Ex-Intelligence Officer: 80% Of Legal Work Could Be Done By AI | Bennett B. Borden

Before he was advising boards on AI strategy, Bennett Bordon spent eight years as a U.S. intelligence analyst, then went to law school, then earned a graduate degree in data analytics. That cross-disciplinary background eventually led him to analyze a decade of billing data from 14,000 attorneys, and conclude that roughly 80% of what lawyers do today is better done by AI.

Bordon is now the founder and CEO of Clarion AI Partners, a boutique firm that advises companies on AI governance, risk, and strategy while also building the AI systems themselves. In this conversation with host Chris, he breaks down:

• Why he left a top-10 law firm to build Clarion instead of trying to "turn the Titanic" on the billable hour
•  The "Iron Man, not Terminator" framework for thinking about AI and human work
• How the legal concept of reasonableness — what did you identify, what did you do about it, can you prove it — keeps companies defensible in an unsettled regulatory landscape
• Why 80% of AI adoption is a psychological and change-management problem, not a technology problem
• How AI governance committees go wrong when they're owned entirely by tech, compliance, or the business alone
• The work Clarion does on child safety, access to justice, and public policy as AI scales globally

A wide-ranging look at what it actually takes to deploy AI responsibly, from the boardroom down to the workflow level.

Find Bennett Bordon at clarionaipartners.com or on LinkedIn.

LinkedIn: https://www.linkedin.com/in/bennett-borden/
Website: https://clarionai.com/

Find Us on:
Youtube: https://www.youtube.com/@SignalRoomPodcast
LinkedIn: https://www.linkedin.com/in/chutchins-healthcare/

Transcript

SPEAKER_02 0:00

About 80% of what a lawyer does today is better done by AI. Generative AI is an existential threat to how we practice law. I'm a huge AI optimist. You can control the behavior of AI systems perfectly. If you can see governance principles, especially down to the actual application, like how I'm using, building, deploying AI, it speeds up innovation, it speeds up your transformation. And governance has always been like a ugh governance, co-pilot, chat GPT, Claude, whatever, right? How do you use that? How do you not use that? And that's really great. Like using these broad enterprise level tools are good. You will get really decent efficiency gains, 20, 30% across all kinds of workflows. The problem is welcome to the Signal Room. Thank you very much, Chris. I'm delighted to be here.

Chris Hutchins 1:01

Before we go too far into some really fun issues, you know, when we're talking about governance, risk, compliance, and all these other things, I want to kind of start with something that caught my attention right away when we we first connected. You've talked about analyzing years of legal work and finding that roughly 80% of recurring attorney tasks could potentially be handled more effectively through AI assisted workflows. Tell me a little bit about that analysis and what you discovered, because I think it really speaks to the amazing work that you've been doing over the last several years. Really kind of ahead of the curve from everything that I I can see. What what did you discover?

