Podcast with Brijesh Anand, Part-2, Regional General Counsel (APAC) at ANDRITZ

[00:00.000 --> 00:12.420] Hello everyone, welcome to another episode of LexDiscuss Podcast, where we interview
[00:12.420 --> 00:16.000] established and seasoned legal professionals.
[00:16.000 --> 00:21.560] You have tuned in to the second part of our conversation with the brilliant Mr. Brijesh
[00:21.560 --> 00:22.560] Arnold.
[00:22.560 --> 00:24.360] I am your host, Abhi.
[00:24.360 --> 00:28.560] Let me start by introducing the man of the hour, Brijesh.
[00:28.560 --> 00:35.280] His career spans over two dynamic decades, leading an indelible mark on the legal landscape,
[00:35.280 --> 00:36.760] from boardrooms to courtrooms.
[00:36.760 --> 00:38.800] He's been there, done that.
[00:38.800 --> 00:44.280] So, buckle up guys, because we are about to unravel the fascinating world of corporate
[00:44.280 --> 00:51.480] laws, decode, strategic maneuvers, and perhaps even sprinkle in a dash of humor.
[00:51.480 --> 00:54.400] So, let's jump right back in.
[00:54.560 --> 00:59.920] My first question for this podcast is, what are the biggest hurdles to wider adoption
[00:59.920 --> 01:06.080] of AI in the legal industry, particularly in large organizations like Android?
[01:06.080 --> 01:09.560] And how to overcome these challenges?
[01:09.560 --> 01:15.300] So Avi, that's a very good point, because while technology is improving, the confidence
[01:15.300 --> 01:20.240] in technology, although improving in some areas, in very niche sectors like the legal
[01:20.240 --> 01:23.040] sector, continues to remain a little shaky.
[01:23.680 --> 01:27.880] I mean, I can give you an example of Android where we have been trying to adopt a lot of
[01:27.880 --> 01:31.480] these AI technologies since last four, five years, seven years in fact.
[01:31.480 --> 01:37.440] Till maybe one, one and a half years back, the technology was just not capable of doing
[01:37.440 --> 01:41.000] what the legal department or lawyers wanted it to do.
[01:41.000 --> 01:45.720] It's only when the large language models came up, it's only when deeper analytics and machine
[01:45.720 --> 01:49.760] learnings improved, we could start benefiting from the use of AI.
[01:49.760 --> 01:54.800] And large organizations like mine and many others started to explore what can be done
[01:54.800 --> 01:55.800] through AI.
[01:55.800 --> 02:00.680] You're right, there are various hurdles still, and those hurdles depend a lot in which milieu
[02:00.680 --> 02:02.960] and which environment you are from.
[02:02.960 --> 02:07.600] For example, a company which is deeply in large contracts, which is dealing with very
[02:07.600 --> 02:13.480] large contracts, companies like contractors, developers, on power plants, on roads, infrastructure,
[02:13.480 --> 02:14.480] project finances.
[02:14.480 --> 02:19.000] There, the issues are very different from, let's say, a retail organization or an internet
[02:19.000 --> 02:20.760] company or a startup.
[02:20.760 --> 02:25.000] So their requirements are different, their expectations from AI are very different.
[02:25.000 --> 02:30.160] While AI is able to, let's say, automate certain things very easily, let's say NDAs, if a company
[02:30.160 --> 02:35.520] requires a large number of NDAs, term sheets, proposals, they are much easily adopted by
[02:35.520 --> 02:37.080] generative AI.
[02:37.080 --> 02:41.400] The same may not be holding true for a large infrastructure contract running into thousands
[02:41.400 --> 02:45.000] of pages with very clever drafting by lawyers.
[02:45.000 --> 02:50.000] So I think whether technology has reached the proportion where it can help that particular
[02:50.000 --> 02:52.080] industry remains a barrier.
[02:52.080 --> 02:56.280] That is one big hurdle in adoption of AI, and this is important because in some sectors
[02:56.280 --> 03:00.600] you will see a quicker adoption of AI, in some you will see AI being adopted a little
[03:00.600 --> 03:03.680] later as in when AI keeps developing in that sense.
[03:03.680 --> 03:07.840] So I don't think it will be a uniform adoption of AI in all sectors.
[03:07.840 --> 03:11.600] Other than that, I think the usual concerns for organizations when it comes to adopting
[03:12.080 --> 03:18.760] cost, availability, efficiency, we have still not reached the criticality of mass, products
[03:18.760 --> 03:23.840] are far and few in between, there are less choices to be given, less people who are trained
[03:23.840 --> 03:28.120] in dealing with these AI because we have a lot of lawyers, a lot of professionals who
[03:28.120 --> 03:32.680] are not so adept in using these technology, many people who really have not been able
[03:32.680 --> 03:35.880] to embrace technology, which will improve as we go forward.
[03:35.880 --> 03:39.680] So that's the other side of the concern and hurdle to adoption of AI.
[03:39.680 --> 03:43.640] Even there are ethical concerns, there is the issue of privacy, there is the issue
[03:43.640 --> 03:48.900] of plagiarism, there is not full confidence in the generation of replies from AI.
[03:48.900 --> 03:53.800] So I think it's a mixed bag on concerns and all of them I think are genuine concerns.
[03:53.800 --> 03:58.600] And personally, I do believe that as we will go forward, as these concerns would be addressed
[03:58.600 --> 04:02.520] and people would embrace technology much more than they are doing right now.
[04:02.520 --> 04:03.520] Well said.
[04:03.520 --> 04:04.520] Well said.
[04:04.520 --> 04:09.520] And that brings me to the next follow up question because, you know, whether it doesn't matter
[04:09.520 --> 04:14.800] how reclusive the legal department is when it comes to adopting the AI.
[04:14.800 --> 04:19.000] Every other department that I'm speaking to, whether it's marketing, whether it's technology,
[04:19.000 --> 04:25.360] they are adopting AI tools left, right and center, and you just raise the point of privacy.
[04:25.360 --> 04:32.160] So as a legal head, what are your concerns about your organizations adopting these AI
[04:32.280 --> 04:35.200] tools and compromising the privacy?
