AI at Work: How iClerk Is Redefining Enterprise Efficiency

In our latest episode of the Disciplined Troublemakers podcast, I sat down with Tom Blair, CEO and co-founder of iClerk.ai, to explore how AI agents are transforming enterprise workflows. We started with the least glamorous parts of business: the critical but repetitive tasks that frustrate teams and drain resources.

Founded in 2019, iClerk builds autonomous, role-based AI agents designed to streamline complex financial processes. These aren’t chatbots. They’re cognitive assistants designed to handle LP reporting, NAV calculations, fund admin workflows, and more. And unlike many AI implementations, iClerk’s agents are traceable, auditable, and secure as these are core requirements for highly regulated environments like finance, insurance, and nonprofits.

Tom’s thesis is simple: most companies don’t need a moonshot, they need relief. His first customers weren’t seeking AI innovation. They were trying to stop analysts from quitting over spreadsheet drudgery. “It wasn’t the brilliance of AI that won them over,” Tom told me. “It was retention.”

This is where iClerk’s approach shines. Instead of pushing a top-down AI transformation, they focus on atomic units of work which are small, well-defined tasks. Each agent handles one job, like compiling a Schedule of Investments or automating K-1 generation. The implementation is fast (often under 100 man-hours), non-invasive (customer data stays put), and low-risk (iClerk doesn’t charge until it works).

This tactical, workflow-first mindset avoids the “religious fervor” often associated with AI adoption. It also enables exponential operational leverage. As Tom shared, “Once a company sees one agent deliver value, they quickly want three more.”

From a scaling perspective, iClerk is expanding along two key vectors:

  1. Market Reach: Starting with VCs, they’ve expanded into private equity, family offices, 501(c)s, and public SMBs as each face similar statutory and reporting burdens
  2. Agent Proliferation: Once embedded, customers find additional use cases. The economics are compelling, such as ten agents for the cost of one professional hire and the ROI is rapid

Underpinning all this is decades of experience. Tom and his co-founders aren’t twenty-somethings chasing trends. They’re veterans with institutional knowledge, deep networks, and a long view understanding of enterprise pain. That maturity has led to a distributed model built on trust, connection, and asynchronous collaboration, run entirely without a traditional office.

As we wrapped our conversation, Tom shared a powerful sentiment: 

“Humans are bad at exponentials. But AI is moving at 4–40x improvements per year. You can’t predict the future. You can only take the next step.” 

That grounded perspective is exactly what makes iClerk worth watching. They’re not chasing AI hype. They’re solving real problems and quietly transforming how work gets done.

And they’re doing it one step and one agent at a time.

Transcript

Scott Shagory:
Welcome back to the Disciplined Trouble Makers podcast. I’m your host, Scott Shagory. Here on DT, we talk about scaling great tech companies and helping CEOs do this with greater ease and flow. Today, we’re gonna be talking with Tom Blair. Tom is the co-founder and CEO of iCLERC dot AI. Founded in 2019, iCLERC provides secure and customized AI solutions for enterprises. Previously, Tom cofounded Performance Livestock Analytics, which was acquired by Zoetis, and ClearAg, an AI weather analytics platform acquired by DTN. It also held senior roles at Trimble Navigation, Sybase before it was SAP, Autonomy Systems before it was HP, and Oracle contributing significantly to mobile, telematics, AI, and database technologies.

Scott Shagory:
Today, we’re gonna be talking with Tom about the challenges and opportunities around scaling AI companies. Tom, thank you for carving out time to chat with us today, and welcome to the Disciplined Troublemakers podcast.

Tom Blair:
Thank you for having me, Scott. It’s a pleasure to be here.

Scott Shagory:
Yeah. Thank you. So, obviously, I’ve provided a little bit of info in the intro there. But to, I guess, maybe flesh out what you guys are working on and and with what you’re comfortable sharing, how you’re doing that. What is, you know, what’s your ideal audience, the key challenge, and and a bit about your secret sauce for Alright.

Tom Blair:
So iClerk, AI is an applied AI company or a vertical AI company Okay. That captures cognitive enterprise knowledge and builds these autonomous role aware agents Mhmm. For, our primary markets have been the financial sector. So both private and public companies, five zero one c’s as well. Anyone that has frequent statutory reporting with critical business processes, often you have, these firms outsource or offshore that type of work, but it’s always professional work that needs to be done by a venture analyst or a CPA. So these are tips we automate accounting processes, critical financial processes, things that need to be, accurate and repeatable and provable. Right? And so things like LP reporting or net asset value reporting or, automating the creation of a schedule of investments from portfolio company inputs that come across your email. Yeah.