SPEAKER_02 1:39

Thank you. Yeah, you know, I I spent my career, as you said, I started out in US intelligence and was there for eight years before going on to law school and then getting a graduate degree in data analytics. And so I've been at this cross-section of the law, technology, data, how do you get insights out of data for all kinds of applications across just about every industry you can imagine? And I was at my prior law firm, I've always spent my career in kind of big law, you know, kind of one of the top 10 firms. And I was uh chief data scientist at a partner and and helped lead the global AI practice at this firm. And it's when Gen AI came out and and we were able to represent those companies. And I had represented, you know, Microsoft and Amazon and Google for years and years. But this was an entirely new technology. This is something that the world had never seen before. And uh the head of the firm asked me to do an analysis on what I thought the effect of generative AI would be on how we practice law today. And luckily we had this massive data set, right? Because lawyers write down what we do every tenth of a minute or an hour, right? And so we had we took 10 years worth of 14,000 attorneys' time entries to analyze what are the steps that lawyers take to put together their legal product, whether they're a litigator or contracts or you know, investment documents, whatever they were. Um, and it really turned out that every lawyer, no matter what their practice area, does six or seven things over and over and over again. And very conservatively, about 80% of what a lawyer does today is better done by AI. And so it was interesting because here I was, the pinnacle of my career, you know, kind of at um, and two things occurred to me at once. One is that generative AI is an existential threat to how we practice law traditionally, that throwing bodies at a problem and charging by the hour, that is just just simply it's the same way we did in the 1700, right? Yeah, and so I decided that I didn't want to spend my the last 10 years of my career trying to turn the Titanic, right? Because the billable hour is deeply embedded into law firm culture. It's the metric by which everything is judged. But also, this new technology can at the same time utterly transform how we practice law. And so I decided to take some of my team and come and form Clarion, which is a very unique little AI boutique law firm. I was very fortunate that a great number of my clients came with me. And so we give advice on AI, you know. So a custom the client will call up and be like, hey, we're thinking about doing this, and we talk through the uh risks and uh how to mitigate those risks, and is that acceptable? And you know, but the things I think the two things that I love the most about what we do are one is is AI strategy, like really helping boards and C-suites decide how to decide to use AI and really taking it from this policy layer, you know, uh down to how do I operationalize this and how do I decide where to deploy AI in my company. And then we're a dev shop. We literally build AI-based solutions for clients, and that being at the cutting edge of how do you design point solutions and agents and multi-agentic systems with orchestration layers. And I I think one of my favorite things to say to CEOs is that you can control the behavior of AI systems perfectly, and they always that that's that's a very provocative statement, right? Like they're they read all the headlines, all the you know, people have gotten into trouble and AI going kind of crazy, but that just means they didn't build it right. Um, and so really unlocking the special sauce that makes a company distinct through this utterly almost magical technology is what really drives me.

Chris Hutchins 6:07

That's that's quite a a transition you you you you've made. It strikes me just listening to you for the last minute or two, love that hearing a lot of terms that I typically would expect to hear from a from a lawyer. Thank you. Uh I think your career trajectory is fascinating to me. I mean you you mentioned you know work in the intelligence uh space and really establishing data science practices and things like that. What are some of the things that you you learned there that helped you really help you to think about how the technology can be be brought to bear in a in a way that's really, really data supported and driven and helping people to really understand the utility of these tools and technologies? I mean, all the buzzwords are just they're just that, unfortunately. I don't think a lot of people really have a high level of understanding. I don't think you can get it being a hobbyist.

SPEAKER_02 6:59

I think I have a somewhat unique perspective on data just because of how my career developed. I had no intention of being a data scientist. In fact, you know, uh in at university, I had kind of put together my own degree studying turning points in history. Like, how do societies go about forming themselves and why do they have one form of government over another? How do they decide to go to war or have a revolution or respond to economic or ecological pressures? And uh specifically really focusing on technological revolutions, everything from uh the wheel to the saddle to gunpowder, right, uh all the way through to the second industrial revolution. And because each of those things had just tremendous impact on the societies that they affected. And so when I was approached by the agency to come work for them, this is the early 90s, right? So 1992. Right. And they said, hey, we're we're setting up this new kind of data analysis shop, and we want you to come work with us. And I'm like, I'm not sure you've got the right guy. Like, I don't really do uh data. I'm not particularly fond of math, I hate statistics, I failed calculus, and they're like, Yeah, we know, but but what we want you for is to understand movements in societies. And so because I had studied things like history and philosophy and economics and sociology and anthropology, and and I had a view that data is a reflection of human conduct. And so if you'll remember in the early 90s, what was going on is the World Wide Web and email was just coming out of academia, cell phones were just coming on the market, and so we were we were creating a new kind of data that really didn't exist in the world before, a very personalized data contrail that individuals were giving off. And the agency rightfully saw that that that contrail has only become more voluminous and more enriched of a different variety as time went by. And so I was responsible for was helping take the incredible data talent of these people and look for signals, you know, who's a good guy, who's a bad guy, whatever that meant at a particular time. Could we understand their behavior? Could we predict their behavior? And could we influence their behavior? Now, this is what every social media company, every cookie, every app, everybody does it now. But at the time it was pretty cool technology. Because I came out of the kind of studying all the ologies, um, I've always felt that data is a reflection of humanity, like how we go about being humans. And so it's it's given me a very different perspective than someone who was raised tech or someone who was raised lawyer, right? It's kind of the combination of those things that I think has been the unusual part of my career.