[04:35.200 --> 04:37.560] So privacy remains a big concern.
[04:37.560 --> 04:42.120] What is privacy of your customers, of your organization, of issues, especially legal
[04:42.120 --> 04:43.120] departments.
[04:43.120 --> 04:47.360] You're concerned giving all your problems, all your documents, all your legal strategies
[04:47.360 --> 04:51.800] on a server, on something that may be liable to be hacked.
[04:51.800 --> 04:56.840] Someone from, let's say, a sales department, maybe would be happier sending the details
[04:56.840 --> 05:01.800] of the proposal through an AI, a lawyer, maybe more skeptic, I would say, to put all those
[05:01.800 --> 05:02.800] documents.
[05:02.800 --> 05:06.440] So when I say talk about privacy, there are two distinct areas of privacy.
[05:06.440 --> 05:11.360] One is privacy of the work in general on data as privacy as we understand of lawyers, of
[05:11.360 --> 05:14.640] people that we are talking about, their personal data and so on.
[05:14.640 --> 05:18.800] And the other is the overall issue of data breach, of confidentiality, of your, you know,
[05:18.800 --> 05:23.960] litigation strategy, of your maybe intended purpose of terminating an agreement or writing
[05:23.960 --> 05:24.960] something.
[05:24.960 --> 05:29.120] Those are large concerns and we need to address them with intelligent use of AI.
[05:29.120 --> 05:32.360] I don't think we should stop using AI because of these concerns.
[05:32.360 --> 05:37.920] I think on the issue of privacy and data breach, large companies have really taken a lot of
[05:37.920 --> 05:39.960] steps and that gives a lot of confidence.
[05:39.960 --> 05:45.400] I told you last time, I think that we are, for example, implementing this Microsoft Copilot
[05:45.400 --> 05:48.760] Network, which has access to all our data.
[05:48.760 --> 05:54.760] Everything that our company has in their contracts, there is access to this Copilot, to that data.
[05:54.760 --> 05:59.480] But we are fairly secure with it because we know the systems of privacy, confidentiality,
[05:59.480 --> 06:04.720] data breach with Microsoft are so strong that I am not afraid that what is there in my computer
[06:04.720 --> 06:08.280] would be used by machine learning and would be transferred to someone else in one way
[06:08.280 --> 06:09.280] or the other.
[06:09.280 --> 06:11.200] So those issues are improving a lot.
[06:11.200 --> 06:16.080] I think larger companies are improving that a lot with cloud servers.
[06:16.080 --> 06:19.720] You know, these days everybody is using Amazon Clouds or Google Clouds.
[06:19.720 --> 06:21.520] So you're not talking about smaller companies.
[06:21.520 --> 06:25.900] So these companies are very wary of this fact and they're taking a lot of steps to
[06:25.900 --> 06:27.040] cover this.
[06:27.040 --> 06:33.000] And I believe that remains a problem, particularly for GDPR compliant companies where you have
[06:33.000 --> 06:35.200] other issues of GDPR coming up.
[06:35.200 --> 06:39.240] So a lot of European companies, for example, are a little more wary on the data that they
[06:39.240 --> 06:42.020] are giving, collecting, putting in this system.
[06:42.020 --> 06:44.320] But I think these issues are improving.
[06:44.320 --> 06:48.700] Short answer would be to partner with organizations, with companies where we have a proven track
[06:48.780 --> 06:53.340] record of privacy, who have technological advancements to the extent where data breaches
[06:53.340 --> 06:58.540] can be prevented and use the technology wisely is what my answer would be.
[06:58.540 --> 07:00.460] Very wise words, Rakesh.
[07:00.460 --> 07:03.500] As I asked in the last question, everybody is using AI.
[07:03.500 --> 07:05.020] That's just a matter of fact.
[07:05.020 --> 07:10.940] Now in addition to privacy concerns, do you see any other legal risk as a legal head that
[07:10.940 --> 07:15.380] might be coming up with more and more adoption of AI tools?
[07:15.380 --> 07:20.100] The legalism is a big risk that may come up.
[07:20.100 --> 07:25.500] The other legal risk that I see is if you are using AI to, let's say, review a contract
[07:25.500 --> 07:31.380] document, AI may not be able to only give you the right summary, right excerpts.
[07:31.380 --> 07:35.700] And if you rely too much on the summarization of document, you will lose sight of various
[07:35.700 --> 07:37.100] clauses here and there.
[07:37.100 --> 07:42.940] For example, if I have right now a contract review tool where I wanted to extract limitation
[07:42.940 --> 07:48.140] of liability or a penalty or a confidentiality severability clause.
[07:48.140 --> 07:52.980] One thing is if it is drafted plain and simple, the other is because we are dealing with lawyers
[07:52.980 --> 07:57.700] and legal drafting keeps improving, legal drafting keeps adapting to the changes.
[07:57.700 --> 08:01.740] The contracts will be written in a manner where it will be more and more difficult for
[08:01.740 --> 08:04.020] machines to catch those clauses.
[08:04.020 --> 08:08.780] The words will be changed, the language will be modified, there will be difficult cross
[08:08.780 --> 08:13.100] referencing and something that you could have said earlier in four lines in one clause,
[08:13.100 --> 08:18.060] you would now say it in seven clauses spread across five sections, a combined reading of
[08:18.060 --> 08:21.660] which would mean there is no limitation of liability, for example, but the machine may
[08:21.660 --> 08:22.740] not be able to read that.
[08:22.740 --> 08:25.620] So that is a big problem also that needs to be seen.
[08:25.620 --> 08:29.660] Yeah, that's a very interesting point that you raised last time.
[08:29.660 --> 08:34.060] And, you know, I'm very naive on this front because I haven't, you know, deal with a lot
[08:34.140 --> 08:39.060] of corporate negotiations, but a very naive perspective of contracts would be that, you
[08:39.060 --> 08:41.940] know, they are there to make everything clear.
[08:41.940 --> 08:46.780] They are there for each party to understand what their roles and obligations are so that
[08:46.780 --> 08:49.620] at the end of the line, there is no dispute.
[08:49.620 --> 08:53.780] But the interesting point that you just raised, what is the practical reality?
