Tanya Kohen: There is a question sitting underneath most enterprise AI conversations that doesn't get asked directly enough. Who actually leads this? Not who approves the budget, not who manages the vendor, but who takes ownership of where AI goes and how it gets built into the way the business runs. In most organizations, that question gets deflected, it becomes an IT project or a consulting engagement or something the CEO champions in a town hall and then hands off. The people closest to the operational function, the one who know exactly where the manual work is and what it's costing, rarely end up in the room where decisions get made.

Treasury is an interesting case. It sits in the center of how value moves through the business. It touches payments, forecasting, risk, banking relationships, working capital. If any function has both the visibility and the incentive to lead AI adoption across the enterprise, it's treasury. And yet that's not typically how it plays out.

My today’s guest Bojan Belejkovski is doing it differently. Bojan, thank you for joining me today.

Bojan Belejkovski: Thank you for having me.

Tanya Kohen: A few words about our guest's background.

Bojan Belejkovski is Vice President, Finance & Treasurer at Voltava, where he leads treasury, post-M&A integration, finance risk management, and AI adoption across the finance function. He has led treasury through some of the more demanding projects and restructurings of the past decade, including the $7B bankruptcy and integrations and $1.5B credit facility restructurings, and projects like TMS implementations across SAP, Oracle Fusion, and Kyriba.

Bojan is the author of treasury 2.0: Future-Proofing Finance with AI. He sits on the AI 2030 Advisory Board and treasury Masterminds board, and consults independently as well. His own framing of the work is "I turn finance from a reporting function into a decision engine."

Thank you again for joining today's conversation, Bojan, and let's just jump in straight into the interview questions. Before we get into specifics, tell our listeners a little more about you and about what you're doing today, what you're working on, and what your current involvement with AI looks like and how did AI become such a central part of your work?

Bojan Belejkovski: I've been in the finance world for over seventeen, eighteen years now, and I actually started out in legal. My academic background is in international and European law, and for a while I was thinking about European policy work. In my mind it was always about frameworks and processes and building something that sticks long term. But somewhere along the way, I started working in finance, then I took operations management course at the Wharton School of Business, and I remember this specifically because it was the “aha” moment for me and re-rewired me on how I think about my own career. I spent a good chunk in strategic roles across a few different industries: automotive, renewables energy, gaming, sports and entertainment, real estate. That is a lot, and I think the range actually matters because treasury looks different from the other, even within the finance looks different in any of these industries, and it forced me to think about the fundamentals fast rather than someone settling into a specific company playbook.

And I've been treasurer since 2019. And even today as I sit in that position, I think I try to expose myself to learning more and more, and I'm going from law to getting an executive education, to getting an MBA, to getting CTP and all these courses. And why is that I think it is important and relevant to the AI part of the story because I got myself into the AI space, like maybe everyone else, through ChatGPT when it came out a few years ago.

I think curiosity paired with ChatGPT coming out pushed me into the AI space and started experimentation and hands on almost immediately, and then shaping things out of curiosity into how can I apply this hands on, that I mentioned, implementation and experimentation eventually turned me into something more formal for me. So I started taking courses, specializations on AI, and I started playing with AI myself.

That led to writing the book, the “treasury 2.0: Future proofing finance with AI.” Writing the book was actually great for me because in my mind I was like, “OK, I've done this and this and this”, but it forced me to organize my thinking in a way that my day-to-day work never does. So you can improvise your way through a pilot and to a project, you cannot improvise through just explaining to someone else in writing. And again that led me to be interested in AI, AI consulting, focusing on frameworks, governance, workflows, and centering around AI pretty much in anything I do today.

Tanya Kohen: That's a great example of how curiosity, taking courses, just being open to learning new things and then experimentation is a good path into truly getting new skills and bringing additional value to your area of expertise. Thank you for this story.

