Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Ep 39: Legal Process Orchestration: Sagi Eliyahu, CEO & Co-founder, Tonkean (w/ cohost Akshay Verma)
70 Plays8 months ago
What is the role of tech in making legal and business teams work better together? How can process orchestration improve your intake process and seamlessly connect silos in your organization? And what is it like to build a tech company that serves legal professionals?
Sagi Eliyahu, CEO and co-founder of Tonkean, didn’t set out to make a platform that automates and streamlines processes for lawyers. But focusing his startup’s point of view on elevating people over data allowed him to identify the issues that legal teams—some of the most people-centric and complex teams in an organization—need to resolve to do their work better, faster, and with more time to focus on solving problems.
Hear Sagi discuss the meaning of process orchestration, his experience immigrating to the US and starting a tech-company, and where he thinks AI and machine learning are headed. And for the first time ever, Spotdraft COO Akshay Verma joins Tyler as co-host and draws upon his background in legal ops at Coinbase and Meta to contribute keen
insights into the conversation.
Read detailed summary: https://www.spotdraft.com/podcast/episode-39
Topics:
Introduction: 0:00
Understanding how the immigrant experience shapes your business acumen: 6:16
Founding Tonkean and focusing your sales efforts on legal: 15:13
Defining process orchestration and contrasting it with automation: 21:13
Working with customers to orchestrate processes: 26:26
Taking a pragmatic approach to using AI and machine learning: 36:44
Contemplating when data sets will be strong enough to give AI reliable judgment: 45:15
Book recommendations: 58:12
What you wish you had known as a young founder: 1:00:55
Connect with us:
Sagi Eliyahu - https://www.linkedin.com/in/eliyahusagi/
Tyler Finn - https://www.linkedin.com/in/tylerhfinn
Akshay Verma - https://www.linkedin.com/in/akshay-verma-esq/ SpotDraft - https://www.linkedin.com/company/spotdraft
SpotDraft is a leading CLM platform that solves your end-to-end contract management issues.
Visit https://www.spotdraft.com to learn more.
Recommended
Transcript
Systems Design vs. Human Needs
00:00:00
Speaker
The systems are designed for the data and not for the people. You need all the information in order to create a contract. You have to have it. That is the data that you need to collect. That's where we already see running in production in our customers with the iPhone door can do so much of the heavy lifting if you do feed it some of the context, some of the policies, and let a person just say what they need.
Spotdraft's Impact on Legal Efficiency
00:00:34
Speaker
This abstract is brought to you by Spotdraft, an end-to-end contract lifecycle management system that helps high-performing legal teams become 10 times more efficient. If you spend hours every week drafting and reviewing contracts, worrying about being blindsided by renewals, or if you just want to streamline your contracting processes, Spotdraft is the right solution for you.
00:00:54
Speaker
From creating and managing templates and workflows, to tracking approvals, e-signing, and reporting via an AI-powered repository, Spotdraft helps you in every stage of your contracting. And because it should work where you work, it integrates with all the tools your business already uses. Spotdraft is the key that unlocks the potential of your legal team. Make your contracting easier today at spotdraft.com.
Tech Collaboration in Legal Teams
00:01:22
Speaker
How can legal and business teams work better together? What's the role of technology? What's it like to build a tech company that serves legal professionals and where is the industry headed? Today, we're joined by Sagi Eliau, founder and CEO of Tonkin, which simplifies complex enterprise level business challenges with a particular focus on legal process orchestration and intake. And we're going to get his definition of both of those terms.
00:01:50
Speaker
This is also a special episode because it's the first time that I'm joined by a co-host, my friend and colleague now at Spotdraft, Akshay Verma, who is our COO. He's got a background in LegalOps himself as the former head of LegalOps at both Coinbase and at Meta. Akshay has also been on the abstract once before, back in season one, prior to joining Spotdraft.
Sports and Business Strategy Insights
00:02:16
Speaker
Thank you both for joining me today.
00:02:18
Speaker
Oh, thank you for having me. It's a pleasure. I get to be a co-host and that changes the entire approach here. Now it's two against one on the other side.
00:02:30
Speaker
Well, two on one. Okay. Sports. That's a good place to start. Did you like that transition auction? Very nice. I cheated up for you. Something that all three of us have in common actually is that we are all 49ers fans. I think the two of you might be even slightly more dedicated than, than I am. How did you follow the draft? How do you think they're positioned for, for next season?
00:02:54
Speaker
Well, to be honest, actually is the true number one fan. I'm a newer fan, although, as you can tell from my professional life, I'm all in when I'm into something. So definitely became like a religious friend in the last probably five years. Yeah, so draft, you know, look, honestly, it's end of quarter for us. So I wasn't as in tune as I usually is, but I was mostly worried about
00:03:21
Speaker
them not trading away are good friends in the wide receivers. But other than that, it looked like it was good. But I'm actually curious what you think, actually.
00:03:32
Speaker
Yeah. Well, you know, it's interesting. Like we actually, we certainly connected over legal ops and legal technology stuff, but like, I don't know if it was at the conference last year, but like I was really impressed at someone who came to this country so late in his life to be such a diehard American football
Sagi Eliau's Immigration Story
00:03:51
Speaker
fan. Like American football is not a global sport by any stretch of the imagination. So I do, I was, I was just really impressed with his depth of knowledge. So yeah, we connected over that. In fact,
00:04:02
Speaker
That was on a Thursday. There was a Thursday night game that night. It was Niners Giants and the conference had set up the ballroom so that we could stream the game. And so there was probably about 15, 20 of us after the conference was done. They got to watch the game and we spent some time together.
00:04:17
Speaker
Yeah, like I was not. I wasn't terribly concerned. I didn't think they were going to trade anybody. It's just this group is too good to keep together for one more year when financially you don't have to trade anyone away. It's a business like we're going to be talking about business in a minute here. Like the NFL is a business. Teams are a business. You combine that too much with emotions and you get yourself into trouble. So but that being said, I didn't think it made any business sets to trade them away. But as you look at the draft,
00:04:43
Speaker
And I think there's so many analogies here for us. They're both going to be playing for the present as in like when now they have the roster to do it, but also simultaneously prepare for the future when key pieces of that roster will be gone. It's not a question of if, but when they'll be gone and how do you replace those pieces? How do you keep continuity? What are the key pieces that you're not willing to let go of? I mean, those are all business decisions.