[08:53.780 --> 08:59.500] Is practical reality more geared towards making things fair and simple for everybody to understand
[08:59.500 --> 09:04.020] or the practical reality is tilting toward the other side of x?
[09:04.020 --> 09:06.020] I don't think there is a distinction between the two.
[09:06.020 --> 09:12.420] I think once you are drafting contracts or any kind of legal documents, the basic premise
[09:12.420 --> 09:15.420] is to make sure everybody is aware of everything.
[09:15.420 --> 09:19.820] So I start with the assumption that people are trying to fool the other parties by clever
[09:19.820 --> 09:20.900] drafting.
[09:20.900 --> 09:25.660] The basic tenet of output contract review and draft would be that people are trying
[09:25.660 --> 09:26.660] to make a good draft.
[09:26.820 --> 09:29.940] They are trying to make obligations clear, like you said, that they're trying to make
[09:29.940 --> 09:31.740] things very ocean.
[09:31.740 --> 09:37.300] The problem which we see is not so much of a problem of trying to dupe somebody or trying
[09:37.300 --> 09:41.020] to sneak in a clause where the other party won't be able to see.
[09:41.020 --> 09:45.860] But ultimately, we must respect the fact that a contract is ultimately a compromise between
[09:45.860 --> 09:47.860] what the two parties wanted.
[09:47.860 --> 09:52.580] So both sides negotiate hard and try to get it stuff which will help them try to get in
[09:52.580 --> 09:57.540] clauses which would prevent their risks, which would not escalate issues for them, which
[09:57.540 --> 10:01.220] would curtail costs for them, which would make things simpler for them.
[10:01.220 --> 10:05.220] So contract is not a literary piece of work where you want full clarity.
[10:05.220 --> 10:09.700] Ultimately, it's a compromise between both the parties of what they have agreed in a
[10:09.700 --> 10:11.980] heavily negotiated contract.
[10:11.980 --> 10:16.900] And many times I would say that if you have good lawyers on both sides in a contract negotiation,
[10:16.900 --> 10:20.120] you would try and get your perspective one way or the other.
[10:20.120 --> 10:21.760] So that is where the problem comes.
[10:21.760 --> 10:27.680] So if I, for example, say that I want a confidentiality clause, for example, the easier way is the
[10:27.680 --> 10:28.680] other side agrees.
[10:28.680 --> 10:33.040] We both agree on a simple lane vanilla language confidentiality clause that will be confidential
[10:33.040 --> 10:34.040] for 10 years.
[10:34.040 --> 10:35.320] These are the exceptions.
[10:35.320 --> 10:37.160] This is what my jumped relief will be.
[10:37.160 --> 10:39.520] This is the penalty, so on and so forth.
[10:39.520 --> 10:43.480] But if my other side starts saying I will not agree on a confidentiality clause for
[10:43.480 --> 10:48.920] X, Y, and Z reasons, I still try to get whatever best that I want from that clause.
[10:48.920 --> 10:53.560] And the only way to then do it would be clever drafting, to put something here and put another
[10:53.560 --> 10:57.280] thing in another clause and try to get as much as you want, so the other side would
[10:57.280 --> 10:58.520] try the same.
[10:58.520 --> 11:00.960] I think the intent on both sides is very right.
[11:00.960 --> 11:04.080] We start with the premise of making a good contract.
[11:04.080 --> 11:08.480] But ultimately, negotiation means that both parties are trying to get what they want and
[11:08.480 --> 11:11.800] they will try to do it in one way or the other in ethical manner.
[11:11.800 --> 11:15.760] I mean, this is just the clever negotiation, clever drafting, there's nothing wrong with
[11:15.760 --> 11:16.760] that.
[11:17.400 --> 11:22.000] Absolutely, very well explained and obviously everybody on paper stays right away on the
[11:22.000 --> 11:26.480] ground of ethical manners, but yeah, they have every right to use the smart techniques
[11:26.480 --> 11:28.120] and that's fair as well.
[11:28.120 --> 11:29.920] Very well explained, Rajesh.
[11:29.920 --> 11:32.400] Now moving on to the next question.
[11:32.400 --> 11:38.320] Given your reputation as a problem solver, you think Khandraj should give you a legal
[11:38.320 --> 11:41.200] superhero costume?
[11:41.200 --> 11:47.160] I think my job is to effectively advise my company and solve their problems in whichever
[11:47.160 --> 11:52.360] best manner the legal department can and support the business in its continuous growth.
[11:52.360 --> 11:58.080] The onus on general counsels these days is much more than just being a lawyer.
[11:58.080 --> 12:04.120] They are also business enablers, they are also risk mitigation officers, they are also
[12:04.120 --> 12:06.780] ethical officers, compliance officers.
[12:06.780 --> 12:11.900] So I think the legal departments and the general counsels these days are entrusted
[12:11.900 --> 12:14.580] with a lot of responsibilities.
[12:14.580 --> 12:19.540] They are kind of a guide philosopher to the board on a lot of issues, whether it be legal,
[12:19.540 --> 12:23.160] whether it be related to compliance or business strategy.
[12:23.160 --> 12:28.420] And I think they are a hero or not, or a heroine or not, based on how much are they able to
[12:28.420 --> 12:34.500] give this advice, whether they give it wearing a shirt and pants or any ordinary clothes
[12:34.500 --> 12:37.220] or a cape is then something which one needs to see.
[12:37.220 --> 12:42.580] But I think the legal counsels should be superhero enough to solve the problems of their business
[12:42.580 --> 12:47.020] and to advise their board, to advise their management in a manner to navigate best in
[12:47.020 --> 12:49.660] the complex legal milieu in which we live.
[12:49.660 --> 12:50.660] Absolutely.
[12:50.660 --> 12:56.180] I believe that the legal role is not appreciated as it should be, but well said.
[12:56.180 --> 13:00.620] No, I just, sorry, just to say, I think the role has changed a lot.
[13:00.620 --> 13:01.620] Right.
[13:01.740 --> 13:04.980] Today where we are seeing the role is much more appreciated.
[13:04.980 --> 13:10.100] I think an effective legal department is very much valued by the management.