In most organizations, people who can build AI solutions don't carry the creational friction. And people who carry the friction don't lead the build. So when you walked into a room of executives to make the case for treasury owning this work, because not only you're experimenting with AI — you're also advocating for it in everything that you do. So treasury owning the work, not AI, not a vendor, but what was the argument that actually shifted their perspective?

Bojan Belejkovski: The tension was always there and it's pretty simple. If you think about IT, it knows and owns the code, and the code for a solution, but it doesn't necessarily know the outcome that the treasurer needs, and vice versa — the treasury team and the treasurer know exactly what the outcome needs to be. But I think historically we have not been able to build it ourselves within the treasury, we've always had to translate our needs into specs, handing it to a developer, waiting and answering questions and all that. So I think that was the gap that AI closed. And it is for the first time, I think, that AI closed this gap. So one person, someone who actually understands the treasury problem, can also define the need and can get the most out of the building themselves. I can give you a concrete example.

So if you're managing roughly, let's say, 1520 banking relationships, and that used to mean logging into different portals, pull balances, multiply that across multiple team members, and then you're in a mess. And we've lived through this, it's a very specific treasury problem, or was a very specific treasury problem. It's never going to be a priority for IT, may not even be for your CFO, because things are working and “let's plow along the way”, and there are no structural gaps and the company is doing whatever it needs. Because you might need to spend two hours, but everyone can still see the cash balances. Now, what actually shifted is the executives, with the executives it's not necessarily a single moment, I think it's seeing, as time went by, proof of concept and things playing out in front of them. So visible time saved, money savings. And it's not just a pitch deck, it is something that has worked in even in a limited way, but changes the organization.

So the reception is not uniform, we can be honest about it. From my personal experience, I've worked with forward-looking executives and they understand even within the first meeting, when they see the demo — they clap and say, “Good job, how we can scale this.” I've also worked with more traditional ones, less comfortable to accept the change. We've always done it mentality. So you need more time, it's less impactful. But I think even there, it is able to build this in a solution so that you can make your case to lead, to put treasury in a position to lead.

Tanya Kohen: I like how you talk about it, that it's ROI first. As you know, we're really close to finance, I'm also treasurer in my past, so I agree 100%. When ROI is there, it's a different conversation. But I like how you also mention a pilot. So doing something, showing some example of the outcome, I think it is strong. And from where I sit currently, working with treasury teams and with CFO, I can agree 100%.

That's what moves the needle when people start realizing that AI is going to help us close certain gaps in our process, in our work, or be the decision-supporting engine for us so we can achieve what we always knew we wanted to achieve, but quicker. That's a great starting point.

Let's talk about the pilot itself. A lot of treasury leaders would struggle to know where to even begin with something like this. So where did you start and what was the first use case that you went after and why on that one specifically?

Bojan Belejkovski: I started with use cases. On my end, personally I was very curious and I started using AI tools broadly. That's when I started to get myself more and more interested in consulting as well. But what I was looking at is the market was saturated even from the beginning with cash management and forecasting. And even if you look at the publications out there, they would say that 60-70% it's all about cash forecasting. And I tried to avoid it quite honestly because again it's a crowded space, well funded, purpose built, and I was just not interested in building a mousetrap that is not necessarily needed. So instead I focused on problems that everyone is trying to resolve because treasury in one way or the other manages them. But they're not as interesting maybe as forecasting and variances on the forecasting and and such.

So the first things I got working on were foreign exchange and insurance risk management, and also started to work on credit agreement compliance and modeling. Because all the companies have either foreign exchange, insurance risk, insurance risk for sure, and almost everyone has a credit agreement. That's what I was thinking, that's an underserved area. But also separately — accounts receivables, which, as you know, we worked on a pilot with WaveAccess — it is a great example because together we found that the data is scattered across different ERPs and different customer systems, and it is never clean, and if it is, it is very rare.