00:05:09
Speaker
that the front office make with our head coach. And I think there's so many analogies to that in starting a business, building a business, running a business. So I will always say that sports are like the ultimate analogy for life. And if you don't like sports, fine. I think you can still understand the analogies.
00:05:29
Speaker
No, absolutely. I completely agree. And I think probably on the personal level, the pain of last year is also something that you cannot relate to. Yeah. You lose a big deal either to a competitor or something happens. You're right there. You're right at the finish line. We've all felt that way too many times in my career. And yeah, it hurts. But then you learn from it. And again, I'll draw the analogy.
00:05:57
Speaker
They took some of the learning lessons from the Super Bowl loss, and that's part of how they drafted the players that they drafted. Not to get too technical, but our receivers could not get off of press coverage. They couldn't get open in the Super Bowl. That was hard. Now they've got some guys who can maybe do that. So we'll see if it helps.
00:06:16
Speaker
Going back to something that you alluded to earlier, Akshay, that the two of you also have in common is the sort of immigrant experience. When you were saying you were impressed that Psyge picked up becoming a huge sports fan, American football fan.
00:06:31
Speaker
later in in life and I'd really like to explore how that's shaped how you think about doing business how you think about building a business sort of from from nothing in the in the case of Tonkin so you why don't we start can you tell us your story tell us why you decided to come here and and how you decided to start a business and
00:06:53
Speaker
Yeah, well, I think, honestly, maybe to tie even back to the last point, real quick, I think the, you know, the reason I got so much into American football, by the way, is partially because it's so much similar to business, I felt so I'm I grew up actually, so I grew up in Israel, right? And Israel, you know, we follow what we call the real football. You know, you know what American
00:07:19
Speaker
it would call soccer, which I loved. And all my life, I was a very diehard sports fan. I would be 10-year-old, go to Haifa, which is a city up north, which is an hour away from where I grew up, but was the biggest city in proximity. And I was just a diehard fan.
Immigrant Resourcefulness in Entrepreneurship
00:07:42
Speaker
I would go every weekend to the field and
00:07:44
Speaker
and watch the game with the loudest group of people. And yeah, when I moved here, I kind of learned the game of American football. And what I really was kind of drawn to is the combination of strategy and tactic and
00:08:00
Speaker
teamwork, but also individual genius talent. And that to me is like the kind of the perfect combination of what you need to do. And also like how well adjusted it is to be able to put ads in, which is just one that you know, you appreciate.
00:08:19
Speaker
But no, back to your main question. Yeah, so I moved here, honestly, funny enough, on a whim, actually. So I was in my previous company. I was part of a small, very small startup that got acquired by an American company. They wanted to start an R&D center in Israel after they acquired us. And I was running engineering, so I had the opportunity to build an R&D site in Israel. So I grew the team there for a few hundred people, and then I moved
00:08:46
Speaker
They offered me to move here to the Bay Area and I sat on a wheel because I was actually here every five weeks or so. But then the CEO and I was with the executive staff back then, but I was only managing the Israeli team. So they say, why don't you come here and take over?
00:09:02
Speaker
a bigger role and kind of be part of the official, and it was a public company, so it was a good opportunity. And I literally came with a suitcase. I did not think I would stay here. It was a one-year type of thing, what I told myself. But I ended up, after a year, living and starting Tonkin. And we celebrated nine years, two weeks ago.
00:09:28
Speaker
So it's been a, it's been a, it's been a roller coaster and a great ride. But I think, uh, I always knew that I'm in a way, I always knew I'm going to end up here in some way, uh, or from, I grew up on entrepreneurship, you know, whoever worked with Israelis tech industry, you know, it's all like the startup nation type of thing. It's our idols are people that kind of build companies and, and in the tech industry. And, um,
00:09:58
Speaker
Silicon Valley was like, you know, the number one place to be. So honestly, if you ask, you know, a young kid version of me, this would be a sort of like a dream versus like something that would be practical. But as I as I got older and got into the workforce, it, it was almost like inevitable. I my mom had had tried to fight off. Yeah. But for me, it was kind of obvious.
00:10:27
Speaker
Did you feel like as you were starting the business, and I'm also curious to hear from Akshay on this as well, just being in the Bay Area, being in tech, did you feel like your immigrant experience was an advantage or did it create challenges? Was there a community that you felt that you could lean on, all of those things? Tell us a little bit about how that informed your journey.
00:10:51
Speaker
Well, I think first of all, I do think there's a lot of the Venn diagram between an immigrant and entrepreneur actually is pretty high. You know, you sort of like you have to be resourceful. You're coming into a space that that you bring different perspective and experience to by definition.
00:11:10
Speaker
But in a way, you're trying to fit in, but in your own way, it's a weird, basic parameters in which you operate in. And in the same way, though, you strive on being able to do that matching. So finding what are you very good at, and what is the ecosystem needs, and how those match together. And by the way, and you're aiming at something
00:11:39
Speaker
bigger than what you had. And it doesn't mean that you came from a bad place is just might have been in a great place already. But your inspirations for one reason or another are bigger than where you came from, which kind of caused you. And of course, you know, there are people that immigrate from unfortunate reasons. But of course, obviously, I'm focusing on immigration by choice kind of thing. And then when you think about the
00:12:04
Speaker
the ecosystem or the support system to your point and whether people are there. I think this is something that is very special for the Silicon Valley. It was very clear even before the pandemic. I think it's coming back now, especially in the last year I've been feeling it. But yeah, I've been able to connect with folks. For me personally, there's a lot of Israelis that move to the Valley and they always kind of
00:12:28
Speaker
there to help. And as you know, I'm trying to pay back the favors now to wherever I can. But honestly, you put enough people that are trying to make big things happen in the same place, you know, they're just gonna push each other. So I think this is very special, very special thing about here. But I think most of those hubs, that's why they create the the Renaissance, you know, of the old is like, they're all gravitating into the same place, because
00:12:55
Speaker
people that are gonna want the same thing. And so he's trying to kind of create this this community. Uh huh. It's I mean, it's been that way for 40 plus years. I'm coming up on my 40th year in this country. So I came here much earlier in my life. So I'm a little bit I'd say I'm one degree removed from like the direct pick up your life and move experience because I came as a kid. And that was my parents experience more than it was my me even though I was an immigrant. But it was like that back then. And my dad
00:13:22
Speaker
moved to this country with nothing, like literally nothing except his scholarship for his master's. And then he got a job at Intel in 1986, where Silicon Valley, the H1B program was exploding at that time in the United States. And so he got a job at Intel and that's what the feeling was like here in the Valley back then, 1986, almost 40 years ago.