[13:10.100 --> 13:15.180] You have umpteen examples where you have now general counsels sitting on the board of companies
[13:15.180 --> 13:20.280] and with a heavy voice, with a vote of, I would not say veto, but a significant voice
[13:20.280 --> 13:22.740] at the table, a seat at the table.
[13:22.740 --> 13:27.100] But ultimately we should not forget that business would run the business and legal is a support
[13:27.100 --> 13:28.100] function.
[13:28.100 --> 13:32.500] There would obviously be a compromise and there would obviously be business wanting
[13:32.500 --> 13:37.420] to take a business decision and legal is there to support them on how to navigate that business
[13:37.420 --> 13:38.700] effectively.
[13:38.700 --> 13:40.340] So I think the role has changed.
[13:40.340 --> 13:42.500] I think there is a big seat at the table.
[13:42.500 --> 13:46.980] I think the opinions of general counsels are valued and things are changing very fast as
[13:46.980 --> 13:48.740] we are going along.
[13:48.740 --> 13:49.740] That's good news, Brijesh.
[13:49.740 --> 13:52.980] I'm happy to hear that this has been your experience.
[13:53.020 --> 13:58.540] Now moving on to the next question, as you are working as a regional counsel for the
[13:58.540 --> 14:04.820] EPAC region, my question is what unique challenges do you encounter regarding legal compliance
[14:04.820 --> 14:09.060] and risk management across different jurisdictions?
[14:09.060 --> 14:13.900] So I think there are various nuances and this becomes important because what is the region
[14:13.900 --> 14:15.180] that you are looking at?
[14:15.180 --> 14:19.780] For example, if I talk about my region, Asia Pacific, we have countries which are very
[14:19.780 --> 14:20.780] advanced.
[14:20.780 --> 14:25.020] We have Singapore, we have Australia, we have New Zealand, Japan and we have countries which
[14:25.020 --> 14:26.020] are not so advanced.
[14:26.020 --> 14:30.940] You have Myanmar, you have Nepal, countries at different level of development on different
[14:30.940 --> 14:34.540] levels where they are on the transparency index, on different level where they are on
[14:34.540 --> 14:39.660] corporate governance, how much laws have evolved in each country and I think that's the biggest
[14:39.660 --> 14:41.940] challenge when you are managing a region.
[14:41.940 --> 14:46.340] The most obvious answer to that would be countries have different laws and you have to know what's
[14:46.340 --> 14:47.980] a problem in which law.
[14:47.980 --> 14:51.540] Data privacy, for example, we were talking, may be a much bigger issue in Singapore and
[14:51.540 --> 14:55.140] Vietnam than it might be in, for example, Myanmar.
[14:55.140 --> 14:59.580] Corruption, anti-bribery, may be a much bigger issue in Singapore compared to what it might
[14:59.580 --> 15:02.060] be in another jurisdiction.
[15:02.060 --> 15:06.220] Language may be a big hurdle when you are negotiating contracts in Japan vis-a-vis whether
[15:06.220 --> 15:09.100] you are negotiating contracts in Australia.
[15:09.100 --> 15:12.500] So there are various issues that arise when you are managing a region.
[15:12.500 --> 15:17.660] The biggest one of them is the level of development of law, the level of availability of good
[15:17.740 --> 15:23.260] councils who understand the subject, who are ready to negotiate with you at the same level.
[15:23.260 --> 15:27.180] And that, I think, is most important for a company like ours or for any large company
[15:27.180 --> 15:29.820] which is operating in various jurisdictions to navigate.
[15:29.820 --> 15:34.580] The laws, the cultural nuances, you know, I'll give you an example that long back I
[15:34.580 --> 15:39.220] was told that because I'm the Council for Asia Pacific, why would you take also the
[15:39.220 --> 15:44.180] responsibility of negotiating some contracts in Pakistan and I told them this may be a
[15:44.180 --> 15:45.180] cultural issue.
[15:45.180 --> 15:50.180] I'm negotiating a contract for a European organization with a Pakistan lawyer because
[15:50.180 --> 15:54.500] of the historical reason why they may not take every objection that we have in a very
[15:54.500 --> 15:55.500] constructive manner.
[15:55.500 --> 16:00.140] It may become an ego issue at some point of time and I told them it was best that somebody
[16:00.140 --> 16:02.860] else does this negotiation rather than me.
[16:02.860 --> 16:03.860] And that was only a hunch.
[16:03.860 --> 16:08.220] I may be completely wrong, but these things have to be kept in mind.
[16:08.220 --> 16:13.380] Sometimes we've had experiences where we are into negotiation and the culture of negotiating
[16:13.380 --> 16:14.380] is very different.
[16:14.580 --> 16:18.540] You know, we were, I would not take name, but we were in a country, a very advanced
[16:18.540 --> 16:22.140] country in our region where we used to go there for a negotiation.
[16:22.140 --> 16:26.100] We used to reach on Monday and we had flight back generally on Friday night so that we
[16:26.100 --> 16:27.420] could come home.
[16:27.420 --> 16:32.220] And for four days we would negotiate something and we would reach some kind of an agreement
[16:32.220 --> 16:36.300] and on the fifth day when our flight was like three hours to go, somebody else would just
[16:36.300 --> 16:40.540] come up and say, I don't agree with any of this and please, we will start again, we will
[16:40.540 --> 16:43.260] work on a Saturday, we will do this again.
[16:43.260 --> 16:45.620] You know, you have to overstay your visit.
[16:45.620 --> 16:46.620] It's a negotiation strategy.
[16:46.620 --> 16:47.620] I understand.
[16:47.620 --> 16:50.460] They're trying to last minute put pressure to get something.
[16:50.460 --> 16:51.940] This will not happen in another place.
[16:51.940 --> 16:56.540] This will never happen, for example, in Japan where there's a lot of respect for each other's
[16:56.540 --> 16:59.380] time and each other's commitment that someone has given.
[16:59.380 --> 17:03.500] So there are these issues that always keep coming up and we should be mindful of.
[17:03.500 --> 17:09.140] So I think laws, culture, language, all of these are issues that one needs to be in large
[17:09.140 --> 17:11.340] jurisdictions.
