How Top Legal Teams Use AI

Shamil Malachiyev: Hello everyone and welcome to this week's episode of The AI Floor. My guest today is the General Counsel at Ironclad. Most people in legal either build these tools or buy them. She does both. She helps shape one of the world's biggest AI contracting platforms, and then she lives on it too, using it for her own company's contracts. The builder and her own toughest customer. In the past, she also burned out of big law, walked away to teach 6 a.m. spin classes for a year, and came back to rebuild all the way to General Counsel. So please welcome Jasmine Singh.

Jasmine Singh: Thank you so much for having me. I'm thrilled to be here.

Shamil Malachiyev: Thrilled to have you here. Can we start with you telling our listeners a bit about who Jasmine is and what path she's walked to get to where she is right now?

Jasmine Singh: I actually love answering that question, because I feel like I have lived a thousand lives as a lawyer. I currently sit as General Counsel, but I started my career in a very different place. I was a litigator in big law firms in San Francisco for a number of years. As you said, I found myself burnt out and really wanting something different in my career, but unsure what that was and how to find it. So I took a break and became a spin instructor, and I'm sure we'll talk more about that. What that helped me realise was that I wanted to come back to the law in a way that allowed me to connect and continue to find my joy and my passions, by way of mission-driven companies and helping to build things. So I transitioned in-house, first to 24 Hour Fitness, which was a really nice connection to my personal passion around fitness, health and wellness. I cut my teeth as a commercial lawyer there, then moved to Pinterest, where I led global commercial and scaled my team. Then I was Deputy GC at Patreon, which is a creator-based payments platform, and then took my first job as General Counsel at Binti, an early-stage SaaS company that makes software for child welfare agencies. And then I moved to Ironclad a few years ago.

Throughout all that I've remained really active in volunteer efforts. I'm on the board of a community foundation called Change Lawyers. I'm also on the board of a nonprofit law firm called Public Advocates. I started a decentralised philanthropy initiative whereby we raise money for nonprofits to provide direct services. And since you asked who I am, not just what I do, I'm also a former competitive hip hop dancer, which has played a really formative role in my life.

Shamil Malachiyev: So how competitive are you as a person?

Jasmine Singh: Pretty competitive, and mostly with myself. I actually have to stave it off sometimes, because it can be a little unhealthy.

Shamil Malachiyev: At one point it does help you burn out quickly in situations where you don't find meaning for yourself. But at the same time it helps you scale fast and learn quick.

Jasmine Singh: Totally. Push hard. I grew up playing team sports. I ran track, I played basketball, I was captain of my respective teams. I take a lot of the lessons I learned as an athlete and as part of a dance team to my work. Some of those lessons are the value of discipline, hard work, never giving up, moving at a fast pace. But also the realisation that I really do go a thousand percent, and in those sports I knew I needed to dial it back, and the same is true in dance. If you go full out all the time, it isn't good for your end result.

Shamil Malachiyev: When you compare it to sports, I feel like the last three years across all industries we're seeing this sprint where everybody has to become an athlete and run as fast as they can. What has that journey been like for you, having to go through so many shifts within industries?

Jasmine Singh: Honestly, it's been a lot of knowing when to push and when to pull back, which is really hard when an industry is running at a fast pace. Starting out in law firms, what we're taught is you work to the client's demands. There's no moment where you actively pull back or say no or limit your work out of self-preservation, because the expectation is that if a matter is going to litigation, and I was a litigator, it's a hundred percent of your effort and your time. I think when I moved in-house was the first time I heard people say, no, balance is okay. You don't have to give everything a hundred percent. The expectation is not perfection. The expectation is what the business needs to get the job done.

So finding a way to adjust how I was working to my circumstances was a lesson learned. And to address your question, the pace of the industry is changing so fast, and it demands that we run at a quick clip and don't get left behind. Even within this, there's balance to be had. It's not don't adopt the technology. It is absolutely adopt it, learn it, run alongside it, but be mindful of when and if you need moments to breathe, and have the agency to do that yourself, because nobody else is going to tell you if it's okay or when it's okay. That has to be you.

Shamil Malachiyev: I like that you have big law experience, in-house legal team experience, and you're also, I could say, part product manager, using the software and directing the software that all of these people across industries use to work through their own contracts.

Jasmine Singh: Yeah, go ahead.

Shamil Malachiyev: How do you see what's moving the fastest, and how is everyone trying to keep up with the changes?