That kind of fragmented low visibility workflow is exactly, I think, where AI comes into play and adds a lot of value. The project we worked on — the accounts receivables pilot — I would say, it reinforces something that I already believed in. The value itself isn't necessarily picking the flashiest possible use case. It is picking the one that is causing probably the biggest headache at the moment and is also a low-hanging fruit.

So I'll just add one more thing. If things had to boil down to a starting point, I'll basically say three filters. If it's already being solved by an existing vendor at a scale, I don't necessarily get interested in, and I do walk away. If the data is fragmented enough that a human is doing the manual work between systems, I think that sends a strong signal and, what you and I picked on when we worked together, “can I get to a visible proof of concept fast in weeks, not in months and quarters?” Because speed of proof is what actually builds the executive trust as well.

Tanya Kohen: Well, this is a really valuable takeaway, I think. If nothing else, I would recommend that listeners take away these three tips from you: how to approach, how do I understand is this the pilot material, so to speak, or no. Because this conversation about buy versus build has been around for a while and now with AI surfacing again, like do we build a tool, do we buy a tool? But actually let's solve the real problem. Buying or building is not the problem, the problem is the business performance improvement or just continuous improvement or just fitting the goals of growth for the company. It's always a business goal that is being solved. So let's just use technology to help us solve it, right?

And then there are certain pockets of opportunities, even with all the great products existing. But as you just said, is there a manual process? Is there a way to get to the decision making quicker or with more data? Because no one wants to make gut-feeling decisions. We want data supporting all our decisions, but do we have time, as treasurers specifically with smaller teams? We don't always have time for all of this. So I agree.

Also you mentioned AR in the pilot that we did. I think the cash cycle and working capital management are the areas where treasurers sometimes don't just get enough capacity and time to dive into. Which is sad because the cash cycle is the blood of each company. I think that's the area where we can work cross function a little more with the broader finance teams and AR/AP and all of that. Thank you for mentioning this example.

You've made a case for treasury to lead this internally, but there is a difference between running a successful pilot inside your own function and becoming the model to the rest of the business. How does this conversation land with other functions? And what does it actually take to move from treasurers doing something interesting with AI to treasurers leading how we do this across the enterprise?

Bojan Belejkovski: If I can summarize this in a way, it's not necessarily pilot to building enterprise model and everyone knocking, you know, all departments coming and knocking on the door at once. I think it happens in a more organic and honestly slower way than people from the outside might think of. Most companies now have some kind of AI task force or SME group. I know, I'm in one. And these cross functional forums where people share what they're trying is probably the best start.

For example, once my results become visible, I guess real-time savings, dollar, or time saving process improvement, and people actually getting excited about AI, and I get it not everyone will be, but most. Other functions also want to understand and they want to approach you, and honestly I think it's pretty much daily at this point where people want to understand how I approached a certain problem. So it's less about treasury issuing some kind of a top-down mandate. We don't have that authority anyway. I think it's more about becoming a reference point and especially now with these AI capabilities, when it's imposed (I don't know if imposed is the right word), but when it's mandated, imposed, however we say, from above, you get compliance, not conviction. Compliance probably evaporates in the moment, when attention is moving elsewhere.

What I'd also add here is that becoming that reference point internally comes with responsibility and it's not I guess anticipated early on. Once people start coming to you, you decide to be the standard setter, maybe de facto standard setter, even without the title. The governance, the framework thinking and all I mentioned earlier becomes even more important at this point because whatever pattern treasury sets, functions are going to copy, and mistakes are included, of course. So I think pilot to enterprise model, there are certain steps, and I've become a lot more deliberate in documenting not just what worked but what failed, what the guardrails are and precisely we know what ended up happening and I know that other teams are going to take this and build on top of this, and they avoiding mistakes that I potentially made is gonna make us more successful from an enterprise standpoint.

Tanya Kohen: Again, the better the communication is internally between different departments, the more everyone benefits from it, from learning from each other and, as you just said, just exchanging experiences about what worked for whom and then being on the same committee. And it should be, I think, be more formal, so not just enthusiasts talking about what we can do better but proactively brainstorming and trying to work together on building something — that's the key.

