Overcoming the AI Governance Bottleneck: The Future of Autonomous Workflows

Hey, everyone. Welcome back here to Techstrong TV. Hey, my next guest for you is Darren Kimura.

Darren is the CEO and president of a company called Ai Squared. Let's welc- welcome him to Techstrong. Hey, Darren.

How are you, man? I'm doing well, Alan. Great to be here.

Thank you, and it's great to have you on. So Darren, we're gonna talk all about Ai Squared, we're gonna talk about AI and jobs, but before we talk about all that, let's talk a little about you. Our audience is gonna be listening to you.

What, what should they know about you? Well, I'm- I'm originally from Hawaii, so I was born and raised in the Hawaiian Islands. Um, I grew up initially on a plantation, so I started my career, you know, cutting sugar cane.

And then- Really? in the very earliest days, when I was about 12 years old, I was an electrician. My dad decided to start a company, and it was a contracting company, so I was a little guy, and because, you know, I was 12, I was the guy climbing through the roofs- ...

and doing all the wiring- Mm-hmm ... at the edges of the dusty- You fit in good into the tight spaces. Re- I could fit into the tight spaces, exactly.

And, you know, it was a great experience. I mean, I had to wash the vans on the weekend and clean the floor of the warehouse. I mean, all of those kinds of things, all of those things that small business owners do every single day, and, and it was a great start in my career.

Uh, from there, I went into electrical engineering and began designing power systems, and one of the things that I started to do very early on, this was about 19 years old, while I was a sophomore in college, was, doing energy efficiency designs in buildings. So that means I would go into a building and I would figure out ways to have it use less energy, better air conditioning, better lighting, and the idea was a perfect idea at that time because we were facing energy crisis, energy shocks, high energy costs. So the business just grew and grew and grew, and g- out of Hawaii, we grew, and we went to the Pacific Northwest, and ultimately, 14 offices around the United States.

" I never thought I could build a company to sell it. I always thought, you know, you build a company to run it. Right.

Yeah. So that kind of became my journey. Uh, sold the business and said, "Let's do it again," I was in my early 20s, and started a solar company at the right time, and, you know, this was before th- the global solar markets came online, and invented a solar panel.

Um, got 14 patents, built factories, and f- installed, systems in 70 countries around the world. It was- Wow ... quite an incredible journey.

Took it public, got rolled up into one of the largest, publicly traded, solar companies at that time called SunEdison, which was really cool. Mm-hmm. And, and then, you know, en- enjoyed or experienced life as a publicly traded company.

Um, from there, I became a venture capitalist. I wanted to invest in the companies. I had always previously run the companies, so I decided to get into, like, investing in companies and really see it from the other side, and invested in about 20 companies.

Um, experienced three IPOs over that journey and, and sold five companies. That's great. It was- 20 would- Very different ...

20 to 3 is a great ratio, man. It w- it was, it was, very different, but I realized, though, it wasn't for me. Like, I'm not a great investor.

I'm, I'm a much more, enthusiastic operator, so I really, I love going into the businesses and creating value and figuring out product market fit, and how do you grow quick, and all those things. That's really where I, I really get, you know, kinda tuned in. So I, ended up joining one of our portfolio companies, and this was in network management, so we sat, we had servers that would basically sit across a network and watch network traffic.

Sure. And, that company went incredibly fast. We went from zero to 50 in about four years, and got bought again by a private equity firm.

Uh, and then I got into cybersecurity, started a cyber company, and, and today, here I am at Ai Squared. Very cool. What a great, what a great- ...

journey this has been, man. Good for you. What, what island are you from in Hawaii?

You know, I'm from the Big Island, so most people- Really? know us for the active volcano. Love the Big Island.

Yeah. Yeah. Yeah, Hilo.

I'm from the Hilo side of the Big Island. Oh, sure, on the, on the rainy side. The rainy, exactly.

Yeah, that's the other thing we're known for is the rain. Yes. Yeah, Hilo's, uh...

So I, I love Hawaii. I f- so I s- first went to Hawaii actually February, it was 36 years ago, on our honeymoon, and I went back five, six, seven years in a row after that. Then had kids and we, you know, it's hard going from the East Coast to Hawaii with little kids.