Jasmine Singh: What's moving the fastest is frankly just the pace of technology development. How quickly models are being developed, then how quickly products are built on top of those models, then how quickly those products iterate and deliver new and different use cases. With the advent of AI and the ability to code using AI, there is just way more out there to consume than ever before. The good thing is AI helps you synthesise, shorten and consume it a little more easily. The challenge is knowing where to direct your focus at any given moment, and not feeling like, my gosh, I need to look at that thing, and that one, and that one, and not diving deep into anything. That's the dangerous part of this circumstance: not giving anything true focus.

Shamil Malachiyev: Not long ago I released an episode with Xavier, who's also a General Counsel, at a company called OneAdvanced. The thing he pointed to a lot is how the dynamic between in-house teams and outside counsel firms is changing, because law firms are very consultant-heavy. How do you see that relationship evolving? I believe the six-minute increments and all of that are on the path to disappearing and evolving into something else.

Jasmine Singh: I'm of the mindset that each player in the legal ecosystem will continue to have a space. They'll just modify how they approach and occupy that space. What I mean for outside counsel in particular is that outside counsel brings multiple things. One is an understanding of a pan-market view: how others are doing it, how peers in the industry are doing it, how peers in other industries are doing it. Sometimes that input is useful for me and there's no way I could get it but for going to outside counsel. The value of a firm that knows what a lot of people are doing doesn't go away, in my opinion. The second thing I go to outside counsel for is specialisation, expertise in areas where people on my team, or me, just don't have a solid understanding of what the law is. That expertise also doesn't go away. Those firm lawyers have spent years cultivating that deep knowledge and skill set.

What I see changing is my ability to frame my question differently and go in with a lot more understanding than before. If I didn't know a lot about a topic, like most lawyers I'd do Google searches to try to figure things out and get enough information to frame a better, more pointed question. AI has supercharged that ability. Now I can do a ton more research and synthesise way more information. Sometimes I'll say: these are the facts and circumstances I'm presented with. What questions should I ask my internal business partners to better understand, based on what I believe the legal risk is? Then I go ask those questions and get the answers. Then I say, now I need a synthesised view of the outstanding questions. Then I can go to outside counsel and say, I have this problem, I've done some initial research that suggests this is the relevant information, now can you give me a more time-boxed answer to that question based on this investigation. I've heard outside counsel suggest that this can go two different ways. On the one hand, I think it helps them to...

Shamil Malachiyev: Sorry. I was going to ask, do you think it evolves to somebody having to take responsibility for the final advice and how credible it is? We can do a lot of research, but what we need the consultancy for is to say, okay, this is correct, and put their stamp on it. We take responsibility that the research you've done is the correct one.

Jasmine Singh: Correct. You know, there there are there is a reason why we go to outside counsel on certain high risks, high stakes things, right? We want people who are experts and who are willing and able to put their stamp and seal of approval on the information and advice they're giving us because they have the lived experience of having enforced it or having seen it in action in other companies. That accountability comes by way of their lived experience telling us that whatever advice we're getting is helpful, or their extreme diligence in looking into the issue, which is frankly something a lot of in-house teams don't have the bandwidth to do. We're more leanly staffed. We have a lot more on our plates than the ability to do a deep dive into some nuanced regulatory scheme, for example.

Shamil Malachiyev: And if we look at internal legal ops teams, these are the primary customers of yours in all the other companies. How do you see their work? What are the biggest parts that evolved with AI, and what are the requirements that changed for the people working in it?

Jasmine Singh: First and foremost, for legal ops teams, what's changed is they have to have a deep understanding of what AI tools are out there, how they work, and which are right for their team's workflows. You have to understand what workflows your legal team is working on, which of them are automated, which are semi-automated, which are entirely manual, and whether tools now exist that could eliminate some of the thrash or inefficiency in any of those systems. And then thinking not just about the system to be built but the end goal of building it. I see that as within the purview of legal ops, and it's changed the game, because now they're not just helping administer what existed before. They're doing first-principles thinking about how to reimagine the entire way we've been delivering this legal service. The best legal ops people have always thought that way, but now there's new technology at their fingertips constantly asking for their attention and their evaluation of whether it can actually do that.

The other part of what's changing is their own ability to synthesise and report information from within their teams. To run an analysis of all the tickets you've ever received and highlight the top five themes across them, how many days it took to respond, and then build an agent to deflect all of those questions so we never receive them manually in a Slack channel again. We've got a bot that answers those questions on our behalf. That's the type of work ops professionals have always been doing, but now they're doing it in a totally different way, because they can synthesise and digest that information more easily.