Tom Blair:
So you can think of it as truly cognitive processes that occur that comprise your business. These agents, like I said, they’re repeatable and they’re traceable. They use multiple types of models. You can think of an agent as a set of actions, and each action could be, apply a different model. So you could have a deterministic model if you’re doing, you know, numerical processing, for instance.

Scott Shagory:
Okay.

Tom Blair:
Or you could use different language models or other models based on cost or performance or even jurisdiction where you are in the world. Right? Mhmm. The beauty of these is that I always refer to it as a change in epochs. I think I’ve been through a couple in my long career. Mhmm. I was early in the database industry, which affected every application in every sector in every country in the world. And I traveled a lot for that through the eighties and nineties. We went through the cloud and mobile epoch, which, again, every system of record changed with a new architecture and new tools available.

Tom Blair:
And you’re seeing this again with these AI tools. Think about agents as tools. It is the use of these tools and the ability to comply and to use different models to adapt to custom workflows. So it’s really a change. You’re able to implement these very quickly. They leave the customer data in place, for instance. The integration is that these AI agents can act as an index so you can point at all the data. And so the implementation is really quick, hundred man hours often.

Tom Blair:
And so, probably the most profound change in the industry they’ve seen over the last in 2025, just over the last couple months, has been the willingness to engage. People have gotten over the fear hump and are starting to look at this mainly because of the economics of them. You know? They’re typically, you can our pitch line if we replace 90% of the work for 10¢ on the dollar. It’s basically, the users set the price, whatever you know, they’re executing this work anyways. They’re using CPAs at $400 an hour or they’re using IT staff to create reports or venture analysts to collect portfolio data to put together an LP. You know, the list goes on, and that’s really mundane, repetitive work, but it’s very critical. So these agents, you can effectively have 10 agents for the cost of one human. And so that lowers your bar of entry.

Tom Blair:
If you just focus on a single agent that solves a single problem, people can get their head around that. They could start to see the value proposition and the time saved and then start to see the other value that AI gives you is the ability to generate new things and to do discovery and one off queries and such. So it’s really an age agent, operating environment. It allows you to rapidly well, that’s what we do is we don’t even charge for integration. We go in within a one quarter period. You can automate a workflow, and pay for it at the end. Right? It really changes how professional services are delivered. It’s having a profound effect on that industry.

Scott Shagory:
Yeah. No. I can certainly we’ve all all of us who have you know, even on a very simple or superficial level can see what those those changes are as everyone in the in the, in the industry or broader IT space looks at orchestration differently, looks at data differently, looks at, you know, the API functionality differently. There’s all sorts of things. And I think what’s helpful in being productive is that you don’t have to do something super complex. You can pick a workflow or a process that is, as you said, mundane, standardized in some way, shape, or form, whatever that SOP happens to look like for company a or b. And then, you know, one step at a time, just relieve pressure from people’s, you know, backs or whatever the team looks like. Because usually a lot of that work, not only is it cumbersome, but it just takes an enormous amount of time and most teams or many of them are stretched.

Tom Blair:
So our first customer had problems with retention. They’re in Edmonton, actually. And, you know, nobody wants to get their MBA and and, you know, join a venture capital firm and then and then spend your time chasing down a port of cofounders for their balance sheet to put together as an Excel spreadsheet that generates Yeah. An LP report. You know? And and and that retention is what drove the first adopters of this, actually. It wasn’t the economics or the ease of use and brilliance of it. It was basic, you know, staffing problems.

Scott Shagory:
Yeah. Yeah. Or just the pressure. It was what I could describe as a level 10 bleeding neck problem that somebody had that was just causing them an enormous amount of discomfort and, you know, AI agents or whatever the solution happens to be can actually go ahead and solve that, solve their problem.

Tom Blair:
They’re role based. You treat them just like you would an employee. Right? And so Mhmm. They don’t have this autonomy even though all agents have this level of autonomy. Mhmm. They are role based. And, you know, if you have an agent serving up, you know, a private Wiki content for your knowledge base Yeah. They might have more latitude on how they respond.