[17:11.340 --> 17:12.340] Very well said.
[17:12.420 --> 17:16.700] And you know, my next question is, in fact, a follow up question to this is, do you think
[17:16.700 --> 17:22.060] generative AI can help the legal function in this particular area?
[17:22.060 --> 17:23.280] I would definitely think so.
[17:23.280 --> 17:29.500] I would think there is huge scope of technology in general and AI in particular to help these
[17:29.500 --> 17:31.820] multi-jurisdictional problems.
[17:31.820 --> 17:36.540] For example, AI would know from machine learning, this is something that I was promised since
[17:36.540 --> 17:40.780] last three, four years now, that there are tools which can tell me which governing law
[17:40.900 --> 17:45.820] should I use in which kind of contracts or not, because they would analyze all the cases,
[17:45.820 --> 17:50.260] past cases, and be able to tell me that it's better to agree on an English law rather than
[17:50.260 --> 17:54.620] Australian law, for example, or choose this jurisdiction over that based on the recent
[17:54.620 --> 17:59.980] cases, recent laws which are coming up, or even, you know, give predictions on where
[17:59.980 --> 18:05.180] your dispute would land, how much time it would take, what can be the outcomes of these.
[18:05.180 --> 18:08.500] So there are a lot of areas where AI can help in this.
[18:08.580 --> 18:13.220] This is very important because, like I said, if you are having a large region, if you're
[18:13.220 --> 18:18.340] dealing with multi-jurisdictional issues, you know, these days, the transactions themselves
[18:18.340 --> 18:22.220] have become so complex, you have equipments coming from one place, raw material coming
[18:22.220 --> 18:26.980] from another country, sales is happening in high seas, the project is located in a third
[18:26.980 --> 18:29.860] country, the financing authority is in a fourth country.
[18:29.860 --> 18:31.900] So it's too complicated there.
[18:31.900 --> 18:36.820] I think it's very, very large scope where AI can make a task much more easier.
[18:36.820 --> 18:43.340] Translations, adaption of clauses, telling you the rules and laws of language, prediction
[18:43.340 --> 18:48.620] of disputes, contract reviews, a lot of things can be done with the help of machine learning
[18:48.620 --> 18:49.620] AI.
[18:49.620 --> 18:55.260] For example, even simple tasks like drafting a letter, if I have to draft a letter to a
[18:55.260 --> 19:02.300] client in Seoul, for example, or any other jurisdiction, to understand the nuances of
[19:02.300 --> 19:07.700] how this letter should be versus what I write, by help of an AI is much easier.
[19:07.700 --> 19:11.940] It will be culturally more appropriate, it would use language which would be not offensive,
[19:11.940 --> 19:15.380] for example, to that different culture, you know, which may be very different and which
[19:15.380 --> 19:17.260] may be something that we were not aware of.
[19:17.260 --> 19:22.660] AI can help you in dropping these cultural issues, in language issues, in legal issues.
[19:22.660 --> 19:25.580] So I believe there's a lot of scope there.
[19:25.580 --> 19:27.500] Great use cases mentioned there.
[19:27.500 --> 19:29.900] And I have actually two follow-up questions.
[19:29.900 --> 19:34.580] You mentioned the use case where the AI could give you a prediction that, you know, okay,
[19:34.580 --> 19:36.980] using this clause may not be the beneficiary.
[19:36.980 --> 19:39.820] So what should be the source documents for these things?
[19:39.820 --> 19:45.020] Should the AI tools focus on what kind of contracts everybody is signing these days
[19:45.020 --> 19:50.100] or should they focus on the decisions that particular country has passed, where all they
[19:50.100 --> 19:51.780] should be looking at?
[19:51.780 --> 19:58.300] So I think case outcome prediction or predictions based on how your contract would perform in
[19:58.340 --> 20:02.620] case of a dispute is becoming a big feature going forward.
[20:02.620 --> 20:07.660] And like you said, yes, one starting point is analyzing the recent judgments.
[20:07.660 --> 20:12.860] So for example, if I have a fitness of purpose clause, which I'm giving, AI can very well
[20:12.860 --> 20:18.620] tell me and with a large amount of digitization of both cases, large amount of digitization
[20:18.620 --> 20:25.380] and interconnection of all the reported cases, of all the laws, new rules and regulations,
[20:25.380 --> 20:31.140] it is very easy for AI to tell me how many fitness of purpose cases were actually listed
[20:31.140 --> 20:36.660] in the past 10 years in that particular jurisdiction, for example, and what was the damages awarded
[20:36.660 --> 20:40.820] and what was the threshold of proving fitness of purpose, for example.
[20:40.820 --> 20:45.220] So fitness of purpose in Australia, for example, may be very problematic because if you are
[20:45.220 --> 20:50.820] proven that your goods were not fit for purpose, the damages may be very high compared to another
[20:50.820 --> 20:54.460] jurisdiction where you may have very few cases on fitness of purpose.
[20:54.460 --> 20:58.740] It's more on contractual breach, actual breach where damages are sought and fitness of purpose
[20:58.740 --> 21:02.580] would not have been bought as a defense in many of the other jurisdictions.
[21:02.580 --> 21:07.500] So case outcome predictions based on the damages already awarded, case data, how many times
[21:07.500 --> 21:10.940] this concept was promissory estoppel against the government?
[21:10.940 --> 21:13.580] How many times has this concept been litigated?
[21:13.580 --> 21:16.020] How many times have courts upheld this concept?
[21:16.020 --> 21:21.580] And these are things that can help us in predicting the outcomes of the cases very much possible.
[21:21.580 --> 21:26.220] A large amount of data has been digitized and we should take advantage of this fact.
[21:26.220 --> 21:31.180] We should take advantage of the fact that most of the jurisdiction's votes are now very
[21:31.180 --> 21:36.180] transparent in the verdicts that they are giving you sophisticated tools to help you
[21:36.180 --> 21:37.780] in searching of these documents.
[21:37.780 --> 21:40.300] So I think there's a lot of scope there.
[21:40.300 --> 21:41.300] Wise thoughts.
[21:41.300 --> 21:45.740] You raise a very interesting point of such prediction in disputes.