Shamil Malachiyev: So for somebody to be a legal professional right now, you have to understand business processes and be able to map them out and try to automate them, and you have to understand the technology, what AI is and where it fits. Is this becoming a standard pack? Without this, you shouldn't be trying to enter the industry right now?

Jasmine Singh: Yes. I do believe, especially to be a successful in-house lawyer, you can't just be someone who delivers good legal advice. You have to be someone who understands the business operations and the systems, and builds a process to deliver legal advice within that broader system, whatever it might be. You have to have an understanding of which technology you can use to bridge those two things together and deliver fantastic legal advice to your clients by way of a mechanism that actually lands for the client.

Shamil Malachiyev: Do you ever feel that, because you have to think at the cutting edge, since you're shaping a product, and everyone has expectations of making their products AI-first, always thinking about the latest technologies and how to use them for the product now, we as technologists, and I think both of us are, start living in a bubble where we think everybody operates at the same level, everybody listens to the same news? Then you start talking to people in the industry and you're like, wait a minute, isn't everyone aware of all these things? When talking to customers, the people out in the industry, what's the standard adoption level compared to your expectation? What are you noticing?

Jasmine Singh: I'm noticing three groups of people. There are, to your point, the group in my echo chamber that are super users of AI. They use it all day, every day. They know what the technology does, they know its limits and its benefits, they're constantly tinkering. Then there's a second group more in the camp of, I'm being told I have to use AI, I don't really want to, I don't really understand, I'm not the best at this, but I have to, so I'm going to try to keep up. And then there are the people that are, I don't really trust the technology, I don't see value, it's faster and better if I do it myself. They're the reluctant group. Right now I see people mostly in that middle bucket if I go outside my bubble. My bubble is mostly the first bucket, everybody using it and pushing it. But most practising lawyers I talk to outside of Silicon Valley are in that middle area: I know I have to use this and I'm trying my best, but I haven't quite figured it out yet, and if I'm being honest, I'm a little overwhelmed by what I don't know.

Shamil Malachiyev: What would you say to all of those people? Clearly everybody is walking this path right now, whether they want to or not. Some of us have already created those neural connections that help us make sense of these technologies. What would be the path you'd advise? For somebody on your team who joins and says, I want to learn about these things, what's the first step?

Jasmine Singh: Try, try, try, test, test, test. To your question about how I do it on my team, I make two things true. One is I do require usage of AI. So they know it is an expectation of mine. It is clear that trying is required. And then I say that failure is acceptable, right? I acknowledge that not everything is going to be perfect, that there are limitations. I like to have everybody share the stories of when it worked and went well, and share those prompts and use cases with other people. But we also tell the cautionary stories of where it failed, so you have the limitations in mind and know what to do to overcome them. That creates a safe space, because without it there could be a fear of, my gosh, I don't know how to use this, and if I do it wrong and make a mistake, that's worse than where I started.

Shamil Malachiyev: Speaking to a lot of people, most of them are scared, because they're reading about layoffs. They think, if I do my work too well, and this is what a lot of people at the cutting edge of AI say, if I automate everything, what am I left to do? I'm just controlling a system that runs by itself. How do we make sure people see AI as a friend, a tool they can use, instead of all this marketing large companies throw out about AI replacing everyone, AI-first companies that are one person and all these agents?

Jasmine Singh: I'm going to start with what's nuanced in the legal profession in particular: we have ethical obligations with respect to how we can and should practise law. One is the duty of competence, which in California requires that we review the output of AI-generated content. Two is a supervisory duty, that we have to monitor individuals or agents creating work product on our behalf. So we cannot entirely outsource our profession to agents and say, I don't need to look at this anymore. At minimum, what I continue to emphasise is that lawyers who are able to exercise their judgment, their curiosity and their experience in reviewing AI output are going to continue to rise to the top. I don't see a world in which those lawyers don't have jobs, because they have to evaluate whether the output is accurate, correct for the circumstances, right for the deal at hand.

They're also the ones, to your very good point, building and monitoring the systems. Someone has to say, for marketing content there's going to be an agent, and that agent has to review the output against this set of laws. When the law changes but the model doesn't get updated, somebody has to manually upload those new laws or regulations, or whatever changes to the product might have happened, into that agent's review patterns. Without a person monitoring the system, there's no way it can keep up with the pace of the time. Those two things continue to be true.