00:13:46
Speaker
And there were communities of engineers, mainly, from different parts of the world. My parents had friends who were Israeli back then. They had friends who were Pakistani, certainly Indian, Korean, Japanese, you name it. The best engineers in the world were coming here because the companies were paying for them to come, study, move, do all those things. And there was just this incredible energy. It's never really left. It's changed.
00:14:14
Speaker
But it hasn't really left. And that's where your point around when you have a locus of energy and focus like that in this area, really cool things are going to happen. There's plenty of downside too. We're not going to get into that in this podcast. But there's some really, really cool shit that happens. And I don't think that's left.
Tonkin's Evolution in Legal Processes
00:14:37
Speaker
On the immigrant side, I do think it's an advantage. There's disadvantages.
00:14:42
Speaker
I'm an optimist. I'll always look at the fact that you're almost always, always hungrier. You're always been the underdog. You've always faced adversity in some level. And I don't mean like I had a bad day at school, but I mean systemic adversity. And it doesn't have to be racism. It can be a bunch of other things. But you're always used to pushing against that and fighting it and figuring out ways around it and through it and just kind of relying on those instincts to nap. I think those are great attributes.
00:15:10
Speaker
for building a company and building a business and growing it. And starting a company isn't easy. When you started Tonkin, did you expect that you would be selling to legal someday? Tell us about how you got going and how you got to where you are now.
00:15:35
Speaker
Short answer, no. I did not. I did not expect it. When someone says I have an idea, you know, there's this sort of like, you know, jokes that, you know, for you can have a thousand ideas, but the, you know, the work that it takes the people, most people on underestimate. And I think young and I was, you know, I was, I was, I was there too, but early stage,
00:16:04
Speaker
uh, entrepreneurs a lot of times really put a lot of weight on, on, on the idea. And, um, and what I learned is the idea does not matter, but the point of view does. And so I think there's the, if I need to kind of like reframe the concept of, I have an idea, the time by the time it will actually hit the market, it would have changed 10 times. Right.
00:16:29
Speaker
But one thing that is critical for every company that you've seen successful is like, what is the point of view that they have on a problem? And that I found I'm very proud of that, you know, through our nine years now and 10 years, if you kind of count the time before I actually started the company, that I would start thinking about these problems. The point of view did not change. In fact, by every year that passes, I'm more and more confident about the point of view.
00:16:57
Speaker
What is actually though, who are we serving? What are we serving them with? Have changed and refined honestly, and this is not, this is kind of like maybe saying the obvious is 90% of the challenge is timing. What are you coming with and where's the market at? And so when we started talking really the main concept
00:17:23
Speaker
and point of view we had was business processes are not about data, they're about people.
Human-Centric Business Processes
00:17:31
Speaker
But 100% of software at that time and close to 100% of it still today is mainly focused on data. And this is not to say that the data is not important. Data is for a good reason getting a lot of, if not all the attention. But my feeling was a lot of the work
00:17:54
Speaker
within big organization happens in that people layer in where people are need to coordinate with other people and across silos and inside the company and outside the company. And if you run this fast forward a little bit, a lot of the pain that people experience feeling like they're wasting time, feeling like they're not working on important things is when they waste time
00:18:25
Speaker
back and forth, you know, miscommunication, inability, feeling like the politics and the red tape and the back office work is taking a lot of their, you know, sort of like their ability to make an impact. And so when we started the company, we knew that
00:18:43
Speaker
That gap is something that technology can help with. If you if you look at it from a slightly different point, this can we can talk about the orchestration piece in a second. But the legal is actually something that in the last, I want to say, probably five years, we started to focus on. And by the way, recently, also like two years ago, also started to focus also on procurement, which is a lot in common, actually more than more in common than I think
00:19:11
Speaker
both department things. First of all, both hate each other, but they also share a lot of other things. But I just felt like about five years ago is that legal so much
00:19:27
Speaker
centric to the running operation of a company. But also a lot of its impact is around working with different people that comes from different silos that have their own priority, you know, on dependencies. And so in many ways, that human centric process was was a lot more obvious in legal. And that's kind of like how we
00:19:51
Speaker
how we start to work with legal. And then since we obviously created a lot more depth in our understanding of that market, and then the market understanding of our value. But this is not when I started the company, I didn't say, oh, I didn't know anything about legal to be frank.
00:20:08
Speaker
Right it was you know maybe as a user of it definitely when i started the company you interact with. With a lot of legal you know processes and compliance stuff but as a solution or problem area it was mostly. Hey this is actually sounds like what i felt is missing.
00:20:27
Speaker
is heightened in this area of enterprises. And since then, again, five years ago, that's quite a lot of experience by now. But back then, it was completely going filled, and we immediately saw that we were making a difference.
00:20:41
Speaker
I think that's refreshing, probably, for a lot of our listeners to hear as a starting viewpoint, which is to say they often speak to technology vendors. And I would throw CLMs, systems, or solutions in there as well. And it's all about the tech, as opposed to about the people and process and maybe policies, et cetera, that are actually going to enable the tech to work or enable you to extract that key insight from the great data set that you're now going to have.
Orchestration vs. Automation in Legal Tech
00:21:10
Speaker
Maybe that's orchestration, but define process orchestration for us. And I also want you to, you corrected me when we were preparing for this and you said, you view process orchestration as being different than process automation. Tell us a little bit about how those two things connote something different or are really substantively different.