One thing that comes up a lot when I talk to treasury leaders about AI rollout is the budget question, as treasury typically doesn't control the technology spend. How do you build the internal coalition behind the project when you're the business sponsor rather than the budget holder, and how does this conversation shift as you start to scale, if you have any insight there?

Bojan Belejkovski: I think it all comes down to ROI, and honestly that's the language that works. Every solution I've brought forward is framed around some type of savings. In more concrete terms, not just abstractions like efficiency, transformation, freeing up resources and whatever that means. Everything is measured, right? So nobody can argue a clear positive ROI and I think, frankly speaking, most executives have heard enough vague AI pitches by now. So they might even become allergic to that if you don't have a grounded number. And I don't think there are any workarounds because it's pretty simple. Everything is built inside an approved system, approved governance, approved policy from day one. At least that's how I approach things.

I've seen people rush with certain solutions and they hit a wall somewhere within the next whatever period of time short term. I am big on governance and framework and that's the premise of my book as well — stable structure that enables skill deployment for long term success. And I believe that's important because, again, it does not lead you to breaks after you go live, let's say, but actually you have a process that you can trust that is built on already approved systems and apps and frameworks and whatever you have internally.

So building the coalition without owning the budget means learning to talk in the language of whichever executive or stakeholder is in the room: finance leadership, that's ROI, risk reduction, with IT it might be something else in terms of following their process application and an approved system that you have. And as you scale, I think that coalition building that I mentioned gets easier and not necessarily harder because you're feeding on already approved and already agreed-on processes. So once your framework is in place, let's say you build something on Microsoft or whatever the company uses, I think the new use case becomes an incremental add-on instead of a brand new proposal, a brand new pitch, a brand new every single time. So you're just extending the pattern people already trust. In that shift, I'll just end with this, from justifying the category to extending that category is really the game-changer and the whole game once you're past the first, let's say. year.

Tanya Kohen: Hopefully results also support the story, right? Because that's the goal as you move. But that's a luxury. Before we get to the first results, that's the hardest part to overcome. Just keep working together, keep collaborating, keep listening to each other, and then keep building and then we see the results. That's a great strategy.

I'm sure some of our listeners will be treasury professionals who feel that leading AI across the enterprise isn't their job, that their mandate stops at the treasury function. What would you say to them? And looking a few years ahead, what's the real difference between the treasurers who took that broader ownership and those who didn't?

Bojan Belejkovski: I would slow the speed on this one. I'll be honestly direct and share my personal perspective on this. The real risk I believe is losing relevance and eventually losing the job itself. Doesn't sound great, but it is the reality that we're gonna face. This work is up for grabs, even right now some people learning, testing, failing, trying again. Others are doing nothing, waiting for something to become mandatory or waiting for someone else to figure it out first. That's okay, but you don't necessarily follow after that. I'll give you an example here. Very recently I talked to someone who is in a treasurer position at a very large company, and they said, “We got our first AI goals”, and I was expecting to hear some huge workflow that they're working on, then they said, “We are asked to use Copilot.” And I guess no one told them that they're the ones who are getting impacted by changes. And not because AI replaces treasury people outright, but because the treasurers who learned this are simply gonna be trusted more. And the ones who didn't get a chance to do this at work or did not try outside of work, I think their role is quietly gonna get narrower. And I don't think it's what is being said out there is whoever knows AI is gonna replace you just because they know AI. It is more than that because you need to also understand how to work with that AI and guide it to narrow the gap that currently exists.