Chris Hutchins 10:06

That is amazing. I I never really would have uh imagined that you know they recruit the way that they they do. They clearly had a good idea who you were and what you're all about.

SPEAKER_02 10:16

Oh, they keep a pretty good eye on the universities and and who they might be interested in.

Chris Hutchins 10:22

Yeah, well, it's I I'm I don't know. Like I said, it's it's fascinating. I'm I'm really kind of psyched that they found somebody like you because uh you know that that gives you a perspective. I think right now is pretty pretty well needed everywhere. And I know that there's a lot of executive teams that are under a lot of pressure to deal with this AI transformation and get it right. I don't know if I've actually met a whole lot of people that really understand the differences between the different types of AI that we're we're dealing with. You know, you you kind of made a transform transition you know from doing the data science thing and doing the law thing, and you're you're doing some advising and things like that. But what are some of the things that motivated you to really go down the pathway of starting to do some development in building? You know, I understand what you're talking about in terms of the the the shift in the I guess we call it the money model, going from billable hours to some other thing. But I I'd love to hear a little bit what what really drew you into the to this space where you wanted to do some development as well.

SPEAKER_02 11:27

You know, honestly, it was my frustration with constantly being the interpretive layer between the business, the tech, and the lawyers. Right. Because I found that, you know, like what we do here at Clarion, we you know, we are AI advisors, like I said, but this AI strategy, like I love helping a company take itself apart and figure out what it's really about and where to bring AI in. But the development is one of my favorite parts because it's taking the years, sometimes decades or even centuries, of experience and data that lies dark within the siloed areas of a company and finding artifacts of its speciality, right? Like what makes it truly unique and being able to harness those things to create true transformative solutions, right? And it's interesting because the tech people are very talented, right? And they know how to build a function, and but they they aren't necessarily trained in the risk side, right? Like on identifying how this could go wrong or or especially legally. And the lawyers are flummoxed because uh you know the lawyers really, really like to have an one answer, right? That there's they can point to a law or a regulation or a court opinion and say, that's the answer. We don't live in that world with AI regulation right now. It is all over the place and it's gonna remain very unsettled for years. And so being able to communicate to a company that you can absolutely take advantage of AI properly, control it perfectly, and be defensible regardless of where regulation heads. And so that's what really motivated me to create this little company, this little strange animal that we are that are lawyers and also advisors and builders, because it brings together these disciplines that are so essential to make really great decisions about how to use AI.

Chris Hutchins 13:42

I love how you explained it. I've been thinking a lot recently about the end-to-end solutions that are being designed and deployed in different places. And you're talking about some of the connective tissue in between major areas that people aren't seeing and they don't necessarily even know about it. It's like the relationship of one particular data silo to the one next to it.

SPEAKER_00 14:06

Exactly.

Chris Hutchins 14:07

You know, I I could see where this is a really uh important asset to have available is people that really understand how to observe those things and how to put the the right controls in place. Um we have heard governance tossed around a lot, particularly since the beginning of the year. I it seems like I'm hearing it multiple times a day. I don't I don't know that everyone really understands what it is, but I'd love to hear your your how you would define it. I mean, what does it actually mean in in reality to you?

