at the document as a human would look at it to try and find

[00:11:45] [SPEAKER_01]: evidence that it was tampered or edited. So I think that the impact of this is not

[00:11:56] [SPEAKER_01]: it's KYC or it's called know your customer. Whether it's someone looking at customers trying to

[00:12:01] [SPEAKER_01]: provide a loan to them and they need to look at proof of income. So it has a lot of ramifications

[00:12:08] [SPEAKER_01]: throughout the industry but in the legal industry specifically where you want to make

[00:12:13] [SPEAKER_01]: sure that the authenticity of the documents is the utmost importance and that's what we're

[00:12:17] [SPEAKER_01]: trying to make sure. We're trying to look at these documents say hey this has been edited

[00:12:22] [SPEAKER_01]: we have this is the evidence that we have that it's been edited and kind of like what sets our

[00:12:29] [SPEAKER_01]: system kind of apart from other systems that can do that is that it's not just a pass or fail.

[00:12:35] [SPEAKER_01]: We actually give you we pinpoint to the exact location on the document where we think there's

[00:12:40] [SPEAKER_01]: an issue. We give you a description of what we think the issue is and we also get a flag

[00:12:45] [SPEAKER_01]: low medium high of how impactful or how important the AI thinks that that issue is.

[00:12:53] [SPEAKER_01]: So we're determining fraud and determining edited documents it's really challenging.

[00:12:58] [SPEAKER_01]: You have to think about false positives and false negatives as well when you are analyzing

[00:13:04] [SPEAKER_01]: them so we want to make sure that we provide as much feedback as much to the person saying

[00:13:11] [SPEAKER_01]: this is why we think this document is fraudulent but here's the exact reasons and reasoning behind

[00:13:18] [SPEAKER_01]: why we think there's an issue not just a pass or fail which is what most of the other

[00:13:23] [SPEAKER_00]: companies try to do this today are doing. Well that's incredibly cool. For people listening

[00:13:29] [SPEAKER_00]: again I think every business in every industry people are always nervous about tech projects

[00:13:35] [SPEAKER_00]: and they will be asking questions such as what's the ROI, what business value will it offer

[00:13:40] [SPEAKER_00]: the business. So how has discrepancy AI enhanced the efficiency and also the accuracy of legal

[00:13:47] [SPEAKER_00]: professionals? Do you have any real world examples where the platform has made a significant impact

[00:13:52] [SPEAKER_00]: which would just allow listeners to understand that the kind of tangible results we're

[00:13:57] [SPEAKER_01]: talking about here? Yeah for sure well a lot of our customers are using our system to fully

[00:14:03] [SPEAKER_01]: automate their document processing flow so whether they have documents that are

[00:14:09] [SPEAKER_01]: from their customers so whether you're a law firm and your customers are uploading

[00:14:15] [SPEAKER_01]: or whoever in the trial or in the case is uploading thousands of pages of documents and

[00:14:21] [SPEAKER_01]: you need to look for specific information within those documents a lot of times that is

[00:14:26] [SPEAKER_01]: automated nowadays. It is people scouring those documents looking for the specific

[00:14:31] [SPEAKER_01]: information or the specific issues or insights that you specified. We are really

[00:14:38] [SPEAKER_01]: automating all of that. We're trying to remove as much human intervention as possible

[00:14:45] [SPEAKER_01]: and automate these documents and we've seen a lot of really useful use cases in

[00:14:54] [SPEAKER_01]: screening documents in the legal profession. One customer was complaining to us that

[00:15:04] [SPEAKER_01]: a lot of times words that are spelled correctly within their documents but it's a different word

[00:15:10] [SPEAKER_01]: for example there was no way for them to detect that so they use the correctly spelled word

[00:15:16] [SPEAKER_01]: but it's the wrong word for that document and they have teams of legal ops people

[00:15:24] [SPEAKER_01]: whose job is just to make sure that those documents are all okay and these are automation

[00:15:30] [SPEAKER_01]: steps that our platform can handle really easily and because we standardize this information

[00:15:36] [SPEAKER_01]: it's able to recall and it's able to construct the appropriate data in a better way.