And things are moving fast, AI develops fast. If you're waiting too long, you end up chasing the tide instead of riding the tide. And that's expensive for the companies and the CFOs and the CEOs, they understand that and they know that. And I think looking a few years out, the treasures who learned and who leaned into this will expand their seat at the table. I believe that they're pushing in a way the treasury's delegation of authority beyond what is the standard. So if you think about a few years ago, DOA and in today's world with AI, you would see more companies shifting from just credit agreements, routine approvals into real decision-making influence, just waiting on capital allocation and risk strategy, on where the business places its bets in a way.

I'll finish with this. The ones who did not stay confined to execution, and are reliable, but not in the room when the direction is set, they're gonna be impacted, right? Eventually someone else in another function or outside vendor or more ambitious sphere, will end up claiming that mandate instead. So the seat doesn't stay empty, just, I guess, it goes to whoever showed up with more skills.

Tanya Kohen: Well, that's the motivation to start learning for anyone who's not yet started.

All right, rapid fire questions are the final three questions that we ask our guests. Number one for you, and you can say as much or as little as you want: what does an invention mindset mean to you?

Bojan Belejkovski: Invention mindset, I guess to me it means treating every manual repetitive task as a prototype waiting to be built, not just something that you push through. I think you raise your hand if you're in that type of environment and you ask someone how you can get help. And I believe now everything we said here, I think it's the difference between doing the work and asking whether the work should exist in that form at all.

Being mindful of that, and I'll give you an example. That's how I approached FX and insurance risk agents in one of the questions earlier. Seeing a manual recurring process and then asking what if this tool just run it by itself and then provided the feedback you needed without you doing the manual work. I think that an invention mindset also means being comfortable with something new, even half built, but something new and useful and then you kind of take it to the next step and then try to perfect it the way you trust it.

Tanya Kohen: What do you think is missing in today's tech stack for finance and operations?

Bojan Belejkovski: I would say it's not more tools. I think there are a lot of tools for small, medium, and large companies. I think it's now shifting to accountability and durability. Most apps, or solutions, or tools — not all, most that I've seen — I think they're built to go through the next quarter, through the next year end, but not necessarily on data frameworks designed for long-term strategy. So you end up at some point with a solution that works fine and it's aimed to get you through the day. But I think it's creating a mess in the future. That's the whole reason I push for data hygiene with AI. Most finance leaders should think about foundations first, even though they start building the actual AI tools later.

One other thing I will add is one thing that's missing, and I wouldn't say this is the technology stack, I think it's more like the approach and the frame of mind. I think people should start thinking about more tailored tools versus something that's out there already on the shelf with algorithms tested on whatever that's not applicable to you.

Tanya Kohen: If you could improve one thing in a business, not necessarily tech related, what would it be?

Bojan Belejkovski: I'm gonna almost repeat myself, I would say — adoption mindset. Especially in AI culture, and especially for small and medium sized companies, and I've just seen so many. They're the loudest that they want to make a change because they're small and they would use AI, but they're so hesitant to actually embrace AI.

And there is such a big misconception on what AI can do for them, and I think that's ironic because they would be the ones that can gain the most and they can move a lot faster than large enterprises because they — the large enterprises — have that weight of legacy systems and everything that the structure brings, so they just need to get out of that way and start embracing and start thinking. So that's what I think gonna improve.

Tanya Kohen: Bojan, thank you so much for this talk, that was very insightful and interesting.

The reason the treasurer is the right person to lead AI adoption isn't just about knowing where the friction is inside their own function. It's that the treasury sits at the center of how value moves through the entire business. I think we've illustrated this today with the examples that you mentioned and this whole conversation proved that.

That position comes with visibility that most other functions don't have. So, and as Bojan’s work shows, it also comes with a responsibility to take that visibility and do something with it rather than just wait for something else to do that.

So the questions he walked through were how to make the case, where to start, how to build the coalition without owning the budget, and how to turn a pilot into something the rest of the business can follow, those are words going back to regardless of where you are in this journey. Thanks again Bojan for your insight.

For our listeners, subscribe, leave your feedback and your reviews, share your story of investment with me on LinkedIn, and see you next time.