[21:45.740 --> 21:48.540] So have you ever used these kind of tools?
[21:48.540 --> 21:53.740] Because as per my knowledge, such tools have existed in past in the US, but none for India
[21:53.740 --> 21:56.360] and the IPAC region that I'm aware of.
[21:56.360 --> 22:00.100] So what are your thoughts on these kind of tools coming into the future?
[22:00.100 --> 22:04.300] Three or four questions back, I had started saying that we have been promised this for
[22:04.300 --> 22:05.300] very long.
[22:05.300 --> 22:06.300] Right.
[22:06.300 --> 22:08.740] So people have started promising it since last five years.
[22:08.740 --> 22:10.140] Our software can do this for you.
[22:10.140 --> 22:12.100] We will suggest which governing law to take.
[22:12.100 --> 22:13.540] We will suggest this.
[22:13.540 --> 22:18.700] But at least our experience has not been very up to the mark on this.
[22:18.700 --> 22:22.740] But I think where things are improving, and especially in the last one or two years, I
[22:22.740 --> 22:29.500] would say post-COVID and lastly in the last two years, is a lot of these issues are now
[22:29.500 --> 22:30.500] much better.
[22:30.500 --> 22:35.220] Of course, you cannot predict what will be the outcome of a case, but we are at a stage
[22:35.220 --> 22:40.420] where I know that it may be higher risk to have this clause or to have this governing
[22:40.420 --> 22:42.320] law vis-a-vis another law.
[22:42.320 --> 22:47.560] So that's where the lawyer and the technology have to work together and come up with the
[22:47.560 --> 22:48.560] answer.
[22:48.560 --> 22:52.400] So I don't think prediction per se is completely there, although I do believe it will very
[22:52.400 --> 22:53.400] soon be.
[22:53.400 --> 22:56.880] With the amount of data already there, it's just a matter of time that we have enough
[22:56.880 --> 22:59.800] algorithms to churn out these numbers and predictions.
[22:59.800 --> 23:03.400] So I would say if we are talking again in two years' time, the technology would have
[23:03.400 --> 23:05.760] improved much more on this.
[23:05.760 --> 23:10.920] But right now what I'm getting is a lot of good help from technology to either agree
[23:10.960 --> 23:14.560] or not agree on certain clauses and governing laws.
[23:14.560 --> 23:15.840] Absolutely.
[23:15.840 --> 23:20.560] My follow-up question was that, you know, when you are having a litigation or a dispute,
[23:20.560 --> 23:26.880] the external councils are financially incentivized to actually follow on that litigation.
[23:26.880 --> 23:32.360] But if you have a tool which can predict the chances of success in that particular litigation
[23:32.360 --> 23:37.640] and the monetary reward that you can get, then I think as an in-house council, you can
[23:37.640 --> 23:43.680] take a better call as to whether prosecuting this dispute via the help of courts will
[23:43.680 --> 23:46.280] give me an ROI.
[23:46.280 --> 23:51.160] And that's exactly what I, for example, told you with the example of fitness of purpose.
[23:51.160 --> 23:56.640] As a concept, if I know that I'm entering a contract in Australia where fitness of purpose
[23:56.640 --> 24:02.440] in general has been known to be given money, much larger damages, I will try to eliminate
[24:02.440 --> 24:04.760] the fitness of purpose clause by drafting itself.
[24:04.800 --> 24:07.800] I'll negotiate much harder not to give fitness.
[24:07.800 --> 24:09.360] It's just an example that I'm giving.
[24:09.360 --> 24:10.640] Obviously, there'll be winner.
[24:10.640 --> 24:15.640] And I think that is where the information will help, especially corporate lawyers, especially
[24:15.640 --> 24:20.600] councils negotiating contracts, to try and get language in their contract, which will
[24:20.600 --> 24:26.360] help them in case a dispute arises and to avoid any language which can result in potential
[24:26.360 --> 24:30.040] larger losses and damages to their companies in case of such issues.
[24:30.040 --> 24:35.480] And that is where I think data would be very useful to further be analyzed and assisted.
[24:36.480 --> 24:41.200] Right. And, you know, Grijesh, when we're talking about external councils, I will ask
[24:41.200 --> 24:44.840] this question because you have worked in both the capacities.
[24:44.840 --> 24:49.800] Now, when you are sitting as an in-house council, what are the factors that are looking out
[24:49.800 --> 24:52.200] while selecting an external council?
[24:52.200 --> 24:53.120] Is it cash?
[24:53.120 --> 24:54.440] Is it credibility?
[24:54.440 --> 24:56.560] Or is it the capacity?
[24:56.560 --> 24:57.440] I think everything.
[24:57.840 --> 24:59.840] I think two disputes are similar.
[24:59.840 --> 25:03.760] Therefore, the choice of lawyers for no two disputes would be the same.
[25:03.760 --> 25:08.880] If I'm having a much larger and complicated dispute, I would try to get a law firm or
[25:08.880 --> 25:11.160] a council who's very good in that issue.
[25:11.160 --> 25:14.400] Money may not be a consideration for me in such cases.
[25:14.400 --> 25:19.680] But if I have smaller issues, if I have smaller consumer complaints, hundreds of litigation
[25:19.680 --> 25:24.080] a company may have, or vice versa, if a company may not have any litigation, may only have
[25:24.080 --> 25:27.160] one or two litigations, then it would be very different.
[25:27.160 --> 25:30.440] Sometimes you may want to choose individual litigants sitting in the court.
[25:30.440 --> 25:33.760] Sometimes you may want to choose large law firms who have tens of associates to help
[25:33.760 --> 25:35.880] to do that research.
[25:35.880 --> 25:39.540] Cost is, of course, always a consideration when doing this.
[25:39.540 --> 25:43.720] How much you've worked in past, how much success you've had together with those partners,
[25:43.720 --> 25:47.560] with those law firms, with those lawyers is something which is very important.
[25:47.560 --> 25:52.000] So I think it's a mix of all, and it will depend on the kind of dispute that is going
[25:52.000 --> 25:56.080] to come up or the kind of help that you are requiring from the external councils, whether
[25:56.160 --> 26:00.760] it is an M&A deal, whether it is an antitrust issue, whether it is a contract negotiation,
[26:00.760 --> 26:05.160] whether it is project financing, or whether it is a plain and simple litigation in court.