Tom Blair:
But if it’s if you’re generating an LP report, you know, everything needs to be one quick trace of all you need to see, nobody believes in AI. They don’t believe in summaries. They don’t believe in the numbers. Right? They’re nondeterministic by nature. And so when you drop that’s basically what I Kirk is is you drop Turing type functionality that’s traceable, auditable, repeatable off one click. You can always go back to the source, onto, like, a weather machine, an ensemble set of models that process concurrently depending on the task. Right? That says that it’s the melding of those deterministic, nondeterministic functions into an agent that can choose what to do at the appropriate time based on the data. 

Scott Shagory:
Yeah.

Tom Blair:
And then insert a human for review or whatever. Always human for the Right? For the

Scott Shagory:
Even if it’s two or three times or just once or the expertise that respective, a respective person might have. Yeah. Because I think we can all appreciate the value of that for a variety of reasons. Yeah. Yeah. What does, you know, you guys have been around since 2019. Obviously, you’ve been through COVID. A lot of things have changed to say the least.

Scott Shagory:
What do you know, here in, you know, 2025, we’ll say 2026, the next eighteen months, what does scale look like to you and the team, now and going forward? Yeah.

Tom Blair:
It’s been a long road. I sold my last company in April of twenty twenty. Right? Peak COVID. Mhmm. And, you know, I got my degree in AI in 1991. I was working at Kodak in their digital imaging research lab doing semantic image recognition on building an image database, searchable semantically searchable image database. Well, you know, that’s a good idea for, you know, forty years later. You know? But, so, so we really built on ideas.

Tom Blair:
In my career, I have done a couple of things. I had worked at autonomy, I think you mentioned, in Cambridge, in The UK. Did basing inference, probability theory to do image classification and text classification. I was lucky enough to be part of the team that bought a company called Dejima, out of Japan. And, one of their customers was Stanford Research Institute, and they had a multi-agent system that would, at that time, do natural language processing. That’s what they used it for, but it’s the same type of multi agent system. The performance livestock analytics is a cow company and in the commodity space, cows are complex objects. They’re, you know, skinny cows, grain, fat cows, ethanol, and feel.

Tom Blair:
It’s the intersection of multiple commodities. And so Mhmm. We, this I click originally started as a price discovery engine for complex commodities where each agent could be an independent commodity and they would interact, to, to form other derivative complex financial, you know, objects. And so, over that thousand days that we built and spoke to, I’m an old database guy. Yeah. I’m cross sector. I spoke to hundreds and hundreds and hundreds of people across sectors all over the world, and we evolved. One one of the things that happened was the agent agentic workflows and agent nomenclature, the ability to to describe what we’re doing caught up to us.

Tom Blair:
And so, it’s been a perfect, you know, merger of that. And through that process of learning and talking and seeing the techno technological change even in the last thirty six months, we’ve through listening to customers, we’ve uncovered two vectors of growth. Right? Mhmm. And so market reach, we’ve now got sets of agents. Again, little autonomous units work might be data collection, might be a k one report generation. It could be a distributed inventory control. Any unit of work that you’re performing right now, what you find is that, you can do the data collection, the processing, and the reporting are the common denominators across that. So you could start to adapt it to different sectors.

Tom Blair:
So market reach is one of the vectors of growth. And so we started our initial customers. We’re venture capitalists, and now we have private equity firms and family offices and hedge funds and all kinds of private financial services. Five zero one c’s they have, you know, they do fundraising and they do statutory reporting on a quarterly basis. Those types of processes that we can automate. For public companies, there’s 5,000 public companies, under a billion dollars in revenue in The US. So SMB market, which is our primary market. Mhmm.

Tom Blair:
And so they spend between a hundred thousand and $300,000 a year just doing statutory reporting, 10 k’s, 10 q’s, proxy statements, you can automate all of that. Right? And so that’s what we’ve been doing is just using the tools available through Ithaca agents to attack these different financial subsectors. The other thing that we found, the other vector of growth, is the operational leverage that we found. Every customer that we’ve engaged doesn’t stop at one agent. They, you know, like I said, you can have 10 for the price of one Yeah. One professional, human. And so that’s a very compelling value proposition. Once they understand that it’s not scary and you do it quickly and it does provide value, then they find three more.

Tom Blair:
Right? And so, so it’s, these are very durable applications and durable revenue. It’s much like a database in that nobody ever just just disintegrates a database. Right? Larry Ellison is still one of the top five richest people in the world. Right? Sybase still runs 75-70% of all the financial transactions on Wall Street, and Sybase got acquired by SAP, like, twenty years ago. Right? So it’s like they never go away. And so that’s what’s compelling about this and is, you know, two vector growth, you know, a hugely growing market.