00:21:32
Speaker
Yeah, so first thing first, orchestration is something I did, I have been using since nine years ago. And as much as it is still not fully understood, nine years ago, no one understood what I want from their dear life, including the music community or whatever you want. It was like, what are you saying, orchestration? What do you mean? Like, what does it mean? But it is, it is to me, the idea of an orchestra where you have different instruments.
00:21:59
Speaker
And they have different roles. And they can sound really good by themselves. But if you try to put them together, you need something to actually align them so the music, you know, work. And the music sheet might be the policy for the process.
00:22:17
Speaker
But it's still going to sound, choose your favorite curse word, unless there's an orchestrator, again, in a big orchestra. So this is kind of where the idea for the phrase, or I should say, embracing that phrase as the main phrase came from. But it's a similar idea. So when you think about automation, to me, a lot of people, when they think about process automation, they really just talk about process digitalization. So just making it digital.
00:22:46
Speaker
versus being manual. That is not actual automation. It is a beginning of automation. It's what enables you to automate. But automating is really taking something manual and making it not manual. So basically delegating the work of any task to a machine, that in my book would be automation. That would be automating it.
00:23:09
Speaker
So if you think about a factory floor, for example, that used to be 100% tables of manual people, if you take one table, if someone put caps on a bottle, and then you build a machine that does that, you automated that task.
00:23:28
Speaker
Orchestration is not any of those machines, it's the assembly line. Is that essentially to say, how do I move the entire process along through those different deep technologies or big machine in a factory?
00:23:43
Speaker
So the end-to-end of it makes sense, sings together, the music play. And that's why, for example, when I think about CLMs, if we kind of focus on the legal area for instance, the CLMs and the e-billings and the e-discovery, all that stuff, that's not never what we want to do. Now, do we handle contracts sometimes?
00:24:05
Speaker
or a process that relates to contracts. Of course, that's what legal is, but are we the one that would actually do the red line for a contract or handle the life cycle of the contract itself? Not necessarily. What we're trying to be is essentially sit on top and say, where do requests even come in for a contract? What happened when the contract completes?
00:24:30
Speaker
You know, and how do you sort of think about that machine in all of its complexity as part of the bigger, wider, not even bigger, just like wider, longer, end-to-end sort of workflow. And so that's why we partner so well, by the way, with a lot of those vendors. But this is also where it's kind of sometimes confusing.
Enhancing Legal Operations Efficiency
00:24:53
Speaker
This is where it leads me to that idea of intake, which is something that also we started to talk about five years ago, essentially when we started to work with legal and became more commonly used now, which is just essentially trying to describe a version of that assembly line in which people requesting something from a different team. So one team requesting something from another team, then there's a handoff process there that is rarely managed.
00:25:22
Speaker
And again, if you think about the factory floor it makes sense.
00:25:25
Speaker
you work so hard to create this perfect machine to do this job, you assume that the people that were there, instead of 10 people now, you only need one person to take it and put it into the machine. But that is still intake. And when you start to scale that, that's the next thing that is gonna break. And so again, when you really think about what assembly line does is moving things from one machine to the next machine, and then from that machine to maybe a person, and then from that person back to that machine,
00:25:53
Speaker
And so everyone has an intake. If legal is dependent on procurement or procurement is dependent on legal and both of them are dependent on IT and so on, each of them is actually experiencing an intake from the other one. But they each have their own priorities and their own policies.
00:26:11
Speaker
And everyone is urgent. So it's like it's really about orchestrating those dependencies and then extracting the most out of those machines and investments that you made and make it all kind of things together.
00:26:24
Speaker
I love that visualization. Actually you've been in the legal ops space for a while and really seen it grown up and have also done it at very large organizations that have both used outside tech and also built a lot internally. Do you feel like this concept of orchestration is widely understood or accepted or does that still need to be embraced?
00:26:49
Speaker
No, I think the bar is so low for improvement, which is why legal technology is such a fun place to play in. Sugi mentioned that they kind of fell into legal a little bit. I think my take on that is it was actually the lowest bar to overcome in terms of sophistication and need and pain. And those are often the easiest problems to solve up front. I also think it has
00:27:16
Speaker
the largest opportunity for growth in all of these areas. Facebook would have benefited significantly from this kind of thing because it was a massive department. By the time I left, Facebook was 2,000 people in the legal department.
00:27:31
Speaker
all running around doing different things and not talking to each other and different systems and homegrown stuff and on-prem and some in the cloud and what does this do? And there was even like, oh, I didn't even know we had this. When you get that big, it's not even does the left hand talk to the right hand, the left hand does not know that the right hand exists. It gets to be that kind of a situation. So I think conceptually, I don't think I've heard you talk about the word orchestration before in our conversations.
00:28:01
Speaker
That to me, it's a really, really relevant visual for what Tonkin does and the benefit of this. But where it hits home for me is all these different teams have different policies and procedures on how they need things done. Many of them are actually incongruent, but they don't know it.
00:28:22
Speaker
And imagine if you could tie all of that together. I don't know if seamless is the right word or not, but a lot better than what it currently is. And I know we're talking in the abstract. I'd love to hear a specific example or use case where you have worked with a customer or a set of tools to tie these pieces together to create the orchestra.
AI's Role in Legal Orchestration
00:28:44
Speaker
Yeah, absolutely. Well, you mentioned the challenges you've seen in Meta. And funny enough,
00:28:51
Speaker
we saw almost immediately how much more early adapters to this type of technology, big enterprises are, which is usually the opposite, right? And so even from being very early stage and still today is our pride, is just the massive logos that we have, right? Anyone can go to our website and see the flashiness of them, but it's really like,
00:29:13
Speaker
The biggest tech companies, the biggest financial companies, even government contracts, it's really the bigger they are, the more obvious it is. To your question about how much the market is ready for this, the more obvious it is. That's not to say that we can't help smaller companies, but smaller companies usually need to be more strategic thinkers
00:29:37
Speaker
to actually say, oh, we should invest in this now, versus when it's too late, or not too late, but when it, so like, you know, you already kind of felt the pain. I think the most concrete example to me does come back to that concept of an intake. When you think about the simplest way for a legal team to start working with, you know, creating their policies and start working with departments, is that usually they create some sort of a portal
00:30:04
Speaker
Right. And in that portal, they'll put some links together. And at the beginning, you know, when they're smaller, maybe it's fine. But then, you know, things start to get complicated. So they start investing those machines. So they'll buy, you know, spodraft, or they'll buy, you know, a knee-billing system. They'll buy this or that. And now they want to funnel people into their technology.