[26:05.160 --> 26:09.760] So it will depend what is the requirement, what is the capacity of the organization,
[26:09.760 --> 26:14.520] how much are you willing to spend, what kind of lawyers do you have in your city.
[26:14.520 --> 26:19.760] For example, if I'm a company in a tier two city, I may not get so many legal help at
[26:19.760 --> 26:24.700] that place as I would if I am sitting in Mumbai or Delhi or Bangalore or Hyderabad.
[26:24.700 --> 26:28.540] So all of these are considerations that need to be seen.
[26:28.540 --> 26:29.700] Wise words there.
[26:29.700 --> 26:35.140] We know your vast tendency to avoid controversial questions.
[26:35.140 --> 26:40.180] So I'll put you in that uncomfortable position again one more time, and this is the last
[26:40.180 --> 26:41.780] time I promise you that.
[26:41.780 --> 26:46.260] My question is that, you know, as you also shared that when you're working in a law firm,
[26:46.260 --> 26:51.380] you always have that pressure of making it great or bringing it business.
[26:51.380 --> 26:55.860] And if not that, you definitely have the pressure to have high billable hours, like
[26:55.860 --> 26:57.720] if you're in junior position.
[26:57.720 --> 27:03.980] So my question is, with that context, bringing in AI tools or bringing in tools that actually
[27:03.980 --> 27:10.060] increases the law firm's lawyers' efficiency is not in their financial incentive.
[27:10.060 --> 27:11.880] What are your thoughts on that?
[27:11.880 --> 27:12.880] Not necessarily.
[27:12.880 --> 27:17.100] So I was having a very interesting conversation just the other day with a partner of a large
[27:17.100 --> 27:19.300] law firm based out of Singapore.
[27:19.300 --> 27:23.340] And something like this had come up in that discussion and I'll quote what he said, and
[27:23.340 --> 27:24.340] I really liked it.
[27:24.340 --> 27:28.500] And his thought was that till now, we've been having a kind of a pyramid structures in law
[27:28.500 --> 27:33.060] firm where you add a few partners and then you have larger salad partners and then you
[27:33.060 --> 27:37.040] have much larger number of senior associates and large number of associates and interns
[27:37.040 --> 27:39.140] and everybody who was doing work.
[27:39.140 --> 27:43.180] And companies were built according to the rates for the associates, early rates for
[27:43.180 --> 27:45.260] the associates, senior associate partner and so on.
[27:45.260 --> 27:48.780] And partner were billing lesser hours obviously compared to an associate.
[27:48.820 --> 27:55.140] A lot of bill was generated because of research, because of initial graphs, associates spending
[27:55.140 --> 28:00.100] 100 hours to understand the project and get to know versus the actual advice that the
[28:00.100 --> 28:04.460] partner was giving would have only spent, let's say, 10 hours in that contract or in
[28:04.460 --> 28:05.460] that transaction.
[28:05.900 --> 28:09.700] Now, with the help of AI, what will happen is this pyramid structure will break down
[28:10.020 --> 28:15.540] and that 100 hours that the associates were doing would be done by AI in much lesser number
[28:15.540 --> 28:16.540] of hours.
[28:16.740 --> 28:22.260] But the effective time for analyzing what has come out of that AI would be done at a
[28:22.260 --> 28:23.900] much higher level of skill set.
[28:23.900 --> 28:27.900] So partners, retained partners would be billing in more hours in that sense.
[28:28.220 --> 28:34.180] I think we would come into an era where law firms would do a lot more value add as partners.
[28:34.660 --> 28:37.900] I don't think the billings will come down drastically, although they will.
[28:37.900 --> 28:44.460] For example, M&A is our biggest examples of this use of AI in analyzing data rooms in
[28:44.460 --> 28:48.980] large M&A transactions would really bring the cost down of due diligence.
[28:49.180 --> 28:53.860] And more importantly, it will bring down the time also of these due diligence of M&A
[28:53.860 --> 28:56.340] documents that are done on target companies.
[28:56.660 --> 29:01.180] But then the senior councils with the senior partners, senior associates would have much
[29:01.180 --> 29:04.180] more time to analyze these issues and give valued advice.
[29:04.180 --> 29:07.860] So I think that would be something which I am looking forward to and which will happen.
[29:07.860 --> 29:11.340] And I don't think that there would be lesser work, lesser bill.
[29:11.620 --> 29:14.500] I would say there would be more value add by senior professionals.
[29:15.380 --> 29:19.020] So external lawyers are not going to make less money for sure.
[29:19.340 --> 29:25.740] I don't think anybody, I don't think we should look at AI in terms of a tool for
[29:25.740 --> 29:26.860] reduction of money.
[29:26.860 --> 29:31.260] We should look at it as a means of value add or doing things better.
[29:31.500 --> 29:35.620] Of course, reduction of money where it is important would happen and should happen at
[29:35.620 --> 29:39.020] certain places where work would be done more efficiently.
[29:39.020 --> 29:40.700] Obviously, number of hours would be less.
[29:40.700 --> 29:45.860] For example, something like I said on M&A, but the value that you will generate in doing
[29:45.860 --> 29:50.260] the same amount in lesser hours, the value that you would generate even on the cost of
[29:50.260 --> 29:55.460] licenses and the cost of these softwares, so that all has to be taken into account.
[29:55.460 --> 30:02.140] And it would be very good for especially lawyers to do larger value add with the amount
[30:02.140 --> 30:09.180] of data crunching being done by AI rather than by 10 interns or 15 associates or junior
[30:09.180 --> 30:14.300] associates doing the same, maybe sometimes more prone to mistakes or missing documents
[30:14.300 --> 30:19.460] and so on. So I think everybody in this game would be better value add, better services,
[30:19.620 --> 30:23.460] a more efficient result, including, of course, cost reduction also.
[30:24.500 --> 30:26.660] Completely agree with you there, Brijesh.