SPEAKER_02 14:39

So there's there's many different layers of governance, right? And and so if you think about all the way up to the board level, right, the board has a responsibility to derive sufficient information from management to guide the strategy of the company, to know that the company's in compliance, that risks are being understood and addressed, right? Management is meant to carry out that strategy. And so it's beyond policies. Like we the very best kind of form of governance that I have found in in larger institutions is like an AI working group or an AI committee. And the reason that for that is be is because AI doesn't naturally live in our current structure of the C-suite, right? If it's lives in tech, it's too tech forward. If it lives in compliance, it's way it's this massive breaking mechanism. Um, if it lives with a business, it has a potential of being really transformative. But you have to have all of those perspectives. And so creating a AI working group, a committee, a steering committee that's properly represented from all of those perspectives, because all of those perspectives are valid, where we find the problem is that when that group is unbalanced. So if you've got it led by the business, sometimes the risk and the tech side are less represented. If it's compliance led, then all you're talking is about risk and not opportunity. And so it's really important to get that group right and who sits on it and how it does its business. And it's important to have policies, right? Like we will do this, we won't do that. This is kind of our ethos around how we use AI. But all of that is just table stakes, right? It's just the beginning layer because what you really need to do is have a way to understand and evaluate the risks and the benefits of a particular AI use case. And so this can be anything from enterprise-wide, so co-pilot, chat GPT, Claude, whatever, right? How do you use that? How do you not use that? And that's really great. Like using these broad enterprise level tools are good. You will get um really decent efficiency gains, 20, 30 percent across all kinds of workflows. The problem is that those tools are available to everybody. And so you'll get this first mover advantage, but basically then everybody catches up, and when everybody's special, nobody's special, right? And so what really is critical to taking advantage of this almost magical technology is to realize that just like we figured out in practicing law, there are certain tasks that are better done by AI, and there are tasks that are essentially human. And so I say companies do stuff, right? And they do stuff in a series of steps. And those steps are made up of tasks and a workflow. And so figuring out which one is better to my AI, building a point solution, which is very easy to do, put the controls around it, prove that it's reasonable and that it's acting the way it should, handle the handoffs between the human and the AI, right? And all of a sudden, you are we call it supersuiting your company. And so we use this analogy all the time, right? Think the Iron Man, the Terminator, right? So you're not gonna unplug a person and plug in an AI solution and it take over all the work. That's just really not typically how it works. What you're doing is taking people who you have spent years training their experience, the experience of the company, and that's hidden within the people, but also within the data, and you're gathering that together and building these point solutions so that all of a sudden everybody's got an Iron Man suit or a Superman cape, right? Like, and so if you can imagine augmenting and extending the natural acumen that you've developed in your company, wherever that acumen lies, that is how you truly transform. And that gets us back to the governance, right? Because you have to be able to understand the risks that you're cre introducing by using AI in a particular situation. How do you reasonably mitigate those risks? How do you prove it? And it's that proving layer that will keep you safe and defensible and agile regardless of where regulation ends up.

Chris Hutchins 19:41

This is a a challenge that I've I've seen people struggling with. And it's it comes down to how do you get um a common understanding um from all the disciplines that you need to have in place to do this governance activity. People have committed their entire careers to their professions, to the areas that they're they own, but it's not necessarily a natural leaning from for people to want to get involved so heavily in data now and in AI technology, it's kind of complicated. I mean, and I imagine that's probably one of the challenges that you you helped to solve, right?

SPEAKER_02 20:21

Yes, and we always start the same way. Like we meet with board and or the kind of C-suite and get everybody on the same page. What do we mean by AI? What does AI governance look like? What regulatory rubric are or industry are you in? What controls already exist? Like it it surprises me so many times when I talk to management and they'll say, Man, I feel such pressure to do something with AI. And but I just don't know how to think about it. And I'm like, of course you do. You run a company and that company is in an industry, the industry has laws, regulations, industry best practices, whatever it is, and you already have controls in place to monitor those things so that you're in compliance. So you're not inventing governance out of whole cloth, right? You're the questions are if I introduce AI into this particular workflow, what how does my risk profile change? What new risks am I introducing? Can I reasonably mitigate those risks? Which of course you can. You know, AI being a technological solution, it means it can be technologically controlled. And then the the part that everybody leaves out is while you're building these systems, build in compliance by design. So that like when we build a point solution, we build in sensors for lack of a better term that measure the stuff that matters and spit out metrics to prove. That you're following the law or the rules or industry practices or the ethos of the company, whatever it is. And so at any moment you can prove to courts, to regulators, to the court of public opinion that these solutions are working precisely as you designed them to.

Chris Hutchins 22:22

The tendency that I've that I've noticed is that I've worked primarily in healthcare for the majority of my career, but governance has been more of a an academic kind of exercise that people usually get approached with long before there's something to be governed, they eyes gloss over, and then you know people just don't want to be involved. But now we're in a very different place. And I think what you're doing is uh incredibly important. Um one of the things that you've mentioned a couple of times now is you know the defensibility aspect of it. You know, I think that's an important part of the work that you're doing. Talk a little bit about what you know what that looks like in practice to you.