00:30:30
Speaker
What do you do? So, you know, you put another page in the portal and say, no, click on that link when you need this thing, but click on that link when you need this other thing. And then what happened if there are stakeholders within that process that are not the ones starting the process? But, you know, you need, in certain cases, actually, FPNA to say, yep, no, we approve that. You can go ahead and do your part of it.
00:30:52
Speaker
or you wanna get a new law firm to work with you and now you need to use other approvals or such. So you're in a position where really the process itself still happens over email and still happen over maybe Slack and Microsoft Teams but really truly over email. And it's not managed anywhere. And so when you think about the matter management,
00:31:17
Speaker
tooling that exists for legal. They also are very much specific into the world of, I need to build hours for this, versus the fact that, wait, there's actually so much back and forth that happens. And whether this even needs to be a matter, maybe just an FAQ, maybe someone just asking a question, hey, should I even talk to legal about this? Or do I need legal advice? What's even the starting point here?
00:31:43
Speaker
So this is probably the most common way for us to start is essentially sits on top of all those different intake sources, places where people can come and talk to legal and either consolidate them. So some companies decide, I actually want them to only start from one front door. And that's where we can come in with some of our cool AI technology. But more often they're not.
00:32:10
Speaker
That is just not as realistic because as much as you put a pretty door, people are still going to email legalatcompany.com. People are still going to send physical letters, which is another big use from a mailroom perspective.
00:32:28
Speaker
So Tonkin basically can connect to your legal inbox, can connect to the SharePoint forms that you have, or replace the SharePoint form with a Tonkin form, can connect to the mail room, maybe the G Drive or SharePoint, whatever the actual file's in, and then do the analysis of what is this request and where does it need to go.
00:32:49
Speaker
And then say, oh, that's actually a contract request. OK, let's kick off a CLM process. Oh, this is actually an FAQ request. I don't even want to create a matter for that. Let's let me try to pull it from a policy, just answer it and bring it back. Funny enough, you know, that's already exists, that concept. When you think about customer support, you have two, one, two, two, three, you know, those type of things. Legal is a customer support function. Don't tell them that.
00:33:17
Speaker
don't tell me. But actually, it is important for them to know that they have that part of their job, but they are not staffed for it. Yeah, right. They're not staffed for it from a research field, but also not from a methodologies perspective.
00:33:31
Speaker
So now they're trying to do their work, but 50% of their time sometimes, even if you want to be kind of optimistic, 15% or 20% of your time, you're spending on basically doing a 2-1 support. No sane person that has an engineering team would make their engineering team be a 2-1 support.
00:33:51
Speaker
They would never do any real job. But that's essentially what happens for LIGO. There are the tier one support and the tier three and the experts and everything in between. And so it doesn't make any sense. And so this is probably the most common place where we start. And then you can get a lot more complicated than that because the downstream orchestration is also real. So you finish, you sign the contract in the CLM. What happens then?
00:34:17
Speaker
Literally what happens from a company policy what do you want to happen if it's a if it's a commercial contract what you want to happen if it's an employee contract what you want to happen if it's a litigation problem what you want that like that can actually yield different downstream effect and we do a lot of those two but those are
00:34:36
Speaker
harder to imagine for many folks versus the customer support example is much easier to grasp. And that's usually where we start from. The front end. So think of it like the old school operator, the switchboard for telephones, right?
00:34:53
Speaker
they're moving the pieces left and right and connecting the dots. Exactly. And then that's where you start. And then if you are truly connected there because we are integrating with everything, we are now, let's say, again, we took that request, put it in a CLM. We at that point handed off there. But the point of orchestrator and the difference between an integrator and an orchestrator is that we're shepherding this in the background until it's done or as things
00:35:20
Speaker
change because you might say like if this is stuck on a certain step in the CLM for more than two weeks or whatever the business logic is, maybe you want to pay back
00:35:33
Speaker
whoever started it, maybe you want to escalate it, maybe you want to create a ticket somewhere, you know, whatever it is. And again, they have those policies. Every company has those policies, but they're on paper in a war dock, you know, like 60 pages that no one really knows what to do about them, which again comes back to the difference between orchestration and automation. This is about what is the logic
00:35:59
Speaker
versus sort of like, what are the tasks to automate? We very, very clearly tell some customers, like, look, I don't actually think what you need is talking. I think what you need first is to have, first of all, a policy of what you want to do, and then you should have some basic tooling for you to do your job, and then
00:36:24
Speaker
Now let's talk about how your job is integrated and connected to the rest of the organization and that's when you need the orchestration. We've experienced that too. I think even more so today we think that making sure that customers are ready for technology or ready for a system is maybe the most important step.
00:36:44
Speaker
You brought up the word AI or the term AI, the letters AI. And we can't have you on the podcast without talking about that. I think that Akshay and I have, I don't think I'd call it a contrarian view, but I think we have a pragmatic view around what we think that maybe like more traditional machine learning tech that we use can do today and what it can do well.
00:37:09
Speaker
what generative can do today and how it can be helpful, but probably not a full replacement for thinking, judging,
00:37:20
Speaker
human beings, we need to checking outputs is super important and try to design workflows and systems to prompt that sort of check or judgment. I'm curious how you think about AI and in the context of this orchestration, because I'm sure it could be very helpful. But you also probably don't want to say take all of your company's policies and put them into a system and let the
00:37:46
Speaker
let the AI write the whole orchestration process. Yeah, how are you thinking about this?
00:37:52
Speaker
Yeah, from the way you describe it, we might actually be very aligned on the way we think about it. I also think about it in a very, I think, programmatic way. To me, this is just the latest iteration of technology. I think it is, can do incredible things. But I think, you know, internet was incredible too. And, you know, I think, you know, iPhone was incredible too. And when I think about, obviously, AI, everyone are like now, but like, everyone,
00:38:19
Speaker
that has been in the industry, know that AI has been in this for like 40 years now. It's really just that the large language models have made their big breakthrough in the last, what is it now? A couple of years now or even? A year, about a year and a half. A year and a half. To me, when that's, you know, and this might be like a very, my tech friends are probably, you know, gonna laugh at that, but I see it as an iteration of almost like regular expressions like regex.