And one thing I really don't want underemphasised: there is no such thing as right legal advice in the abstract when it comes to business. It is the best legal advice for the circumstance. Only individuals can know the full circumstance: we've got to get a deal done by tomorrow, the counterparty already rejected X, Y and Z, I know the lawyer on the other side is sensitive to this argument, so here's how I'm going to position my latest redline to make sure we get to agreement today. That's really hard to outsource to AI, because there's no way it can understand the human emotion involved in closing that deal.

Shamil Malachiyev: Interesting that you mention human emotion. Thinking of watching Suits, for example, it really emphasises the role of relationships. It's who you know, how you build relationships, because business is still between people; companies are made up of people. Do you see that relationship element ever being replaced by more deterministic rules, or is it still important?

Jasmine Singh: I think relationships will become more important than ever in the age of AI, because the human touch required to deliver the best output still matters. Why does software exist? Why does technology exist? To aid humans in doing whatever they want to do with their agency. Outsourcing certain things to technology, whether software or AI, is still in the name of helping. We are the deployers of it, for now. It's intended to be a tool that we wield, and that tool is wielded by us wanting to achieve certain objectives. So I see human emotion and human relationships continuing to drive it, because they're the source of the motivation. They're the reason those things exist. The secondary component is that relationships help people find and understand their why and their impact more deeply. People want to feel that what they do matters, and if you divorce emotion or relationships from your work altogether, it's hard to derive meaning from it.

That's not to say we won't one day have a world where agents are negotiating with agents to come to terms on contracts. I can see that world. But I can also see a world where agents go back 45 turns and I jump in and say, pick up the call. Hey, our agents are going nuts. We can close this deal right now. We do not need to wait for them to fight this out. Why don't we just close it? Because I'm motivated to get it done.

Shamil Malachiyev: Interesting. If we imagine the future with agents, and I'm in software, so it's similar to what's changing in legal: how do people get into the industry? If you're in a top position and driving the changes, it's fairly easy, because you have all those experiences behind you. What about the juniors, people just passing the bar, graduating from university? What's going on with junior positions, and what would you advise people thinking, I just need an education, my diploma, and my future is settled? How do you see their roles evolving within companies?

Jasmine Singh: I see their roles evolving specifically in a way that leverages the experience, judgment and curiosity I just talked about. Those people have to bring those things to bear to do their jobs better, to build the systems, to wield the technology, to deliver legal advice by way of a process that's better for the circumstance. If what they were doing as a junior associate was redlining contracts, and now technology does that on their behalf, they have to bring the judgment to review those AI outputs and say, actually, that redline doesn't make sense here, or doesn't apply very well here. Or, actually, I need to fix my playbook, because my position keeps getting redlined out of that agreement. I've got to tell the AI what to do. Having the ability to understand what's wrong in a contract and why it needs to be better is the game changer. If you have the working experience to know a good redline from a bad one, if you have the lived experience of working in a business to know the counterparty would never go for that term, it's non-standard, we should not propose it, those things matter now more than ever. So I tell juniors especially: bring those experiences to bear. They're relevant and they're going to set you apart.

Shamil Malachiyev: That makes me think of two things. First, whenever I hear there are huge layoffs because of AI, and then coming to this topic, I keep hearing lawyers are being replaced by lawyers who really know AI, business process automation and a bit of technology. All the juniors are a lot more connected to the technologies. Back when iPhones came out, it was harder for people ten or twenty years older to adopt them, whereas when you grow up with all of these things it's easier. So maybe that's the big shift: people who don't want to touch AI might find themselves doing something else, while juniors come in with the technical skills required for the work. And the second: four years ago you had one contract and two days to redline it, go through it, understand the context. Now everyone says it takes three hours, so we expect you to do two or three contracts a day. The amount of information you have to keep in your head, the number of decisions you have to make, is three, four, five times what it used to be. How do you stay away from burning out with this much mental load?

Jasmine Singh: This is actually a great place for AI to take on some of that mental capacity on your behalf. Using AI as a chief of staff, as a work-tracking system, so it can help organise your inbox, organise your Slacks, keep track of what work you have pending, is a vital way to survive in this world. If you're not using AI for anything else, at least use it to organise the massive amounts of work you have to do. The second thing I'll offer, and obviously this is because I work at Ironclad, is that platforms like Ironclad let you log into your dashboard and see all the contracts that are pending and their status, go into the audit trail and see the last activity on this contract. Having a place to go and say, I don't have to remember every single one of these myself, I can go back to the system and know who was involved, what happened when, what the ultimate document was, what changed between version two and version six. If we didn't have platforms like that, it would be impossible to keep everything required of us in our minds at all times.