SPEAKER_02 23:04

So the thing to understand is that this technology is so advanced and is advancing so quickly that regulation and law just simply cannot keep up. Right. It's always the case that the law is slightly behind what we can do in the real world. But when the pace of technological change is so fast, we there's just almost no hope. And so you you come up with these broad-based principles, you know, fair, accurate, reliable, transparent, but it's really difficult to get beyond that abstraction layer down to specifically how do I put that into practice, right? And so the the idea of governance, I try to help leaders understand that it's like when you're driving down a road. Like I live in northern Utah and there's lots of mountainous, round, why windy roads with cliff drop up, drop-offs and all those things. And it's the governance that actually makes you allows you to go faster and move more nimbly. If there weren't lines on the road, if there weren't warning curve ahead, if there weren't speed limits, if there weren't guardrails, you would actually move slower through that environment. But with the correct governance pieces in place, it actually protects your momentum. It allows you to move faster instead of jerking and starting, and oh, what's around this corner? And oh, I gotta stop and look at it, right? And so if you can see governance principles, especially down to the actual application, like how I'm using, building, deploying AI, it speeds up innovation. It speeds up your transformation. And governance has always been like a ugh governance, right? Like nobody wants to hear about going to another governance training. Um, but this is where it really matters because there the law is very unsettled. But luckily, there is this principle in the law that all our Western society's legal system is based upon is this idea of reasonableness, right? It's a common law principle from the 1700s that basically the the reasonable person standard, the reasonable company standard that says what would a reasonable company do based on the information they had at the time? And what we see coming out of the federal government in the US, from like the EU AI Act, from state government legislation, is this idea that what that means is when you're using AI, what reasonably foreseeable risks did you identify? What did you do to reasonably mitigate those risks? And how can you prove it? And you can use those three questions on every whether it's a point solution, an enterprise solution, and whether it's an internal facing or external facing. If you fall back on this idea of reasonableness, you can build AI multi-agent systems with orchestration layers, always coming back to did I act reasonably? And that will always keep you out of trouble with courts, public opinion, the regulators, whatever it is.

Chris Hutchins 26:24

That's really an important piece of what organizations really have to be thinking about. You know, and I I know there's a lot of push for efficiency, and you know, there should there's definitely some massive opportunities there. But what one of the things that caught my attention is is something that you know it's it's been kind of an important part of my own career path. And it's really because of the of the ability to try to make things better for people. You've done a significant amount of work in things that are more public interest to human impact related. Uh, some of your most important work has focused on issues that I care about. I'm sure all the listeners care about. I'm talking about child safety, suicide prevention, public policy. Uh talk about some of the work that you're you're doing in that space. Because I think it that to me that's a really unique layer uh that that you have that you bring to the table. We're talking about technology and law, and this is so many different parts of the society around us. But this is this is a whole different thing. I'd love to hear about this.

SPEAKER_02 27:25

Thanks, Chris. It's it's one of my passions, and it's one of the things that Claire Anne works on quite a bit because this really is the world's most transformative and disruptive technology that we've literally ever experienced. And there's not an area of our society that it's not going to impact, and we are just at the very beginning of what that impact is going to be. But it has tremendous potential in education, in life sciences, in pharmaceuticals, in access to justice. And, you know, for the first time in human history, we have the ability to place world-class education at the hands of every child on the planet with AI. And but we've got to get that right. And so I've done, we do quite a lot of work with state and local governments, with K-312, academic institutions, higher ed, lots of access to justice. Um, how do you bring AI to make courts more efficient, to make the justice system more available? But especially around child safety. We are, you know, within the next few years, 75% of the workforce is going to be millennials or younger, people who have grown up in this very technological screen-driven attention economy. And that has had some really negative effects. Um it is not a coincidence that the world's most connected generations are also the loneliness and have the largest mental health crisis. And so working around public policy on how to make this technology different is incredibly important. And the the piece that's missing, and this is one thing we work on a lot around the world with consortia and responsible AI groups, everybody kind of agrees on these broad principles. If you look at the NIST AI risk management framework or the EU AI Act or some of the state laws, these ideas of be fair, accurate, reliable, transparent, explainable, great. But what does that mean? And what does that mean in my specific industry? You know, fair or accurate means something very different in a credit underwriting algorithm where you're measuring risk to what I'm going to charge somebody, um to compare that to a social media post. Like what is fair or accurate or not harmful? What does that mean? And how, more importantly, do I mathematically measure that? There's not agreement in the world about that next layer down. And so one of the things that we are heavily involved with is research around how are different ways to design AI-based systems and measure and monitor their behavior, what's most effective, and what is it I'm measuring, and how can I do it mathematically, right? Because you can't have a person looking at everything that we create. And so that's going to be the next real breakthrough in AI how do we monitor and test its behavior so that we know that it is acting precisely how we want it to.