00:38:48
Speaker
in a very dumbed down way. It is orders of magnitude stronger, but from a utility perspective, it is the ability to really understand and manipulate text, especially specifically generative. Obviously, when you apply it on video and audio, it creates some other craziness. But specifically for our world, the biggest leverage here is better understanding people.
00:39:17
Speaker
And this is, I think, this is how I think about it as a tool. There is, and actually perfectly aligned with, I've never would imagine that that's how we gonna get there. I'm very honest, but this is essentially kind of like the gap, I think, I felt again, nine years ago, where I felt that the systems are designed for the data and not for the people. And I mean by that, you wanna have it,
00:39:45
Speaker
You need all the information in order to create a contract. You have to have it. That is the data that you need to do, that you need to collect. So how do we design systems today? We design them as a form.
00:39:59
Speaker
Here are the fields that I need. You, person, transpose your thought into those fields so I can run the process. Where LLMs, and that's where we already see running in production in our customers with the iPhone door, can do so much of the heavy lifting.
00:40:17
Speaker
if you do feed it some of the context, some of the policies and then let a person, by the way, not even knowing because they're just sending an email, but let a person just say what they need. Hey, I have this issue and I have this problem and I can now better than before, but you could have done it before too. We had something we call training set. So we kind of did it part of it, but it's like, again, it was orders of magnitude.
00:40:45
Speaker
worse than what LLM can do, which is being able to say, oh, what you want to do is this. And wait a second. This thing requires 10 fields. I as the LLM can actually pre-fill five of them already just by what you told me. And now I can ask you a couple more questions and then kick off the processing sports draft because I now already have all the information I needed.
00:41:09
Speaker
in like a two second thing because it makes some of those assumptions. Of course, the person needs to review. And so I think to your second part of the guardrails, I think about it in two ways. I think about it in no one, like literally when you go to Starbucks and you order coffee, they will repeat to you.
00:41:32
Speaker
what they heard. Oh, so you want the tolate with almond milk? No, not almond milk, oat milk. Oh, okay. I get it. That's fine. So no one actually trusts even people to just go and do something. So you got to have that review mechanism. And if you don't have it, you're just bad at your job. It has nothing to do with the technology.
00:41:55
Speaker
Now, the second part of it though is, and this is something that this is a very specific way of how we think about orchestration, but even automation is that, again, if you compare everything to people, it's easier. Even if you hire an intern and you give them a job,
00:42:13
Speaker
you're not going to give them access to the kingdom, to an intern. You're going to create the playground for them to be creative, to do the job, and ideally it will be actually hands-off, but within the scope of that,
00:42:30
Speaker
of that task, right? You know, I don't know, something dumb, but like, I don't know, go scrape LinkedIn and put it in a spreadsheet, right? Whatever. Like, you're gonna give them access to that, but you're not gonna give them access to the CEO necessarily LinkedIn, right? That's just like too much of a risk.
00:42:46
Speaker
It's just something that, again, if you're smart about it, if you're a small company, maybe you'll do it. But if you actually need to be compliant and all that stuff, you won't. And so I think the way we think about where AI can help more than what I already discussed, which is understanding people, there are a lot of
00:43:05
Speaker
gray areas and fat in between steps, it's an overkill to actually codify them. But the main structure of what needs to happen when the hard-coded policies, those needs to be coded. I don't think those make sense.
00:43:24
Speaker
Not everyone has billions of billions of billions of recorded video hours like Tesla has for Autiguide that can go and I don't know how much of you guys are in tune with their stuff where they moved from having partial neural net and partial hard coded stuff to full neural net. But I really think that only can happen.
00:43:46
Speaker
when the structure of what is good and what is bad is very clear. And I think most companies, there's just not enough data for this to be true. There's not enough data to say, this is good action to take, this is bad action without having some of those ground rules implemented there. So I think for many, many years, I think at least internally for how processes happen, it will be a combination of the two.
00:44:16
Speaker
think about it as, again, back to my factory floor example, as a new cool machine or a new cool technology that will allow machines to work much better than they worked before versus a fundamental shift of
00:44:33
Speaker
of how things would be done. And therefore, that's where people are like, oh, our jobs are going to be taken away, blah, blah, blah, blah, blah. People need to learn history a little bit better. But it repeats itself every time, literally every time. Every five years, there's something that people, jobs are going to take away. And then really, in the bottom line, jobs are
00:44:57
Speaker
roles might change. And of course, some people will lose their job, but it's not because of the technologies, because maybe they didn't.
00:45:04
Speaker
They took it for granted and they didn't understand that look, change is inevitable and you need to adjust to it and adapt to it. It's just one-on-one Darwinism. I think people do a lot of things out of fear and lawyers especially where this is a big topic of conversation. I see it on LinkedIn like every other day. OAI is going to replace lawyers. The hell it is.
00:45:30
Speaker
It might replace the lawyers who aren't actually doing lawyering work. And there's a lot of that. That's not their fault. That's the legal profession. That's legal departments. That's law firms. Those are the culprits there. So how you do things will change. You said something which is really interesting around you use the analogy of driving, autonomous driving. And it requires immense amount of data.
00:45:57
Speaker
But that's what judgment is. Human judgment is not inherent. It's based on data. I will do this because of my past experience and the thing that's in front of me and all of these different data points. And we say it's good judgment, not because of the decision that's made, but because of the outcome. It's only good judgment if the outcome is good for that situation. It's not because of the decision that was made.
00:46:23
Speaker
So to me, that's where AI is no different. It's a surrogate for human decision making with significant limitations, except that it has massive faults in the outcome right now because it doesn't have enough data.