Shamil Malachiyev: Can we talk about Ironclad? Most viewers and listeners will have heard of Ironclad. What was the platform like three years ago, when the news came out? What did the company try to implement to change, how did they see the platform evolving, and after you joined, what specific changes have you seen that you like the most and see the most value in? Maybe we can touch on the vision as well.

Jasmine Singh: Three years ago the platform was this amazing end-to-end tool to facilitate the acceleration of contracts at every level. Not just intake more easily and quickly, but legal review, execution, storing in a repository and getting insights. The supercharge over the last three years has been optimising each of those steps to make them better, faster, cheaper. We started on the negotiation front with our Jurist tool. In the last few years we've been working on how to support commercial lawyers in all the tasks they have to do around contracts, whether that's risk review for triage, an understanding of what obligations are in an agreement, summaries of agreements, plain-English translations, comparing versions of contracts. Jurist does all of this. We've really tried to give lawyers access at their fingertips to more features and products that let them do their work a little more quickly and easily. And the second part of your question, actually, I'm going to need you to remind me what part two was. I'm sorry.

Shamil Malachiyev: When you're evaluating, okay, we have these things AI can do now, how do you know which ones to use? There are so many things you can do.

Jasmine Singh: Yes, great. To decide what AI use cases to optimise and offer to our users, we really thought about the friction points. We've deployed agents at those various friction points, or we're in the works of deploying them. One is AI Assistant, which lives next to your Ironclad instance at any time. You can type in: find me contracts that have this condition true, or find me contracts with this counterparty. Having the ability to prompt your repository in natural language to give you results faster, easier, better. But also deploying an agent at the front end to help intake contracts. Historically, when you have a contract for review, you go into a workflow, you manually fill out all the questions, you attach the contract and submit it for review. Now, with things like the intake agent, you just upload the contract and all of those questions are answered on your behalf. All you're doing is validating whether they're true or false, and then it gets submitted. Then a tool like Jurist auto-applies a playbook based on conditions that are true in the intake form, proposes redlines to an attorney for review, and pings when they're ready. And now suddenly the process of intake to redline is a minute when it used to be potentially days, right?

Then a contract gets reviewed, negotiated, executed. What we also know is that where you started with the contract and the intake form may not match the contract when it's fully executed. The data may have changed. So we've also been working on something called archive agent to update the data fields, so if you need to glean intelligence from that contract later, you can. By way of natural-language search, the data in the repository is parsed differently, so you can find the more global view of what you're looking for when you ask your repository whether certain things are true about your contracts. So we're really trying to reduce the friction we've seen in the process and in the experiences of our customers, on the AI side in particular.

Shamil Malachiyev: When developing these technologies, in all industries, from manufacturing ERP systems to anything, whenever you give people AI features, they say fantastic, they make the job so much easier. But actually persuading people to start using them turns out to be quite a challenge. People don't like change. When you do something you know, it's comfortable. When something new comes in, it's risky. How does Ironclad try to help people go through and make that change?

Jasmine Singh: Part of change management, whether internal or external, is painting a picture of what's possible and explaining why certain things are true. Here's why we created this feature, and here's what we think it will enable by way of a different, better future for you and how you work. Once someone understands what might be possible and why it's being offered, step two is: now let me teach you how to use it. Make it easy by way of support guides, documentation, training sessions, live meetings where you get together with other customers doing similar things, live sessions where you hear from Ironclad's amazing customer success and customer support teams on how to optimise and use certain features. Enabling people with the information and skills they need to go achieve what we've painted for them as possible is the connection to how we encourage people to iterate and change.

Shamil Malachiyev: What parts of the platform are you personally most excited about?

Jasmine Singh: So many, but what I'll tell you about in particular is the prospect of what we're working on with Jurist, especially what's on our roadmap this year. Jurist is a redlining agent, and I've talked a lot about that. What we're trying to do is leverage the rich data sets we have from customers to make the redlining and negotiation process seamless. Earlier I mentioned intake forms are where the circumstances about a deal live: who the vendor is, if it's a vendor deal, how much the contract costs, what exactly you're getting. Lawyers also need to know whether you're sharing PII with a counterparty, whether you're giving them access to company systems, things like that. That context, that awareness of what's happening in the deal, actually drives redlines and drives the application of certain provisions in playbooks. So being able to connect how Jurist applies redlines to that contextual awareness. And secondarily, based on an understanding of your own company's precedent: how have you contracted with counterparties like that before, in that industry? How have you contracted with that exact counterparty before? What do your own contracts say about how you like to agree to terms or where you push back? I think those things are going to change the game for commercial lawyers in negotiating, because it won't just be, let me remember the last time I did a deal like this and try to come up with the same redlines. The tool and the technology foster the ability to do that in a scalable fashion.