Yeah. Um, but once they got old enough, maybe 16 or so, we started doing our annual trek again, and, you know, I've been to Hawaii now, I don't even, I couldn't tell you how many times, but I don't know, at least 25, something like that. Wow.

Um- More times than me probably. Well- Since I've been back. Yeah ...

we, you know, I mean, Maui is our favorite, though. Yeah. We've been to all the islands, and we used to stay in West Maui, and since the fire, we, we kinda bummed out, you know, that- Right ...

it was a terrible, terrible disaster there. But anyway, w- it's a great story and w- so much success, man. Congratulations.

That, that's a, that's an amazing journey. Let's talk a little bit about Ai Squared, right? Obviously, it has the word AI in it.

We're gonna assume AI has something to do with it, but give it, give us the Ai Squared kinda rundown. Yeah. Yeah.

You know, so I joined the company as the, company was looking to grow, and the, the investors were looking for someone who had scaled a business before, and obviously, we all know what's going on with AI. It's going so fast right now. Companies are getting bigger faster than ever in history, and this was obviously about two years ago, so this was before, you know, we have the benefit of hindsight now, but they knew what was coming.

So I was brought in, and one of the things, or a few of the things that really caught my attention was the background of the company. And what, the team did was they were all part of the NSA and they were all responsible to build secure data pipelines to run, machine learning models and create top-secret reports. So it's a very secure platform.

Um, what it does is we go in and we connect to all the different locations that a company might store your data. So for example, if you're using Google Drive or OneDrive, some organizations that are larger could be using things like Snowflake or Databricks or even AWS. We basically just s**k all that information into the platform, and then once it's in the platform, we can go in and activate a bunch of different ML models or AI models to make the data more usable.

We, we apply intelligence to it. It's done in a secure way. And then we add things like, trust and governance.

We make sure that the results are compliant with whatever your industry requirements might be. For example, in the healthcare industry, they have HIPAA. So you wanna make sure that the results coming out of a, the, the large language model are gonna be compliant with HIPAA.

So we can do that. These are all components of our system, and then we take all of the insights, and we put them where you work. So, for example, it might be in Teams or Zoom.

com or even in a website. It doesn't matter to us. We take all of the different...

We call them applications, right? Data applications, and we put them where you want, and we do it in a very secure way. So it was such an interesting proposition because the company had this rich history in dealing with the most sensitive data in the world, building these really incredibly, robust and secure data pipelines and then working with enterprises, which is my favorite thing to do, to make their end users, make their salespeople, make their finance people, a lot more efficient, work faster, reduce errors.

Um, just the perfect situation for me and, and as you know from my past, I love to go into companies at that inflection point when the industry is taking off, when the business is taking off. So I just saw all of those things here in Ai Squared, and I was just thrilled to join the company. Very cool, very cool.

Um, how, how... So you joined, you said about two years ago. Is that-- How old is the company, though?

About five years. So they, they... So this predates the whole ChatGPT kind of explosion.

Yeah. So we call that the, the transformer model component. And this company was really, um...

It, it's still within the AI era, but at the earlier days, to your point, there were still large language models, just not as smart, but also a lot of data science models back then. Sure. You know, things that can do predictions for example.

A lot of ML, ML kind of stuff. A lot of ML. Yeah.

Yes. Pattern matching. So we can do everything from your traditional ML and, you know, bring it all into large language models as you know them today.

The platform is designed to be very robust. So as models change, we can actually swap models in and out. We actually have some technology that can figure out the right model for you, depending on whatever your question is or the, who the user might be.

Um, so it's a very flexible model designed for, for growth, designed for the future. You know, so Darren, I, I've, I've founded a bunch of companies myself, venture backed and so forth. O-one of the things I learned about working when you come out of the, the DOD, the whole...

You know, I mean, the federal marketplace, but DOD and agency specifically, and I, I learned this lesson the hard way, unfortunately, is that we wound up making a product that was so specific to the DOD and agency mission, and it, that just didn't lend itself as well to like... You know, we, we went the same way, enterprise DOD. Even the enterprises didn't want, you know, milit- And it was a cyber company.