00:46:38
Speaker
So when do we get to a place where there is enough data in these things? What are your thoughts on that? Well, if we continue, here's where I defer, I think, in my worldview on this. I think some things take long to change, and some things just have to change. So if we take Tesla, for example, and you have all of these training sets into how to ride a road,
00:47:05
Speaker
right, and be autonomous driving. But then the same CEO and founder actually end up, you know, building a different way of transportation. That data would be useful, but it won't be enough. So now you would need to have a different kind of modification on that data set and change it.
00:47:23
Speaker
I think the corporate world, even in B companies, is changing too often, and the policy decisions change much faster than what laws of the road of traffic are.
00:47:40
Speaker
And so it's not even the judgment is, has so much corporate context to hit sometimes that it's like the data set is, is it become a foundation layer? And this is why I think you don't even need to wait for it in a way. I think you can actually extract a lot of value from LLMs today because they know English.
00:48:00
Speaker
but in, you know, or many other language, but like, you know, the focus on English and then the rest of it, you sort of like need to be specific and kind of guide it and prompt it or kind of have additional training, but that even not as calm. So I actually think in one way, so I'm kind of like saying, I'm giving both answers. So in one way, I actually think we have enough data for it to be significantly improve the output of legal teams right now, like literally tomorrow.
00:48:28
Speaker
But I think the difference is, to me, is even just what the real role of such departments are, like LIGO.
00:48:38
Speaker
I don't, you know, when I think about my, my console and even the name, the name console is I don't need them to review an NDA. Yeah. Just a waste of time. I also, I also don't need them to, to review a massive MSA that is red line as I obviously that's, I prefer to do that than having them review an NDA. But I see that.
00:49:03
Speaker
uh you know goes lower and lower where where i do need them is to actually have an advice and a strategic sort of thought that is not that that data doesn't even exist yeah it's just you know it's just someone that would have been with me for the last nine years yeah and knows also kind of like where i want to go as a company and and help me think through the risk main
00:49:29
Speaker
point here is think. So if I want to validate the risk, maybe I can talk to chat GPT right now. And it will have the knowledge to give me answers, but I still need to come up with the questions. If you ask chat GPT, what should I do?
00:49:45
Speaker
Yeah, that would be nonsense. If I ask specific questions like, hey, what happens when this happens and this happens and how likely it is to happen, then it will blurb something that I can learn from. And it might be better, by the way, I have my own thought about when is it better than a search interface versus a chat interface. It's not always true. Sometimes it might be better than a search interface. Sometimes it actually would be worse. But either way, it is a great tool to extract
00:50:15
Speaker
you know, information, but not necessarily, you know, to go about that advice piece. Now, I'm not saying that you can't, right? You can hook it to your brain eventually, you know, and then have that context or hook it to all of the record, all of your calls over the years and all that.
00:50:33
Speaker
On the theory level, it can happen. But on the practical level, I don't know. I don't want to talk to Siri or Alexa on a public space. No matter what.
00:50:47
Speaker
I just don't want to do it. It just doesn't make sense to me. I also don't want it to, I just don't want it to be, I don't want to record every call that I have in the company. You know, we use recording tools and we use it a lot actually, but still there's some, you know, that I just don't want to do it, right? And I think legal would understand it the most, right? There's a reason why retention on data. So like there's a counter counterintuitive here of like where
00:51:14
Speaker
How can you make that work well? But I think it's a positive thing because I think the goal of technology is to take away literally the first technology you take a stone like a monkey take a stone and break a coconut.
00:51:30
Speaker
Why do you do that? What's the value of that? Well, it creates leverage for you. You get more force out of it. But you also don't break your wrist. You take the stuff that are painful and you don't want to do, and you create leverage out of them. And any technology can be on that same scale. And I think what people misunderstand is that you are going to stop doing something.
AI's Impact on Legal Work Dynamics
00:51:53
Speaker
If you want to enjoy the leverage,
00:51:55
Speaker
It means you're going to stop doing something that you did before. What are you going to do with that time? What are you going to do with that energy that you saved? Are you going to break more coconuts than the one that it took you to do before? Maybe you break four, and so now you can have a family. I'm making things up, but you know what I mean? It's a matter of leverage, and so I think a good advice
00:52:16
Speaker
for people listening in, especially in legal, is just to stop with the conversation of whether it's gonna happen or not. It's definitely gonna happen because as a consumer of legal services, I don't want to pay for checking an NDA. I just don't wanna do it. It still happens way too much.
00:52:37
Speaker
Exactly, but that is the transition that people need to do and start instead building the skill set. If they don't feel comfortable, they have the skill set, they do have time to build those skillset, but not if they wait two, three, four more years because then it will be too late. I think part of the problem is people think of their roles often as a set of tasks and not impact.
00:53:00
Speaker
And AI, like any other tool, technology, is going to replace tasks. But that is not your job. That is not your career. That's not your role. Those are things that you do. Exactly. And we are human beings. We have the capacity to learn new skills. We call it upskilling, upleveling, whatever you want to call it. Anyone can do that.
00:53:27
Speaker
And, you know, I mean, this is this to me, this is the crux of legal operations. Anyway, like this is why we exist in the first place is there are all these tasks across the legal department that are far too manual. There are drain on the business. We don't talk about that enough. They're not just a drain on the lawyers. They're not just a drain on the groups. They're not just a drain on the legal department. They're actually a drain on the business.
00:53:51
Speaker
And the sooner you make that connection, you say, hey, we can remove, that's the number one lever, get rid of these tasks. We don't need to do these tasks anymore. Number two, here's an entire set of tasks. And the value chain here is getting higher and higher that we can now automate through whatever tool, technology or process. Then it's like, hey, now we're getting closer to our core strategic work. This is fun.
00:54:16
Speaker
We made our business better and people are happier. That to me is like the paradise land for a legal department and a business. We are far, far away from that still, even with AI that's here right now. I think you said something that I think is actually
00:54:37
Speaker
I want to double click on even further the remove part. To me, it's funny because it's very visual. Example actually in the finance and procurement world, but I think it's applicable here as well in the legal world. I think everyone should think about what is that equivalent. I call it the RFP conundrum.