Shamil Malachiyev: Looking at so many changes going on for people in companies, you have your own story of burnout in big law. I can imagine that with trying to keep up with all these technological advancements, there are a lot more people going through something similar, when expectations are tremendously high, when they have to start dipping their feet into areas they didn't touch before, like technology, business processes and systems. Using the example of how you took time for yourself to refine yourself and come back stronger, could you share a bit about that personal process and journey?

Jasmine Singh: Thank you for asking. What's interesting about my experience taking a break and becoming a spin instructor is this. When I first did it, I thought I had ruined my professional career. It wasn't as though I was confident I'd be able to come back and get a job in-house and know exactly what to do. Instead, I left the practice of law believing maybe I'm not a very good lawyer, maybe I chose the wrong profession, maybe this isn't for me. So for the first time I allowed myself to think about what is for me. Not checking something off the list, chasing the next academic or professional success, but what actually brings me joy. I had never been a fitness instructor, though I was an athlete and a dancer. I said, what do I love? I love fitness. So I literally showed up at a spin studio and said, I love fitness, I'd love to teach a cycling course. I'd been a road biker in the Bay Area, so I was pretty good at it, and I was a dancer, I love music, I think I could do this. Thankfully the studio had an in-house training programme, and they said, go through it and we'll teach you to be a spin instructor. What I knew about myself as a litigator held true as a spin instructor: I believe I can learn anything. I can figure anything out if I put in enough effort, time and dedication.

So I learned to be a fitness instructor, and for the first time I felt real joy in my work, and I was actually being paid for it. It's not that being a litigator didn't have moments of joy, but every day people were high-fiving and saying thank you, you changed my life, I feel so good about myself. That was a different way of living. What I realised was that my essence requires a job where I feel like I'm building something, helping people, delivering something of value to the world. And the second thing that was true: I continued to hone this discipline of showing up every day and not giving up. So I knew that if I didn't go back to the practice of law, I'd feel like I gave up. So I leveraged my community. That's the second thing I'll offer. Joy and wonder guided me, and community helped me find a way back in-house. I said, I want to try being a lawyer again, but in-house, as a transactional lawyer, not a litigator anymore. I talked to every single friend, every contact willing to talk to me about their jobs, anybody willing to give me an informational interview, advice, look at my resume, pass it along. It was a former coworker who referred me into 24 Hour Fitness, and that was my window back in.

I'll tell you, it was not an easy path back. It's hard enough to move in-house from litigation to commercial work. It's way harder if you're a spin instructor living in Las Vegas. Nearly impossible. I made that shift because I had the conviction of my connection to the mission of the company, so I could tell that story, and because my community went out on a limb for me. I give you that example to say that's how I made it through that moment, and that's how I continue to make it through every moment of shift in my career. I have to have a connection to the mission. I have to care about what the thing can offer. And I have to rely on a community of people to get me through. AI is no different. I truly believe in the power and potential of AI in helping lawyers practise more efficiently and effectively, and really demonstrate the incredible value attorneys bring to organisations. I think we can magnify it by way of AI. I'm in the camp that this is going to make us even better. My goal in life is to find a way to help legal AI transform the lives of lawyers everywhere. And it's not easy to use this technology or know everything about it. There's too much out there. So I need to ask my peers, my friends: what have you been building? What workflows have you disrupted? What tool did you use for this? Compare your use cases, compare notes. There are always ways to optimise, and community is the perfect way to do that.

Shamil Malachiyev: Two follow-up questions. Whenever I talk to somebody who's really far along the journey with AI, they all think they're behind. The more you know, the more you think you don't know. You seem a very self-confident person, which is awesome. Do you ever get imposter syndrome, when everybody's feeling a bit behind? How do you keep your confidence up while navigating it?