Chris Hutchins 30:51

This is such an important topic. And I I don't know that uh the generally speaking, people are thinking this far ahead. And I I I love to I love what you're doing, how you're connecting it to, you know, you're talking about the stakes of you know, for human beings. This is this goes way beyond you know just a technology conversation. And it's really we're not trying to put solutions in there to take care of one's part of the world. I mean, this is this is a all of humanity kind of thing that we we need to kind of coordinate. And I'd I'd love to hear just your thoughts around how you see things evolving from the from a global perspective, since you you really have a lot of uh lot of background to in and you I'm sure have studied this extensively. But where do you where do you see things evolving and and how do we start to see the kind of cohesive or collaborative approach to some of the some of these important issues? Because I I I think about the ethical pieces of it, people define ethical differently, unfortunately.

SPEAKER_02 31:53

Yes, and it's been one of the the real blessings of my life to be able to have worked in all sorts of situations in industry and federal government. Um in the US with people in the EU over the EU AI Act in many different countries, AI ministries that are developing. And the nice thing is that everybody is focused on this problem, right? Everybody is is true recognizes the power of this technology for both good and ill. And just like every piece of technology, you're going to have people who are fighting to use it for good, and you're gonna have people fighting to use it for evil. And like always, that that that tension is going to be there. But fortunately, I'm a huge AI optimist. Like I really do believe that the benefits to health education, um how we educate people, all of these things are going to be more efficient, more accessible, more available. Um, it's a wonderful democratizing um technology. And so I do see some of the greatest minds and hearts on the planet working to try to define and try to understand how to control the behavior of this technology. It will be interesting. There were all there will always be this battle between good and evil. There always has been, right? And so watching how countries, nation states, non-nation-state actors are using this technology and defending against the evil uses of this technology. But I think it will make government more efficient, all of government services more efficient, it will make um, like we said, education, healthcare, financial services, employment. There's just so much potential here. I do see it unfolding that there are going to be leaders, business, industry, NGO leaders who are leading from a moral perspective, that we can then quantify and translate down to the actual place where the AI touches the world. You know, one of the things that I hear all the time from business leaders is, oh, this AI thing is just a big black box. I don't understand. If people who make it don't understand how it works. Well, yeah, but I don't know how my washing machine works, but I, you know, I know how to make my clothes clean. But fair point. You can always, always measure where AI touches the world. And so if you think and that's really where it matters. That's where the the goodness happens, or where you get in trouble. And so, like uh pick an example of like HR Tech. So a lot of companies are using HR platforms that sort through resumes and decide these people should be interviewed and these people shouldn't. Well, I can take that pile of yeses and pile of no's, and I can compare them and look for discriminatory risk, or am I missing something? And you know, are there unicorns out there that don't fit a pattern? Um, and so no matter what your application is, you can always measure where AI touches the world. And there's usually laws or regulations or guidance about that. And that's where you focus your governance layer. It's where you focus your monitoring and testing, and that's how you kind of stay out of trouble. And so that's how I see this kind of unfolding in the world is that the responsible use of AI, especially in government and in business and human rights and civil rights, is is that is where you're going to see how this developed.

Chris Hutchins 35:55

We can kind of take another swing at you know what what we should be uh expecting in the future. First, what are some of the things that that you see that um organizations might be missing right now that they need to be thinking differently about? Because that I think we get we see a headline, we respond to that. Uh sometimes it's it's the right thing to focus on, sometimes it's just not. Uh what are some of the things that you're seeing that that people need to be aware of and start to focus on as they're leading their organizations?