00:55:00
Speaker
AI, if everyone had eight RFP, people hate to create it, people hate feeling it. If you have an AI that creates an RFP and then an AI that answered the RFP, do you even need an RFP? What is that? What's the point? And I think that's important because for the next immediate future, that's what will happen. You will have an AI that creates an RFP, an AI that
00:55:26
Speaker
that fill the RFP and then an AI that, you know, score the RFP. But then not far after that, someone will come and say, this is all kind of dumb.
00:55:37
Speaker
And so let's have something different that is better. And essentially, this process needs to be removed. And so if you are hanging on to the fact that, well, right now it's pretty hard. I'm making this part up. But if you are like, oh, it's too hard to create an RFP right now. But then that's going to automate it away. And they're like, oh, well, it's pretty hard to validate an RFP.
00:56:04
Speaker
You're looking at the right, the wrong way. Like the question is, is RFP as a concept, is a construct of our limitation, or is it actually create an actual value? And in the RFP example, I believe it is, you want to source from different vendors, you need a way to get all the information. Absolutely. Right. And so it's a construct of, you know, we don't have enough time, we don't have enough, you know, so we're trying to normalize the information. That is, that is,
00:56:34
Speaker
That is a solvable problem. This is not something that is core to any business. It's a way to transform information. What are the equivalent? There are a lot of them. What will remain on the extremes? I think this is something that
00:56:53
Speaker
I didn't come up with this. It's pretty, I think well known, but I really registered to is thinking about things in the extremes. Like if everything is like zero, if everything is a hundred, how does the situation looks like at that point?
00:57:07
Speaker
And that would give you a good idea of the play field. Like if everyone using it for everything, or if no one is using it for nothing in a specific area, what does that look like? And I think you start, but you start imagining, look, that would mean this and this and this, what will mean on my job, if I'm the only one that is not doing it, or I'm the only one that is doing it, and so on.
00:57:35
Speaker
Yeah, it's a very interesting topic. I think it's a privilege for everyone, honestly, being in that place in
AI and the Future of Work and Life
00:57:44
Speaker
time. I think it's exciting, but maybe it's the technologies in me. There's amazing potential.
00:57:52
Speaker
It's not just about the technology. It's not just about AI. It's about, as Akshay was saying, the potential for what your work or what your life could be. This has been such an insightful conversation. I've got a couple of, well, I hope they're fun, questions for you.
00:58:10
Speaker
as we start to wrap up. The first one is something I like to ask most guests, which is I'm a big reader. Is there anything interesting that you've read recently, book or similar that you would recommend to our listeners? I actually mostly read biographies. I find inspiration in
00:58:38
Speaker
People struggles, you know, and how they overcome that. I love, I hate the idea of a overnight success concept. You know, it is, there is zero cases of that. It, you know, 100% of the time.
00:58:53
Speaker
you know, it's either your hard work or your parents hard work or like, you know, it is never like, just like out of thin hair, lottery thing. By the way, if anyone wins the lottery, they usually lose it immediately after. They can appreciate it as well. So I think, so any literally any type of
00:59:15
Speaker
of biography. I just read the Elon Musk biography. I do recommend, I think, that one specifically was a very in-depth one to actually his childhood and other stuff, which again, for me is interesting more than just the recent years, if you actually want to understand it. But you mentioned any other, like in Netflix, they nailed down the doco series thing.
00:59:44
Speaker
from like the Formula One that is like a must-watch, I believe, but yeah, all the way to, you know, they're not doing it for every sport. And I'm still holding my fingers back to the beginning of this conversation that they, last year, that they might've done one for Brock, but we'll see. They still haven't- No, we need Brock to focus on the season, the upcoming season. No, I think last year, I think they already, my bet is they actually already- Filmed it. Filmed that last season.
01:00:17
Speaker
But either way, I recommend the latest Elon Musk one. If you like biographies, there's a great podcast called Founders. I don't know if you've heard it before. It takes the concept of founders very broadly and the host reads biographies and then summarizes them for you and anywhere from
01:00:42
Speaker
45 minutes to an hour and a half, two hours. But the interesting thing is then when you start to listen to more episodes, there's an interplay between them, right? Yeah. It's that's a, that's a good one. Yeah. I'll check it out.
01:00:54
Speaker
A last question for you, Sagi. I have a traditional sort of question that I ask pretty much everyone who comes on. And I'm going to frame it a little bit differently because you're not a lawyer. You're a founder. And maybe this is for the benefit of other hopeful founders, some of whom these days, especially, are former lawyers. You see that all the time.
01:01:18
Speaker
If you could look back on when you first started, Tonkin, what's something that you wish you'd known then that you know now? So I think, you know, we talked about the idea versus the point of view part. I think aside from that,
01:01:32
Speaker
I would say two things. One, it's a known fact, but it is, you got to live through it to fully realize it is that timing concept. It is, you cannot be, if you're starting, if you have a new point of view or an idea, a new idea or something you want to build, you're by definition early. Like if it's a common sense already, then it's not, there's nothing new there. So you're already coming with something new. So by definition, you are early to the market.
01:02:00
Speaker
The biggest important thing is how early are you? Are you a year early?
01:02:06
Speaker
three years early, 10 years early, 20 years early, right? And people talk about market fee, they talk about bunch of things. I find all of them good for some things, but sometimes oversimplifying it to an operational level of what you kind of need to do. People talk about talk to your customers early, but they don't talk about what you need to ask.
01:02:31
Speaker
And so I think to me that concept of if you can try to focus on analyzing by talking to customers, by shipping early and fast stuff, but what you're actually measuring is how far is your vision from the market right now and how long before this is going to be the norm.
01:02:56
Speaker
That would be a very helpful tool in your decision making arsenal is what I think. That's great. Well, Saiki and Akshay, thank you both so much for coming on and producing this first ever co-hosted episode of The Abstract. We want easy on you, right? No, that was fun. I really appreciate it. Yeah, likewise.
01:03:20
Speaker
And to all of our listeners, thanks so much for tuning into this episode and we hope to see you next time.
01:03:39
Speaker
We talk about how LegalOps professionals can leverage technology and where the LegalOps profession is headed. You can also subscribe so you get notified as soon as we post a new episode. And if you liked this one, I'd really love to hear your thoughts, so leave a rating or a comment. If you'd like to reach out to us, our LinkedIn profiles are in the description. See you all next week.