Scott Shagory:
Yeah. I just think, you know, one of the interesting things that I observe around, you know, the agent market, if I may call it that, is that you can. You can just replace what is already known or replace the most painful points of it, however that is. Once we get over the fear of it, you know, HAL 9,000 is not showing up just yet, yeah. Just to relieve that burden And then I think of it as somewhat of a paradox because we always talk about technology in certain ways. It’s just the emotional relief that somebody has, whether it is that report that they have to file or it’s any number of things that just end up being a burden. In terms of and and you’ve highlighted some of it, but I’d like to dig a little bit deeper into it if you’re comfortable, which is that, you know, you’re solving an immediate problem for a team person and whatnot. You’ve described that very, very well.

Scott Shagory:
In terms of, I guess, uncovering kind of the deep intel that gives you, you know, ICLERC, the decision superiority over competitors, you know, what does, you know, kind of what does that look like? Is it about proprietary signals within a team? Or is it about, you know, kind of painful to replicate feedback? Because there are all sorts of loops that one could engage in. But I guess for iCLERC itself and for you and your team, you know, what is that what is that look like?

Tom Blair:
One one of those the way we learn is basically, it is through operational telemetry, through really working with these customers. I’m not a subject matter expert in any of the sectors that I operate in. Right? So I’m not, I’m not a lawyer. I’m not a doctor. I’m not a venture capitalist. And so Mhmm. We rely on the partner in interaction. It truly is a discovery of what their pain is.

Tom Blair:
That’s the first question I asked them: what don’t you wanna do, and what do you spend the most hours on? And they tell us what the problems are. Are. Right? And so that’s been the most valuable one. If you pull on those threads, economically, these agents can attack these niche markets. So you look at, like, fund administration, that might be a billion dollar market, maybe a $1,500,000,000 market.

Scott Shagory:
Yeah. I

Tom Blair:
wouldn’t necessarily wanna build a company that just attacks that, you know. Auto grading for education or k one generation for hedge funds. You know, you don’t wanna build a whole company, but in a hundred hours, I can build an agent that’s repeatable and I can sell it to every partner and every LLC. Right? And so that’s how we’ve been, learning what the market wants. The other advantage we have is that, like me, all of my investors and cofounders and partners are, you know, one, people who I’ve worked with before and made money on my last companies. But Mhmm. More importantly, they’re in their sixties and really well experienced. I’ve got you know, they’re old and have four years experience and, you know, lots of contacts all over the world.

Tom Blair:
One of my cofounders spent twenty years at Goldman Sachs running their global commodities. Another one spent twenty eight years at Bank of America as a quant. Right? And so we have this deep institutional knowledge because we’re not like the, you know, 20 year old who started your company. I’m a 61 year old man who’s been doing this for forty years. It’s my fourth company. And so the ability to, you know, just go through LinkedIn and talk to all 2,000 people that I know Mhmm. It is an advantage. Right?

Scott Shagory:
So Oh, yeah. No. I think institutional knowledge just runs deep. Usually within the financial space, whether it’s on the asset management side, it’s around the commodity side, it’s, you know, elsewhere, even just, you know, iBankers and stuff. That stuff does matter. They they like telling each other stories or having them, scars and all, about, you know, the whether it’s the four days they spent on a deal and were not able to sleep to, you know, particular market crashes and opportunities with a big score or, you know, not a not so big score. But to be able to come together and be able to go ahead and, you know, share that that you and your team have that and can, you know, evangelize on behalf and understand them at a very deep level, I think, just often resonates well, particularly if they are concerned about again, I jokingly referred to it as hell 9,000. But if there is that fear barrier and you can punch through that with what is their biggest pain or set of pain points, and they’re happy to share that.

Scott Shagory:
And all of a sudden, the the, the, the swords or other things come down and the trust level just spikes almost immediately. You know, you’ve got people who are listening to you instantaneously like that, which is hugely powerful because most selling processes just meet a lot of indifference or apathy. And that’s, you know, not not the case here. It sounds like yours. Would that be fair to say?