Jasmine Singh: Totally. That question exists with respect to AI, and frankly with respect to my profession. Yes, I feel that with AI all the time. Though I know a lot, though I'm doing a lot, though I've learned a lot, I still sometimes feel I'm not at the cutting edge, that there's more for me to do, that I'm behind. The way I square those two things is I just keep trying. I never give up. I use that same ethos that's so buried within me: I will get up every day and try my best, I will learn what I don't know, I will try new things I've never done before, I will fail, and I will get right back up and do it again the next day. That relentless pursuit of knowledge is what helped me overcome the imposter syndrome. Though I don't know something today, I sure as hell will know it tomorrow. I'm the person who will go find it out and figure it out.

That translates to my entire career. I've always had this tension of, I'm an imposter, I tricked everybody, I don't know how I got here, but I did, and this deep-seated confidence that I can do anything. Those two things exist within me every single day, in tension, and I have to remind myself that the voice trying to tell me I don't belong is also probably part of the reason I'm so successful, because I'm constantly trying to quiet it. I'm in therapy to deal with that. It's part of the emotion. How do you control that voice and harness it at the same time?

Shamil Malachiyev: How do you make it your superpower? You also mentioned the community that helped you get back on the horse. For people within companies, that's a valid point: if you want to get better and feel you're not as behind as you think, it helps to create a small community within your department or company. Have you done that, creating a small AI council of builders sharing what works and what doesn't?

Jasmine Singh: Yes, in a few different places. One is within my own team. Every single week in our team meeting, people present their AI use case of the week. What are they using AI for that week, what prompts worked, and, as I said earlier, what's a fail? What didn't work, what can the rest of us learn from? We also have an AI course syllabus we created for our own team. We did our own research to figure out the skills we need to learn in the next three to six months, and the classes, course material or information we can ingest to do that. So we set up a syllabus: this month you'll learn about prompting, next month MCP, the month after developing agents or vibe coding, whatever it might be. We've set out a path of what we're learning, and then what we have to do with it to put it into practice. By a certain date you have to have built a named agent or connector and disrupted a certain workflow. Our little team is an AI test case: how are we changing the way we practise and deliver legal services to our customer, which is Ironclad, by innovating with AI?

We have the benefit of partnering with our incredible EPD team to do the same. When we have ideas, we go to our engineering and product partners and say, we're having this problem, we have this idea, we think it could solve it, how can we partner with you? So we get to extend that to engineers and product folks who can bring it to life. And I'm a member of a few GC groups and women's groups where we have AI circles. How did you build this? How did you navigate this complicated thing? This isn't working for me, can you give me advice? We try to meet regularly to talk about those things, and I've found that really useful.

Shamil Malachiyev: What's the best way to find those groups? I feel like everybody needs an AI support group. Claude Coders Anonymous in their region. What's the best way to find the right group for you?

Jasmine Singh: For me, start with the groups you're already part of. If you're part of an affinity group or a bar association, ask whether there's a subcommittee on AI. If not, start one. Or just start an informal group of people looking to talk about AI. Start where you already are and see what exists within those groups. If you really can't find anything, find the organisations you want to be part of that likely have AI components. If you're a GC of a tech company, L Suite is an amazing place to go. If you're an Ironclad customer, you can join our online community. There are so many places with existing groups of people who do the work you do, with subgroups specifically focused on AI.

Shamil Malachiyev: What are some of the coolest use cases you use on a personal level?

Jasmine Singh: First and foremost I use Jurist every single day to redline contracts. I've spoken about that ad nauseam. I think it's amazing to see our company's guardrails put into practice and redlined on live contracts. Some other use cases: I use AI all the time for synthesis and quick summaries. If I get added to a Slack channel that scrolls back for months, I'll drop that channel into Glean and say, give me a quick summary of what's happened, where the issues are, I'm being looped in to answer this question, what information in this channel is relevant to my ability to answer it. I do that version of synthesis with basically everything. Issue spotting: a customer says, I want to build this feature, how do I do it in accordance with whatever the applicable rule might be? I start with AI always. This is the feature we want to build, this is the regulation at issue, what questions should I ask to figure out more about how it might comply or not comply? Using AI as a thought partner is a really strong use case for me. A synthesiser, an intern. And a sparring partner: I'm about to send this email to my executive team saying X, Y, Z. Find all the holes in my argument, then draft what addressing those holes might look like.

Shamil Malachiyev: How often do you have to remind AI to replace dashes? I know dashes are a favourite in legal departments, but now, for it not to look AI-ish, do you ever have to say, please remove them?