They didn't want that military-grade cyber as much as we thought- Yeah ... they would, but we... But, you know, in the meantime, we had a lot of contracts and, you know, business within that DOD world.

Um, but it just didn't translate to, to the civilian market. I wonder... Now, now you guys clearly come from the DOD side of it.

Have you been able to transition to a, a civilian, if you will, a, you know, use case that- Yeah ... well, you got market fit and everything else? No, you're so right.

Because when you build for, you know, mil-spec or mil grade, it's completely different, right? And, and when you build for that, you have components in there that you may not necessarily need in the commercial world. That would be overkill, yeah.

It would be overkill, so you're completely right. Guilty. One, one of the things that I've really done is spend a lot of time, my time in, in, in coming up through enterprise.

So I spend a lot of time with the customers, focusing on user experience, and I think that's maybe where the big shift has been for Ai Squared, where we still can build for mil-spec, and we do that of course for the military work that we do, the DOD work that we do. But a lot of what we do is also focus on the user experience, the UI, making sure that the workflows completely connect with each other and they're logical, familiar look and feel, drag and drop, right? No-code.

Um, it's really important I think these days to go to no-code. And some of that, by the way, is also usable back in, in the, in the DOD, DOW world. Mm-hmm.

One of the things that we started with our mission as being was to help the war fighter be able to deploy AI models, from where they, where they are. So if you think about a lot of times, what happens is war fighters might be in the back of a Humvee working on a laptop, and it's really tough to write code when you're bouncing around in the back of a Humvee. So we had to make the code i-in such a way where it was basically no code, right?

So that we have lines and connect lines to g- you know, images or cards as we call them, and that allows them to be able to create the workflows. That's also very usable in the enterprise too. Now what we're doing one step beyond that in the enterprise is we're actually bringing vibe coding.

Everyone's talking about vibe coding and how you can create an application in hours these days. Well, AI Squared can create the, the AI pipelines, the secure data pipelines by vibe coding as well. So you can go into a prompt and say, "Connect my databricks to this model and put that into Salesforce," enter, and we're able to create the entire code without having to code Python.

I love it. I love it. You know, I-- it's funny, I actually had just finished an article before I came into the studio here to do this with you on, you know, what's the first AI war?

Is it Russia, Ukraine? Is it this current Israel-US with Iran? Is it Iran-Israel from June?

Wh-whichever one you want to pick, we're, we're really starting to see the influence of AI, right, w- in our war fighting. But I, I wanna get away from war fighting. I-- enough wars.

Um- Agreed ... let, let's talk... Well, before we do, just for people who want to get more information about AI Squared, actually, I don't think we mentioned the website.

Sure. What is it, Darren? ai.

We have a lot of AI. AI Squared. Lot of AI Squared.

ai. I mean- Yeah ... that's, it's a good pun.

Um, and that's great stuff. ai. Darren, I want to kind of pivot now to the bigger AI question, right?

And, and it's one that we get. It's one that I write a lot about, I speak a lot about on these videos. Jobs, AI and jobs, right?

If anyone- Yeah ... out there is saying, "I don't worry about how AI's going to affect my job, or whether AI is gonna replace me," they really are not-- they're either not telling you the truth, or they just haven't been paying attention, right? I mean, clearly AI is having a, an impact.

I'm not saying it's gonna take your job. " But, um- That's really good. Yeah.

Yeah. But from where you sit, what do you see? Job killer, job creator, a little of both?

You know, I, I generally think we're on a journey, and, what we saw and probably still see, particularly from corporates, is, you know, a, a convenient excuse to be able to reduce costs, which in many cases w- comes through headcount reduction and, and today that is AI. So you're definitely seeing across corporate America, you know, we're sh- you know, cutting forty percent of our workforce because of AI. But, but if you look really closely and you look in-- you listen to some of these, earnings calls, what you're also finding is that they're realizing AI may not be there yet.