[00:15:45] [SPEAKER_00]: As the pace of technological change continues to ramp up and AI continues to evolve are there

[00:15:52] [SPEAKER_00]: any challenges that you see in integrating AI in traditional legal practices and how are you

[00:15:59] [SPEAKER_00]: helping with discrepancy AI in addressing some of these challenges because you can have the best

[00:16:04] [SPEAKER_00]: solution in the world but the implementation side of it can be quite painful in many tech

[00:16:10] [SPEAKER_00]: projects and also the things like the work culture that needs to change. What kind of

[00:16:14] [SPEAKER_01]: challenges are you seeing here? Yeah so there's definitely a lot of challenges

[00:16:19] [SPEAKER_01]: and especially in this industry even if your product is free there are things known as

[00:16:24] [SPEAKER_01]: hidden costs so even if you offer a free product to a lawyer or a law firm they're

[00:16:29] [SPEAKER_01]: hidden costs or hey if I'm gonna upload these documents to this platform what will they do

[00:16:33] [SPEAKER_01]: with these documents and so that's really something that we have to reassure and we

[00:16:38] [SPEAKER_01]: don't store any of our documents. We have no code implementations where you can just paste a link

[00:16:44] [SPEAKER_01]: on your website and it would allow your customers to upload their documents in our

[00:16:49] [SPEAKER_01]: system or automate the analysis of those documents so making sure that we have ways to

[00:16:54] [SPEAKER_01]: integrate via no code solution and making sure that we're very very responsible with our users

[00:17:00] [SPEAKER_01]: data and then we kind of like a lot of times you have to reiterate that over and over again

[00:17:06] [SPEAKER_01]: because rightfully so customers are very reluctant to believe startups when they say

[00:17:12] [SPEAKER_01]: we're going to be responsible with your data because there's been a lot of evidence in the

[00:17:16] [SPEAKER_01]: past where companies and especially tech companies have not been but we try to go above and

[00:17:23] [SPEAKER_01]: beyond by allowing our customers to purge their data at any point at the time delete all the

[00:17:29] [SPEAKER_01]: documents we also have ways to set up automated document deletion so you can say hey every 30

[00:17:36] [SPEAKER_01]: days I want all my documents deleted automatically so yeah we try to we don't train on any of our

[00:17:43] [SPEAKER_01]: customers data so we try to go above and beyond in terms of protecting our customers data

[00:17:49] [SPEAKER_01]: and then we try to make the integration as simple as possible by making it so you don't need

[00:17:54] [SPEAKER_01]: developers on your team to integrate with our AI solution and start analyzing these documents

[00:18:00] [SPEAKER_01]: you can do it simply by pasting a link and then you're good to go.

[00:18:04] [SPEAKER_00]: Fantastic and I would imagine as well that your analysis will also have to meet specific

[00:18:10] [SPEAKER_00]: and often stringent requirements from legal professionals and indeed their clients is that

[00:18:17] [SPEAKER_00]: challenge to you or are you you've probably seen it all before in your previous life as well?

[00:18:22] [SPEAKER_01]: Yeah for sure it's it is I would say one of the biggest issues with AI as it stands today

[00:18:31] [SPEAKER_01]: is that the person using it has to very specifically define what it is that they're

[00:18:36] [SPEAKER_01]: looking for and what it is that they're trying to analyze and this is where I see that actually

[00:18:42] [SPEAKER_01]: to build a relationship with our customers and try to fully understand their problem

[00:18:47] [SPEAKER_01]: understand what they're trying to accomplish not just build like a cool AI thing but really do

[00:18:54] [SPEAKER_01]: something that solves their problems. That's the approach that I'm trying to go for trying to

[00:19:00] [SPEAKER_01]: really holistically understand what it is the problem that our customers are facing and how

[00:19:05] [SPEAKER_01]: I can solve that problem without the AI come second you know so that's just that's my approach

[00:19:12] [SPEAKER_01]: to it where I just want to solve problems for our customers and AI is just this really great

[00:19:17] [SPEAKER_00]: tool to allow that. And I think law firms have had a reputation for being way way too

[00:19:23] [SPEAKER_00]: slow in adapting to technological change and thankfully I think that has changed over the last