SPEAKER_02 36:27

I see a lot of organizations kind of stuck at the enterprise level tools, right? Like how do we use open AI, how do we use NropEg? How do we use Copilot? That's great. You do need acceptable use policies, do this, don't do that. Yep. Those are but those are just table stakes questions, right? It's thinking one layer down. What is it that truly distinguishes you as a company? What is it that sets you apart from your peers? Where are the artifacts of that that I can access? Because they reside in two places. They reside in your data. So you know, everything from email and documents and PowerPoints or research or you know, wherever they're at, but they're also in the heads of your people. Now, the really cool thing is that AI, especially agentic AI, can capture both of those things. So what we tell companies is that the best use cases for AI are where a task is data-centric, so it's it's based on data information, it is highly repetitive, you do it all the time, and you're applying the same kind of judgments to it over and over and over again. Well, you can capture that data, and you can capture the judgments, whether in written policies or programs, or whether they're in the heads of people. So there's this fantastic concept that Anthropic developed called constitutional AI, right? And it basically the idea is that you take an AI system that is this glob of all of human knowledge, and you contain it within a constitution. And so you give it a an identity. I I talk to uh technologists all the time. I'm like, you are not programming these systems, you're not coding these systems, you're forming the psychology of these systems. And so giving it a purpose. You are a chatbot that is going to interact with customers about this product, right? Give it it it hones its focus into on this is what I am, right? It's prime directive. And then you use very simple code, these are all just plain language prompts. This is how you are to act, right? These are the constitutional laws by which you will behave. Um, things like accuracy, compassion, empathy, whatever they are. And sometimes those laws are driven, those the laws you give it are driven by actual laws. Like especially healthcare is one of the most um regulated industries, right? And so if you're using an agent to do a thing and there's rules around that thing, well, you program those rules in. But then there's this um that's where you can also put the judgment, the acumen into these bots, right? And so, and then you know, the quality control is you put scoring mechanisms around each of those. And those scoring mechanisms are related to the metrics we talked about that you need to prove that it's defensibility. Um, so what I see happening in the future is us finally getting beyond the question of should we use AI or should we shouldn't? Of course you should. 99% of the times AI augmenting humans is the way this is going to unfold in in the future. Yeah. Learning how to contain that acumen, take advantage of that acumen, and then using it to augment and extend it, that is what good governance is, but it's good operationalization of that governance to watch your company put on a big old supersuit and be um a next generation company.

Chris Hutchins 40:43

I I love the the Iron Man analogy that you use. Are you gonna like build a platform that will, you know, just create a suit like that that I can put on? Because I like to be the first.

SPEAKER_02 40:55

We do it all the time. Um, we just basically take you and expand what you're able to do. It's really actually one of the most satisfying things that we do as a company is to watch the vision of a company play out. What the one of the most important things we say to companies all the time is that when you think about AI, it's only 20% of that problem is a technological problem, right? Like where you're gonna apply it, are you gonna technology control it? 80% of it is a psychological or sociological problem. You've got people in your company who feel all kinds of things about AI. A lot of it, some of them are you know at the right side of the bell curve, they're just like rearing a go, this is fantastic. You've got the people on the other side who are scared to death that this is just replacing me, right? But the truth of it is um everybody wants to have a supersuit, right? And so the the communication and the the training and the change management that you do is the really critical piece to making the technology work in combination with the humanity.

Chris Hutchins 42:10

Yeah, that's a that's a great a great point. And I I been talking to people that do coaching and other things a lot in the last few months, and there is a common theme um that that I'm hearing. And you know, I just I think it's just an important fact for for leaders just to really take a look at how you're how you're doing with the communication end of it and how but how well are you equipping that first layer of supervision uh out on the front lines of the organization is oftentimes uh it's these frontline people who are not feeling it. They may have unspoken fears or whatever. The conversations need to be happening, and I think we've we've uh done a great job across every industry in terms of creating the right career ladders that allow someone to be an individual contributor and a rock star at several levels in their career. Um, but oftentimes the only way to retain and promote somebody is to put them in an oversight role. And yeah, not every organization does a great job, you know, training those people and teaching them about creating psychological safety, for example. Yeah. Yeah.