It may not be as accurate, and as a result of that, there's not as much trust. So what they're having to do is backfill that now with humans. Salesforce, for example, about six months ago, nine months ago, had talked about h- head force reductions.

Recently, they talked about rehiring three thousand employees because the AI, agents that they had deployed were not accurate enough, and they now have to go back and verify them. Now, I, I do think that's also a temporary thing. I think over the course of time, these, systems will get better, get smarter.

Eventually, they'll regain, you know, our trust. But it's a little bit of both, right? So as, as you framed it, it's gonna be some companies are early adopters and are really aggressive, other companies are maybe not as much, and, and everything in between.

Yeah. I mean, I see it too. You know, we, we-- as we were talking off camera, I told you we, we really have made a big push with agentic AI over the last two, three weeks.

New. I mean, this is still new. And what I'm already starting to see is, you know, I'll, I'll have this newsletter template perfect.

Mm-hmm. Better than we've ever had. All right.

Let's just run it every day now at six AM, right? Y- you know, all the integrations are done. And every day there's a little drift.

You know what I mean? Every day- Yes ... " And it's like, oh, a- and you know how AI is.

It's very polite. " You know? "Uh, let me correct that.

" Boom, boom, boom. You know? And, so we-- it's...

Is it perfect? No. It, it's far from perfect.

Is it getting better every day and week? Absolutely. I, I think the bigger issue, and this is what I tell my people here at Techstrong, you've got to figure out what this can help you in your everyday task.

What can it take over from you in your everyday task that allows you to do something even more that you don't get the time to really do but could be really important? And the other thing is it's good at doing that task, but it doesn't necessarily think that task up. And- Yeah ...

you as a human have to think. That, that spark of creation, if you will, and then tell it to go do it. That's what's gonna be the difference between human and machine in, in terms of doing this, these things.

And the idea of creating, of building, of coming up with that idea, with that spark, is I don't think AI is gonna do that anytime soon. Um, and, and, you know, that's what I'm encouraging my people, where they need to develop their thinking skills, their-You know, just their curiosity to go do things and explore things and push these boundaries. And, and I see it every single day here, Darren.

I'm, I'm seeing every, you know, every day, 'cause everyone reports to me, "Hey, this is what I've been doing today with it. This is the new thing I taught it. I, I have it doing this, so I'm doing that," right?

And- Yeah ... I looked into the new way of doing it. So it, it is very exciting.

But, you know, the other thing I... Here's another thing I've seen. I don't know if you've seen this.

People who get it, who, who buy into it, they're almost manic about it, right? Like, they don't wanna go home. They come in on weekends, they work from it at home.

They wanna use it and keep pushing and exploring and trying and experimenting. That scares me a little too, though, 'cause that leads to burnout eventually. You can't do 70- Yeah ...

hour work weeks forever and, you know, that- that's a problem. Um, I'm wondering what you... I mean, you're probably, you know, you're seeing it across multiple companies, your experience.

What do you, what do you think about it? Yeah. I think AI is leverage.

I mean, I think if used properly- It's a great way ... it can help you. Exactly to what you were saying earlier, right?

And I think I, I love the way you have this feedback loop within your organization where you encourage the use of it, but then tell me how you're using it, right? And, and iterate from that. Um, it, it allows you to be much more efficient, in particular with some of the more routine projects or work that you would do in your every day, which I think is so powerful, because then it allows the human mind to do what it does best, right?

Be creative, solve difficult problems, think about big picture strategies, and, and that is uniquely human. Maybe one day in, in the future, who knows what's gonna happen, but today, this is where I think we set ourselves apart from AI, which is more predictive, and humans, which are a lot more creative. So, you know, if you think about some of the best things in life, right, these are plays and songs and, you know, just things we enjoy, m- you know, movies.

It's the impromptu moments, right? It's the imperfection in the drawing that makes it so perfect, right? It's, it's that movement that was unscripted or, or that line that was, you know, again, unscripted, but that's what makes it h- human, and that's what draws people in.

And I think, again, that is really uniquely human. And, and that's also what I think engages us, in, in, in that work. So if it's, like, a routine thing, routine codes, to your point, newsletters and whatnot, I think that's fine.