Tom Blair:
It’s the implementation which, you know, feeds to what is the misconception of AI and and but it’s almost perceived as a religion. Right? Something you have to adopt, a change in reasoning style. And it’s not that. And it is and it also points to what we found to be really effective is, you know, because you can implement them so quickly in a quarter, 75 man hours to a 50 man hours or so. Yep. We do that, you know, like I said, without touching their data or messing with their systems. So if they have spreadsheets that can stay in place, if they have a CRM that can stay there. And so the interaction to do a deployment is one hour per week.

Tom Blair:
So we spend good quality time, maybe an hour and a half, once a week for the six or eight week deployment. And through that time, that team is doing their work anyways. Right? Today they have their quarterly report. They’re gonna hand crank it just like they’re doing regardless of whether or not I’m there with them. And through that process, at the end, you have an automated report that’s fully now, you know, traceable and clickable, and it’s dynamic, and it’s awesome. And you have the hand cranked one, and they should if they match, then the integration’s done. Right? They’ve already done their work. And so we can do that whole integration for no cost, and that’s the difference is that you learn you build the partnership.

Tom Blair:
It’s no risk for them, and at the end, they can see the results. And so that’s the operational leverage we found that allows us to get in the door and, you know, and that’s all you need.

Scott Shagory:
Yeah. No. Is most of it cloud based or are they or is there a lot of on prem? I know, the financial space can be a very, if you will, diverse one in terms of what that hardware infrastructure looks like. Yep. Do you guys end up dealing with multitenancy and all the usual things, or does it tend to be one type?

Tom Blair:
It’s truly custom. Ensemble systems like that are abstracted from the rapid technical change that takes place in AI. Right? No company wants to deal with that. They wanna control their costs. They don’t want runaway reasoning costs. So we abstract all that. And with that comes the ability to run locally without a cloud connection. We had, one insurance company in Germany that wanted to keep all their information within the borders so you can have, you know, you can go to that level, at, you know, but, so it’s it it depends on cost, performance, or jurisdiction, but the but you can do whatever you want.

Scott Shagory:
Yeah. No. That’s I’m not surprised to hear that, but I know some have some very specific requirements or just wants for reasons. What do you think? It’s a question I always like to ask because there’s always an interesting answer, which is, what do you think is most misunderstood by the market today about what you and the team or the company do now?

Tom Blair:
I think it is that religious fervor that everyone thinks we need to adopt AI. Almost every conversation I have, especially with public companies, they’ve all been told now by either their boss or an AI committee that’s been formed that they have to do something with AI. And, you know, like, I’m not a, you know, a CPA. They’re not AI specialists. So it’s like Mhmm. And bridging that gap has been difficult for people because they think holistically. They’re not thinking at an atomic unit of work level. Mhmm.

Tom Blair:
And so that’s really why one of the reasons that AI agents will have a profound effect on all industries is that it’s one one unit of work at a time. Right? And with that comes a business process reengineering that you can start to think through new ways to do processes for, for, you know, for fractions of it off of the cost of doing it now.

Scott Shagory:
Yeah. I mean, I even think of the work that I do, which is relatively recent since I left my corporate role. It has changed in a number of ways as I have utilized, first, you know, just the the simple, you know, log in, play with chat GPT, play with Grot, play with to then moving on to browser based systems to going ahead to doing something else, and I haven’t had to change a great deal. I’ve just been able to utilize and do things much more quickly or just, again, leverage the platforms in a way that I hadn’t been able to do as efficiently. So I use chat and the other platforms as a Socratic teacher for things that I’m not either as knowledgeable about or wanna prepare in some way. I could do that before, but it would be a much more arduous task, And it just gets me that much more quickly down the road to ask better questions and to understand something at a deeper level to then go on my own and do other other stuff. And I think that’s,

Tom Blair:
You know, I just run a whole company with just a bunch of engineers, distributors engineers. No employees.

Scott Shagory:
Yeah. Nope. No. It’s it’s quite it is quite remarkable to see, and we’ve only really scratched, I think, scratched the surface of what is is possible and that it, yeah, as a whole will be, I think, very beneficial for for all of us if we adopt it and think about it in in compelling ways. Yeah. Well, why don’t we go ahead, go ahead, excuse me, and take a short break, and we will be right back. Purple Finch Group helps technology CEOs innovating at the edge of chaos create operational fitness across their technology stack and team of talent. Achieving significant and measurable scale, selling your story, and finding time to think and plan can feel like a bridge too far.