[00:19:28] [SPEAKER_00]: five years but how do you envision AI further transforming the legal industry in the next

[00:19:34] [SPEAKER_00]: few years ahead particularly in areas like document analysis fraud detection and overall

[00:19:39] [SPEAKER_01]: legal operations is that something that excites you here? Yeah so I've spoken to hundreds of

[00:19:45] [SPEAKER_01]: legal teams about this and have learned some really interesting stuff. The main thing is

[00:19:52] [SPEAKER_01]: that most medium to large law firms are trying are looking at implementing kind of like their

[00:19:58] [SPEAKER_01]: custom solutions. They are more reluctant to kind of have a third party on their platform

[00:20:09] [SPEAKER_01]: doing all this stuff so a lot of them are looking to build custom solutions and that's

[00:20:12] [SPEAKER_01]: why we offer like white label solutions for our customers as well but a lot of them are trying

[00:20:17] [SPEAKER_01]: to have their law firm be an AI law firm in a way where kind of like the first step in

[00:20:26] [SPEAKER_01]: document analysis for them is this AI tool. So that's something that has been really interesting

[00:20:33] [SPEAKER_01]: to learn about how the legal profession is trying to join forces with AI and not just

[00:20:41] [SPEAKER_01]: kind of like use it as a tool like other digital document systems or they're trying to

[00:20:48] [SPEAKER_01]: make it a part of their business and that has been an interesting thing to learn and

[00:20:53] [SPEAKER_01]: speaking with our customers how many of them are interested in the white label solution in having

[00:20:59] [SPEAKER_01]: our AI product integrated into their kind of like ecosystem and allowing

[00:21:05] [SPEAKER_01]: their customers to review their documents like that. So that has been something that's

[00:21:11] [SPEAKER_01]: been really interesting to me and that has been kind of like I didn't expect them going in.

[00:21:16] [SPEAKER_01]: I thought that they would just want to use it as just a third party tool but seeing how

[00:21:21] [SPEAKER_01]: many want to integrate and want the AI to be part of their ecosystem they want to be seen as an AI

[00:21:27] [SPEAKER_01]: law firm that has been kind of like a really game-changing insight.

[00:21:34] [SPEAKER_00]: Well thank you so much for coming on and sharing your insights today. I think it will

[00:21:39] [SPEAKER_00]: be so valuable to people listening around the world. There's a little reward I want to see if

[00:21:43] [SPEAKER_00]: there's something we could do for you here. Before we started recording you were telling me

[00:21:47] [SPEAKER_00]: how you used to host your own podcast, how valuable it was for networking etc. Now some

[00:21:53] [SPEAKER_00]: of the biggest names in business, VC funding and tech have either been guests or maybe just maybe

[00:21:58] [SPEAKER_00]: listen to this podcast so is there a person you'd love to have a private breakfast or lunch

[00:22:03] [SPEAKER_00]: with and why? He or she might just get to hear this but let's see what we can manifest

[00:22:09] [SPEAKER_01]: together. Who would it be? Who would it be? So for me the number one I just want to talk

[00:22:16] [SPEAKER_01]: to as many customers as possible. There's no one kind of like person I idolize.

[00:22:23] [SPEAKER_01]: I think of every person is a person so that's kind of like a saying that I say like all

[00:22:29] [SPEAKER_01]: people are people so for me the main anybody who has was having a problem that they think

[00:22:37] [SPEAKER_01]: AI can solve that get in touch and but if we want to get like way out there I would love to meet

[00:22:47] [SPEAKER_01]: Sam Altman. I really want to know what he has kind of like what his future plans are for

[00:22:55] [SPEAKER_01]: OpenAI and where he wants to take that into the future so yeah having a conversation with

[00:23:00] [SPEAKER_01]: him would be amazing but like everybody wants to have a conversation with him right now so

[00:23:05] [SPEAKER_01]: I think that would be difficult but more down to earth I really just want to talk to people who

[00:23:11] [SPEAKER_01]: are having problems with documents and analyzing those documents and they just need assistance

[00:23:17] [SPEAKER_01]: or like a direction and where to go you know we don't have to you don't have to integrate

[00:23:21] [SPEAKER_01]: or use discrepancy AI. I would love to just have those conversations and learn what people

[00:23:26] [SPEAKER_00]: are doing. Love that perfect answer. I always say we're all just connected balls of energy.