[30:26.660 --> 30:30.660] And you know, you brought in a very interesting use case close to my heart, which is
[30:30.660 --> 30:36.260] bringing AI in M&A peak and that's a very interesting use case there because, again, I
[30:36.260 --> 30:41.460] have been part of some M&As and completely agree with there's a lot of mundane and
[30:41.460 --> 30:45.580] clerical work going on, which can be easily automated with AI.
[30:45.980 --> 30:49.300] Now, that brings me to another very interesting use case.
[30:49.460 --> 30:55.220] I wish I could take credit for coming up with it, but it was suggested by someone else.
[30:55.420 --> 31:01.420] And he said that a similar problem also exists while doing internal audit, because he said
[31:01.420 --> 31:05.260] that if you look at M&A, there's a large amount of data and that you have to.
[31:05.500 --> 31:10.380] But if you look at the internal audit, which is done by an accounting firm, then also
[31:10.380 --> 31:15.140] they're processing a lot of accounting data, a lot of other information and then coming
[31:15.140 --> 31:19.580] up with suggestions. What are your thoughts on using AI for internal audit?
[31:20.420 --> 31:24.780] I think already being used, it's all about how much information can we gather.
[31:24.780 --> 31:30.700] So you have currently also contract management tools, for example, which are embedded in
[31:30.700 --> 31:36.540] the company's system, which have timestamps, which have a process already there, which
[31:36.540 --> 31:41.460] will show by analysis that how a file moves or how a purchase order moves.
[31:41.460 --> 31:45.100] You have these, you know, fancy terms on procure to pay.
[31:45.180 --> 31:50.820] You have from a time the offer comes in to the time a subcontract agreement is drafted.
[31:50.820 --> 31:56.260] So you have large tools which are already timestamping them, seeing, giving many people
[31:56.260 --> 32:02.380] in the group acceleration, escalation issues and so on, which is all being used to better
[32:02.380 --> 32:07.100] the system. Ultimately, your internal audit is to see whether your system is working well
[32:07.100 --> 32:12.380] or not. And a lot of these tools will show you much faster that these systems are working
[32:12.380 --> 32:17.580] or not working. If I have a flow chart of 10 steps which have to be taken care of and
[32:17.580 --> 32:22.260] if my system is already tracking those 10 steps, an internal auditors job becomes much
[32:22.260 --> 32:27.300] easier, whether it be a tax accountancy issue, whether it be an internal audit issue, whether
[32:27.300 --> 32:32.620] it be a compliance issue, regulatory issue. Ultimately, everything see works on data.
[32:32.620 --> 32:37.260] And if you have timestamp data, if you have escalation points, if you have process flows
[32:37.260 --> 32:41.700] well defined and checkboxes being done according to those process flows, then I think the job
[32:41.700 --> 32:44.100] would be much easier.
[32:44.100 --> 32:48.860] Very interesting insights there. So Brijesh, as we are running out of time, I'll ask the
[32:48.860 --> 32:55.020] last question. What advice would you give to an aspiring lawyer who aim to take a path
[32:55.020 --> 32:56.580] similar to yours?
[32:56.580 --> 33:02.740] Well, I think complete your law properly. I mean, do your hard work and not cut short
[33:02.740 --> 33:07.940] on actual legal learning. I think that's the most important advice. What happens is sometimes,
[33:07.940 --> 33:13.540] especially in an age where we are, if your senior says draft up a letter, draft up a
[33:14.540 --> 33:21.180] and get it drafted there. I would say embrace technology intelligently. Know your law because
[33:21.180 --> 33:25.780] there's no substitute for hard work. No matter how much technology comes in, how much things
[33:25.780 --> 33:31.140] are automated, the lawyer needs to know their law to help get something from the technology.
[33:31.140 --> 33:37.140] So I would advise all young lawyers, embrace technology in a manner which will help them.
[33:37.140 --> 33:42.100] Do use technology. There is a lot of data being given out by technology. A lot of things
[33:42.100 --> 33:47.300] have become easier, and we should embrace it, but we should not forget that it's not
[33:47.300 --> 33:53.420] a shortcut tool. You should embrace it intelligently. I think one should use technology. I would
[33:53.420 --> 33:58.100] advise all lawyers to think beyond the box and things beyond what they are taught in
[33:58.100 --> 34:03.900] law schools and in the early years of their career. And use these time effectively to
[34:03.900 --> 34:08.500] better their legal profession, to better their networking skills, to better their speaking
[34:08.580 --> 34:12.020] skills, and so on. And that would be my advice to all the young lawyers.
[34:13.060 --> 34:20.100] Wise words, Brijesh. And thanks a lot for joining us again on this podcast. And as I promised,
[34:20.660 --> 34:26.260] if you agree to record this podcast again, I'll reveal the reason why we had to do it
[34:26.260 --> 34:32.820] in two parts. And now I think I owe that explanation to everyone. So the reason we are
[34:32.820 --> 34:39.220] doing this podcast in two parts is because I forgot to record the second part of our podcast.
[34:39.220 --> 34:44.580] And thanks a lot, Brijesh, for agreeing to redo this podcast. It was a lifesaver,
[34:44.580 --> 34:48.900] and I got a privilege to spend half an hour and two an hour again with you.
[34:49.460 --> 34:53.300] Thanks. Thanks, Avi. I think the pleasure is mine. I think it is an interesting discussion.
[34:53.300 --> 34:58.820] I really enjoyed it. And we are touching upon some important work which is happening. All the best
[34:58.900 --> 35:04.340] to you and your organization for doing this and taking it forward. I hope to talk to you
[35:04.340 --> 35:10.100] sometime again soon. We hope that too, Brijesh. Thanks a lot for joining us again. Thank you.
[35:11.780 --> 35:18.340] And that, dear listeners, brings us to the end of today's entertaining and enlightening episode.
[35:19.220 --> 35:26.500] I hope you have enjoyed this deep dive into the legal universe with our exceptional guest,
[35:26.500 --> 35:34.020] Mr. Brijesh Anand. His insights have been nothing short of remarkable. Before we part ways,
[35:34.020 --> 35:43.700] remember to subscribe us on your favorite podcast platform. Stay curious, stay caffeinated.