Scott Shagory:
So if you’re a CEO whose company has found product market fit but is struggling or you’re a CEO whose company is looking to accelerate existing momentum, we can help with proven blueprints for optimal scale. Shifting from incremental addition to exponential multiplication can be absolutely achieved with less effort and stress. To get in touch, I can be reached at Scott with 2t’s@purplefinchgroup.com. Now back to the interview. Well, we are back with Tom Blair from iClerk. And let’s go ahead, Tom, and shift gears just ever so slightly. What type of work does the work that you do now give you joy and energy as a CEO?

Tom Blair:
Yeah. During the break, I was telling you that I have just a joyous life right now. I’m Mhmm. You know, later in life, my kids are grown, and, I really get to fulfill this, these sets of ideas that have grown for a long time. So it’s really fulfilling intellectual curiosity and, you know, fighting boredom and all the things that come in your, you know, in your sixties. Right? And, you know, and, but there’s other things about starting companies that, you know, if you ignore the staring into the darkness in the abyss every Monday morning, things like that, there’s real joy in, you know, I work around a lot of students. I, you know, when you’re a poor young startup, you hire a lot of students. I’ve always done that.

Tom Blair:
I also hire north engineers that live in northern cold places because it’s dark most of the year. They work more. You know, it’s awesome. They’re usually cheaper. It’s and so I’ve got all these philosophies. So I’ve got this band of married people across 10 different countries that have over the last thousand days, we’ve formed this great community. There’s just brilliant folks, you know, everywhere from Rotterdam to Edmonton. Right? In the country of Georgia and and, you know, all literally all over the world.

Tom Blair:
All over the world. And and, and we operate through our dog food our own technology. We have agents that record every meeting and transcribe it and do it by role. And so you can time shift. I don’t need to attend my engineering meetings. I can, again, time shift it and the engineers act differently when I’m around. So I don’t you know, I’m not around and then I can see what they’re doing. I and it and it you know, obviously, all the marketing and the sales have been Kevin Bacon ing it through life, just on connections and word-of-mouth.

Tom Blair:
And it’s a new way to run a company. I’ve been a section sixteen officer at a NASDAQ company, and I’ve, you know, started to run and I had hundreds of people working for me. And this is different, and it’s a different time to build a business. And I’ve never been more inspired, including when I was doing research at Kodak in my twenties, and it was just like the perfect world. And so I’m just inspired. It’s fantastic. I get to help young people and watch them grow, and they’re brilliant, and they’re naturally adapted to these tools. I get to speak with people across the sector and always learn every single day.

Tom Blair:
I’m learning all kinds of things that I had no idea about. You know? It’s like walking through the forest. You think you’re alone. You pull over a rock, and there’s, like, a thousand ants there, a whole colony that you didn’t even know about. Right? I feel like that every day of my life. Right? And that’s the beauty, that’s the essence of what I do. I control my own time. I’m pursuing my intellectual, you know, what’s haunted me for four decades since I went to school in the eighties.

Tom Blair:
You know? So, so, you know, I encourage everyone to start businesses to go full risk in their sixties. No. Don’t do that.

Scott Shagory:
Yeah. Well, when you look at the stats, I mean, you know, obviously, we see 20 year olds getting lionized for things, holding teenagers for the same thing.

Tom Blair:
See, I’m not looking for the lionization at this point. It’s a different goal set. Right? You know?

Scott Shagory:
Yeah. But when you look at the stats, you know, there are lots and lots of firms

Tom Blair:
It is.

Scott Shagory:
You know, that we all know and utilize or, again, consume their products, whatever it happens to be, that have actually been founded by people in their forties, their fifties, some of them in their sixties.

Tom Blair:
Yeah.

Scott Shagory:
Yeah. And, you know, all the cofounders were in those age groups and stuff. So it is not just a young person’s game that we perceive it to be. And then when you look globally, usually, certainly in Asian cultures that at age, being older, shall we say, is, you know, perceived as being enormously valuable too as opposed to our youth culture here in The US in particular in the West in general.

Tom Blair:
So It’s much easier to do when you don’t have kids around.

Scott Shagory:
Yeah.

Tom Blair:
I don’t care what they say. Yeah.

Scott Shagory:
No. That’s that, that is, that is fair. Yep. Yep. What looking ahead for you and the team, we’ve covered it partially. Mhmm. What are you most excited about over the next eighteen months or so?