Let's leave that to agents to automatically work on that, work with, with the agents to come up with, you know, the right guide rails for that, and continue to push that forward so you can produce more of it. But then on the other side, to your point, here I live in Silicon Valley, and we do have companies that are looking at, you know, nine to nine workdays, six days a week, you know, 70-hour work, work weeks. Really difficult, but they see it as a, as a very long sprint.

You know, it's, it's a race- Yeah ... to, you know, whatever the next checkpoint is, whether it's the finish line or, or maybe, you know, an incremental finish line along the, the course of this race. So you definitely do have that.

You, you have other countries working in that way as well. China is an example of where that- Sure ... that, that concept came from.

Um, so they are trying to, you know, work twenty-four/seven and, and get ahead of it and build code and deploy apps and, you know, make robots smarter than ever before. So it's, it's a mix. I mean, it, it comes down to, I think philosophy a little bit.

Like, our philosophy at AI Squared is you need to have work/life balance. I think creating, reliable technologies that are, you know, again, balance. Balance and humanity, but also taking advantage of the advantages o- of AI.

I think that's- Of what it offers. Yeah ... of what it could be.

But, but do it in a safe way. One of the things that, that really encourages me about what we're doing, is that we bring in the safety layer. We bring in the, the guide rails.

We bring in the governance to ensure that we're not just having AI running off the rails, which in some cases is what you're seeing. I mean, we've heard, for example, of Grok being, you know, drifted in a way where, you know, it, it could be spewing out potentially hateful things and, you know, companies like AI Squared, our technology are specifically designed to eliminate that from happening. That will never happen through AI Squared.

I get it. I get it. You know, the, the other thing, like you mentioned this, a- and this is again, I have a good vantage point where I sit, 'cause I get to talk to people like you, I get to talk to a lot of different users, producers, et cetera.

What I think AI has created is our attention, emphasis, and money used to go to creating code. Creating code is no longer the bottleneck, right? Have you ever read the book The Goal by, Goldratt?

It, it's an old book from master's MBA programs in the '80s. But, actually, my friend Gene Kim's book, The Phoenix Project, is based on The Goal, and it's about the theory of constraints and bottlenecks i- in the manufa- this was more in the manufacturing, not software. But what we've done is we, we've, we've broken the dam on code creation, right?

Mm-hmm. We- we're creating more code than we ever have created. But, you know, that's the theory of constraints.

As soon as you break that bottleneck, you run into another bottleneck. Yes. And that bottleneck where we are now is, all right, we got all this code.

Governance, security, compliance. How do we test it all? How do...

And at each, I think at each bottleneck, AI's gonna play a place, 'cause you, you kinda- Yeah ... need AI to, to use AI. But we're learning this, right?

One, one, one bottleneck at a time, right? That's, that's what we're going through right now. And but, you know, this is the process.

This is, this is evolution, right? Happening in real time right before our eyes, so. Interesting.

You know, I call that the ironies of AI. Mm-hmm. I- it's exactly what you're describing, where we create AI to do all this work, but it gets to the point where we need AI to govern the work, and then we need AI, so it just kinda continues to build upon itself.

I- I, yeah. It becomes like a self-fulfilling kind of prophecy thing, but- Exactly. Yeah ...

and, and we're seeing it in cyber too, right? I have 25 years in, in the info sec cyber space. You need AI to secure AI.

You need AI- Exactly ... to fight AI, 'cause the bad guys have AI. That's right.

So this goes, you know, it's a bit of a circular kinda thing, but I don't... You know, that's the, this is the world we're in right now. Anyway, Darren, I wanna thank you for coming up here on Techstrong TV.

We probably did more than our 15 minutes. I apologize. It was so much fun.

But, yeah, it was a great discussion. Keep us posted on what AI Squared is doing. Come back anytime.

Let's chat. It would be my pleasure. Thanks for having me, Alan.

Appreciate it. All righty. ai.

AI, you got the squared, like the number two squared- You got it ... dot AI. All right.

Darren Kimura, CEO and, president of AI Squared here on Techstrong TV. We're gonna take a break. We'll be back in a minute.