Jasmine Singh: I feel attacked, as someone who has used em dashes my entire life. Now I'm like, no, that is not a signal that AI wrote it. All lawyers use dashes. It drives me nuts. I'm very sad. But yes, I do have to remove them, because otherwise people assume it wasn't me that drafted it. I'm not shy about using AI. I just don't want people to think I didn't revise it or make it better before I sent it to them.

Shamil Malachiyev: A lot of listeners run their own departments. In conversations with managers on my teams, I'd say we need to start pushing people to use AI. And one came back with, Shamil, we can't push people, we can ask them, but leave them to make the decision. But as a leader I feel I need to start pushing, because in a year's time it's going to be my responsibility, and if you're not up to speed, I won't have a choice but to find somebody who is, because there are going to be a lot of people doing it better. How do you go about that tough love, helping people adopt the right things at a fast enough pace? You're a competitive, sporty person. Other people on your team might not be.

Jasmine Singh: I do it through two things. One is the expectation, the requirement. Everybody on my team knows my expectation, my requirement, my express ask is that you use AI to do your work, in a way that is ethical and complete, exercising your judgment. My expectation of every single person on my team is that you're using AI to do your job. I make that clear. But I couple it with what I'm calling enablement material. I'm not just saying use it, and if you don't you're in trouble. I'm saying use it, and let me give you access to all of these things to help you use it and understand why using it matters. That's why we built the course syllabus. That's why we have an AI charter setting out our expectations on usage. I'm enabling them to use it in a way that's compliant with our rules, how we think about AI, how our team thinks AI can best help us. And I'm giving them access to information so they can use it better, and freeing up time. My expectation is also that you do this course, an hour a week or whatever it is. And then I'm modelling it, and they're modelling it for each other, by coming to team meetings and saying, this is what I'm doing. People get a little FOMO if they're not doing it. They think, my gosh, that's so cool, I want to do that too. It's a combination. If all you're doing is setting the expectation, I don't think you're doing a good job managing the change necessary to push towards AI.

Shamil Malachiyev: I guess you're also leading by example.

Jasmine Singh: Trying my best, yes.

Shamil Malachiyev: What other things make you think we might be living in an AI bubble? That question keeps popping up in my conversations over the last four months, especially when you see companies like Harvey and Legora reach valuations of ten, twelve billion dollars really quickly. The industry is trying to move really fast, but when you look at adoption, people need time to get accustomed to the tools. In your personal view, do you think we're in an AI bubble?

Jasmine Singh: I'll start by saying that for legal tech in particular, there's a real need for the application of AI to what we're doing. It's a high-leverage scenario. So I don't think the need for AI, or the ability to apply AI to legal work, is overblown at all. I don't think that's a bubble, because we're already seeing the value of applying AI to our work and how it can change things. But to the meta point, whether there's an AI bubble around the value of this technology in general, my answer is no, because the tool continues to iterate and prove its value and its potential delivery of success. It's what we do about how AI performs that will dictate whether the decisions we make in relation to AI are the right ones or the wrong ones. We're still early in the innings. We're still figuring out what AI can do, how to use it best, how to get the right return on our investment. So it really depends on what happens in these next months and years around that return on investment and the value to our day-to-day work.

Shamil Malachiyev: As a last question: personally, are you more anxious or excited about the future AI is getting us into?

Jasmine Singh: More excited. That's not to say we don't have to exercise care and diligence in driving toward that future. I absolutely think that's necessary. But the reason I'm excited about what AI has to offer is that it can help eliminate some of the rote work, the administrative labour, the minutia that create burnout in our profession especially. And having someone else, in a profession that can be really lonely, is really useful. I say someone else, but I mean going to an agent and saying, gut check me on this, what's wrong with my argument, ask me the questions somebody else might ask me. Noodling through things with an agent that gives you an instant response can make the feeling of practising law less isolating, less lonely. So I'm really excited about what AI can do for us as lawyers and for how we practise law. It requires a lot of discipline and care to make sure we don't lose sight of the very important personal aspects of the practice we talked about earlier.

Shamil Malachiyev: A really cool perspective that I haven't heard before, alleviating loneliness. Otherwise you have to spend so much time with just a screen and a lot of text, thinking through, and now you can almost have an assistant or a buddy helping you move through it. I want to thank you so much for the conversation today. This has been an absolute pleasure, Jasmine. I want to thank you for sharing with our listeners. I think they got a lot of gold nuggets to take back into navigating this changing world of AI. Thank you so much for taking the time to come on the show.

Jasmine Singh: Thank you so much. I really appreciate you having me. This was a pleasure. Thank you.