SPEAKER_02 43:21

So you know what's so interesting, Chris, is that so when we start to work with a company, we do what we call this discovery phase, where we go in and we talk to the leaders, the leaders of functional areas, but then talk to the middle management and the frontline people. Because what we're trying to find are where are your pain points, where are your choke points, right? But then we always ask this magic wand question, right? If you could poof into existence tomorrow some new capability or some new insight, what would that be? Because it's it's the middle and front line people who deal with this. The process, whatever the process is, every day, they know where the problems are, and they almost always know what the solutions could be, right? And so finding those pain points or clog points and using AI to solve the pain or to unclog the decision-making process or process workflow, whatever it is, right? That's where you just start to see this machine that is a company. A company is just a it's a bunch of cogs working together. And you see the friction start to go away, you see the interaction start to increase, you learn across departments where you've got great understanding in one place and not another, where you can spread that around. Um, and all of a sudden, as a management layer, you're seeing much more deeply into the company and much more information and insight where you can guide strategically where you want your company to go.

Chris Hutchins 45:02

No, I I love that. It is such a it's a key point that we we just need to make sure people are are hearing it uh enough. So I mean, I really appreciate how you address that. As we kind of wrap, um listeners could re will remember only one thing from our conversation today. What what do you hope it is?

SPEAKER_02 45:23

It is absolutely think Iron Man, not the Terminator, right? So, how is it do we take humans, the acumen, the compassion, the moral compass, the creative spark of humans, and and that's what we want to magnify, right? So many people are worried about the replacement of people with AI. Right. Yeah, there's thank goodness the drudgery tasks are gonna go away. They just simply are. Um, AI is better at it. Um, but that's great. It it frees up essentially human characteristics of what we bring to things. I work with a lot of creatives, like even in the fine arts, right? And they were some of the earliest concerns, like, well, if somebody can put in a prompt and get a picture, they don't need a graphic designer. I'm like, but they do because what they're hiring is not what they're hiring is your creativity. You were born blessed with this creative talent, and it's not just can you be replaced? Right, but there are some amazing artists out there who have taken AI as an additional tool in their toolbox, like a new brush or a new um musical uh capability and creating entirely new kinds of art, entirely new kinds of creativity. And there are always we're already seeing this kind of pendulum swing back to I want a human-made thing, right? I I appreciate even more that which is truly human. And so what I want to leave is this message of great hope. There are things that we can do today that we could have never ever done before, and tomorrow is going to be even more new things, and so it is this combination of things that are beautifully, uniquely human that we can make flourish by this utterly magical transformative technology.

Chris Hutchins 47:43

Beautifully said, I'll speak for myself. Um, I I'm inspired. Um, it's really tremendous work that you're doing. And uh, you know, clearly there's a a passion for it and a passion for people. Uh I really love that. It occurs to me, having heard our conversation today, that there's probably a fair number of people out there that would love to know how to how to get a hold of you. Um about your work and connect with you. Uh tell tell us how they can find you and what modes do you like to work with?

SPEAKER_02 48:16

So um our website, clarionai.com, um, you can reach out to Drew's there, find me on LinkedIn, Bennett Borden. My email address is Bennett.borden at clarionai law.com. I love having conversations with people. I love just answering questions. It, you know, if it leads to a project together, God bless. Um, but if but really what I want more than anything else is to help people understand. And and whatever they cherish and believe in that leads to human flourishing, that's what we want to forward.

Chris Hutchins 48:54

That's fantastic. Well, for listeners, uh you definitely look in the in the show notes. We'll make sure that you have uh all the information that you need to be able to reach out to to Bennett and his team. Uh Bennett, this has been fantastic. I really have appreciated your your perspectives and um I learned a lot. So thank you so much for for sharing your your your stories with with me today and for joining me on the signal room.

SPEAKER_02 49:18

Thank you, Chris. It's been a true delight to be here.

Chris Hutchins 49:21

That's it for this episode of the Signal Room. If today's conversation sparks something in you, an idea, a challenge, or perspective worth amplifying, I'd love to hear from you. Message me on LinkedIn or visit Signal Room Podcast.com to explore being a guest on an upcoming episode.