Tom Blair:
I’m, I’m I cannot humans are bad at exponentials. Right? And we can’t see the future. We don’t understand. In my career, you know, the Internet came along and gave us all this data. And just Moore’s law over my career is a million times, you know. And so I look back to where it was then and do, like, image processing, something that you can, you know we’re doing, I think, 16 pixels at a time across four four k by six k images on a spark station. I remember reading.

Scott Shagory:
Yep.

Tom Blair:
Yeah. And and and now, finally, you can see that Kodak sent me to school. How do you find a picture of a boy in a blue boat from five million images? They just purchased a photo stock agency. So we are building the image database, an object oriented image database with semantic image recognition using neural networks. Great idea. Forty years too early. Right? And so when so even though I know that it’s happening and I cannot look out even three years into the future and look at what impact it will have on enterprises. Because these AI scaling laws are I’ve I’ve got the numbers here.

Tom Blair:
It’s like, it’s growing between four and forty times per year. Moore’s law was what, two years every 18 months. And this is Yeah. It was every 18 months. To 40 times per year of inference improvement. And so I can’t even and so we’re trying to you know so far, our architecture has done well with the advent of LLMs. I mean, I had natural language PhDs on staff when LLMs, you know, dropped, and it was like, oh, good. Thanks.

Tom Blair:
And, you know, so how do you continue to keep up with technology when humans don’t change? Technology is, like, changing four to 40 times per year now. And so we’re operating in that gap between the static humans that don’t wanna give up their spreadsheets. You know, there’s 750,000,000. You’re going from 30,000,000 programmers, seven hundred 50 million spreadsheet users to billions of language users. Right? And so somehow, you gotta get from, like, my mom to, you know, over here. Right? And that’s kinda where these agents come in. So I can’t see the future. You know, no one can.

Tom Blair:
My adviser in grad school told me, don’t bet your career on neural networks. Right? He didn’t win the Nobel. You know? He’s not Jeffrey Hinton. Right? Mhmm. So I don’t know what’s right. I’m just trying to, one task at a time. That’s all I can do. That’s what I like about it.

Tom Blair:
It’s just, like, one step at a time. And that’ll end us up somewhere. I don’t know where in thirty six months, but it’ll be interesting.

Scott Shagory:
Yeah. No. I think so too. It’s just, you know, taking that small action, whatever it happens to be, I think is the most important step. And, while it’s a bit of a cliche, it doesn’t have to be perfect as long as it’s good. You know, it’s a bit like, you know, a fitness journey if, you know, one wants one where if you get a little bit better every day or maybe it’s every week. It doesn’t have to be some, you know, leap over the huge puddle. But go ahead and do something.

Scott Shagory:
Yeah.

Tom Blair:
And then, you know, find out where you are at.

Scott Shagory:
Yep. And and calibrate on it and then calibrate on it. See what works with whatever a team wants you to build.

Tom Blair:
That’s right.

Scott Shagory:
And then as they understand it and think, oh, well, we could do this, this, this. You know, John and his team need this, but Susan and his team need this. Could we implement that? And the answer, of course, was gonna be yes. You could go ahead and do that. So I think it’s, you know, not only an exciting time even though some will focus on the fear factor, but I think the next you know, you and I have been around a while. I’m gonna be turning 58 next week. I just think the next ten years are gonna be really very exciting.

Tom Blair:
Yeah

Scott Shagory:
I mean, it’s gonna be genuinely transformative. And I know we say that in tech all the time. But I do think that there’s just going to be it will look very, very different, from even the cycles that we’ve seen. You know, you’ve mentioned the Internet. We could talk about clouds. We could go for mobile. We could do the conventional journey there, but it’s, I think, like you say, whether it’s four or 40 times per year that there’s gonna be there is gonna be some really quite miraculous stuff. We’ll look back. It’s gonna change how far.

Scott Shagory:
And say, you know, in health care or in a number of areas, just, you know, wow. That was an accelerative and how much more accelerative it is and will remain. You know, it’s just not going to that accelerative curve, if you will, will, I think, continue for a long time, which is exciting.

Tom Blair:
It’s increasing.

Scott Shagory:
Yes. It is. Yeah. Yeah. Which is crazy to think. I mean, conventionally with the Incomprehensible.

Tom Blair:
Yep. Even though that’s all I do. I study this. It’s a passion and it’s incomprehensible.

Scott Shagory:
Yeah. No. I think which is what makes it exciting where if someone like yourself who actually has started with it since the early nineties and been thinking about it beforehand and, you know, been in the space for its good and bad points for, you know, four decades now, and you’re more excited than ever, and you can’t predict the future,

Tom Blair:
and you don’t know it’s And it’s still bad to me.