[00:23:31] [SPEAKER_00]: All as one almost and I love how you just want to talk with people

[00:23:36] [SPEAKER_00]: not necessarily to sell them discrepancy AI of course but to hear the problems learned from

[00:23:41] [SPEAKER_00]: each other that's where the magic happens and Sam Altman if you are listening out there let's

[00:23:46] [SPEAKER_00]: see what we can make happen. We're throwing this out into the universe I urge you to

[00:23:50] [SPEAKER_00]: get in touch and speaking of that for anyone listening wanting to connect with you wanted to

[00:23:56] [SPEAKER_00]: find out more information about discrepancy AI. Where would you like to point everyone?

[00:24:00] [SPEAKER_01]: Yeah so you can you can go to discrepancyai.com and send us a message or you can you can email me at

[00:24:08] [SPEAKER_01]: license at discrepancyai.com that's l-i-s-e-n at discrepancyai.com or follow me on LinkedIn.

[00:24:17] [SPEAKER_01]: I am the only licensed Casey so it's going to be easy to find. Love it well I'll make sure

[00:24:24] [SPEAKER_00]: these links to everything so everyone can find you nice and easy we covered a lot today from

[00:24:28] [SPEAKER_00]: automate a document review analysis and analysis using AI converting any document into structured data

[00:24:34] [SPEAKER_00]: that you could just search sort and filter and get insights from the time saved on stuff like

[00:24:40] [SPEAKER_00]: that must be phenomenal but more than anything just thank you for sharing your story today.

[00:24:44] [SPEAKER_01]: Thank you so much Neil it was a blast speaking to you and yeah you did it you asked some really

[00:24:49] [SPEAKER_00]: great questions. I think my conversation with license Casey today has really shed light on the

[00:24:54] [SPEAKER_00]: powerful ways AI is transforming the legal landscape and discrepancy AI is at the forefront

[00:25:00] [SPEAKER_00]: of this change by offering this sophisticated solution that not only automates document

[00:25:05] [SPEAKER_00]: processing but also provides a level of precision and security that's crucial in the legal and

[00:25:11] [SPEAKER_00]: financial industries and his insights today into the limitations of traditional OCR technology

[00:25:18] [SPEAKER_00]: and how this new platform overcomes those challenges with advanced AI have all painted

[00:25:23] [SPEAKER_00]: a clear picture of where the future of legal tech appears to be heading but what stood out

[00:25:29] [SPEAKER_00]: for you most in today's discussion? Was it the potential of AI to uncover discrepancies that would

[00:25:36] [SPEAKER_00]: otherwise go unnoticed or was it more the shift towards law firms that are embracing AI

[00:25:42] [SPEAKER_00]: as integral part of their operations? I think these innovations are driving significant changes

[00:25:49] [SPEAKER_00]: in how legal documents are handled and I think it's exciting about just how much more efficient

[00:25:55] [SPEAKER_00]: and accurate legal practices could become as a result. Before I leave you as we conclude this

[00:26:00] [SPEAKER_00]: episode consider how these advancements might impact your own interactions and your own legal

[00:26:07] [SPEAKER_00]: documents whether in a professional or even personal context. How could AI driven tools like

[00:26:14] [SPEAKER_00]: AI change the way you handle and analyse important document and data? Hopefully this episode has

[00:26:22] [SPEAKER_00]: sparked a few new ideas and provided a few funny building insights along the way but let me know

[00:26:27] [SPEAKER_00]: your thoughts tech blog writer at outlook.com, Twitter, LinkedIn, Instagram just add Neil C Hughes

[00:26:33] [SPEAKER_00]: but as always thanks for tuning in be sure to join me again tomorrow morning

[00:26:38] [SPEAKER_00]: where we will keep exploring a different area of business and technology every single day.

[00:26:45] [SPEAKER_00]: And with that in mind I hope to see you all in the morning okay speak to you then bye for now.