Scott Shagory:
And it’s still very exciting. Yeah. That just, I think, says just so much about what is essentially the potential. Yeah. Yeah. I guess this is maybe the final question here as we wrap up. You know, how is being CEO of the company, if you will transform you? And I know you’ve been CEO of several companies. You’ve sold them.

Scott Shagory:
So you’ve been through the rodeo many times now. This is number four. You know, either as a whole or with iClerk, what do you think has transformed you the most?

Tom Blair:
I touched on it before. This has been a different ride for me because the technology is really different, and I can do things that I couldn’t do with, you know, a dozen or two people before. And Mhmm. And so it’s been a combination of, you know, the nice thing about running a start up is that you really do get to control your time, and that’s a very powerful thing. If you can get to a point in your life or whatever it is where you really do control the bulk of your time, it’s a path to happiness. And, you know, I always look at life as a happiness curve where you’re really happy when you’re young and then you’re, you know, fighting life and having kids and jobs and getting fired and moving around and all that stuff. And then you get back and your kids go away and you’re happy again. And that’s kinda where I am in that happy curve.

Tom Blair:
And, that combined with doing a startup that keeps my brain going and, you know, and, I feel like I’m contributing back by helping people and and actually doing good, you know. So it’s it’s it’s, you know, it’s bridging that gap between idea and execution. You feel like you’re actually creating something. You can see an end result that’s tangible. It’s a win-win type of technology where the customer gets good value, and it’s, you know, just exceedingly fun to learn about these tools continuously. It’s just amazing. So

Scott Shagory:
Yeah. No. That’s cool. I always think I say this only half jokingly because it is absolutely true. What you’ve you’ve highlighted is I said, you know, they ask me sometimes what’s great about AI or AI agents? I said everyone gets a dopamine hit much, much more quickly than they ever would have. You know, a decade or more ago there were some projects. I’m sure we’ve, you know, both been on projects that lasted months and you didn’t know whether they were actually gonna pan out or not or what the outcome was going to be. Were we going to achieve the ROI? And here, everyone, we we it’s all compressed and we end up dealing with the fear right away.

Scott Shagory:
We then move on to actually the work itself and the implementation. And then you get to see the results pretty quickly and then iterate on it with very little discomfort. And I think once you get over that initial fear of a client or prospect, all of a sudden everything just starts to flow because all of us know or all of the team members know what the why is. And then we just simply move forward. We know who the owners are, and it gets solved that much more quickly, which is what most humans want is they want the solution very, very quickly. They wanna lose 10 pounds in two days. They wanna achieve some more, you know, marital bliss in some short period of time. Whatever it happens to be, there’s

Tom Blair:
Yeah.

Scott Shagory:
You know, AI at least as a solution can actually truly do that even if it’s a small step that becomes a set of bigger steps.

Tom Blair:
So agents take that immediate gratification down to one step at a time. And that’s, you know, that’s something that we can all get our heads around.

Scott Shagory:
Yeah. No. That’s great. In, as the final question here, how would you like people, those listening, those who will see this on YouTube, how would you like people to get in touch with you?

Tom Blair:
Sure. I’m at, my email is tom@iclerk.ai. It’s I c l e r k dot a I. And you can go to our website, check it out. I’m always happy to talk to you. You know, we’re in a period of discovery with these technologies. All of us are. Right? It’s a partnership.

Tom Blair:
And so it’s a fun way to learn. It really is. So, you know, please feel free to reach out.

Scott Shagory:
Yeah. That’s great. Well, Tom, thank you very much for carving out time to talk with us today on the Discipline Troublemakers podcast. It’s been a great chat. Always a wonderful opportunity to learn new stuff that I didn’t know about before, a new angle and things. And to all of you, watching and listening out there, we will be back with another episode very soon. Cheers.

Tom Blair:
Thank you, Scott. K.

Scott Shagory:
Thank you for tuning into the Discipline Troublemakers podcast, and we hope you found the episode valuable. If you did, please subscribe to the discipline troublemakers podcast on your favorite platform and leave a review. Putting this podcast together is a group effort, so I’d like to provide a shout out to Heartcast Media and their team for all of the assistance they provide with each episode. If you wanna get in touch with me directly, I can be reached at Scott with 2. is the sponsor of this podcast. We have another podcast episode out very shortly. Stay tuned.