Is the Treasurer a dying breed?

Christian Gjerde has spent more than 20 years in corporate treasury, working across multinational environments where funding, liquidity, risk management, systems, and operational resilience sit at the heart of the finance function. He is currently Head of Treasury and Investor Relations at Elopak, a Norwegian company listed on the Oslo Stock Exchange. Across his career, he has seen treasury technology evolve from traditional system support toward AI-enabled decision support. We asked him what that shift actually feels like from the inside.

Every other voice in this series – the academic, the technologist, the AI itself – has argued from a position of relative comfort. They describe the transformation of treasury from a safe distance. Christian Gjerde brings the perspective of a practitioner with more than two decades in corporate treasury. His experience spans complex multinational treasury environments, where the function is not just a cost center – it is infrastructure for companies navigating growth, volatility, funding needs, currency exposure, and operational resilience.

He has followed the emergence of a new category of treasury solution: AI-native platforms built from the ground up rather than retrofitted. The conversation that follows is the most grounded in this series – the view from a practitioner who has seen treasury systems, processes, and priorities evolve over more than two decades.

“AI will not blush if it makes a mistake. I will. And that responsibility will always be mine.”
Christian Gjerde, experienced treasury professional with more than 20 years in corporate treasury; currently Head of Treasury and Investor Relations

Q1.  Setting the scene

Before we talk about AI: based on your more than 20 years in treasury, how would you describe the operational reality of many treasury functions today, and what does that mean for their capacity to focus on anything strategic?

Across my career, I have seen many treasury functions remain more underdeveloped on the system side than they should be and being too reliant on manual processes, and the tool that we all want to move away from, but can’t Excel. Companies may have a treasury system in place, but it is often not fully implemented or developed. From a system support point of view, the setup can therefore be quite basic – basic accounting, basic registration of financial instruments – with too many processes still manually based, even today in 2026.

The operational reality in many treasury functions is still more manual and limited in terms of data and data insight than it should be. That creates capacity challenges in small teams. It creates less timely decisions, because a lot of the data treasurers make decisions on needs to be collected and analyzed before they can actually do anything with it. And it also entails significant operational risk – if anyone is on sick leave, it can create a bottleneck immediately.

This has been a recurring theme throughout my career, having worked in treasury for over 20 years in bigger and more mature teams. That is where I believe treasury functions need to move: better system-based processes, built on something that also captures future potential within the AI space. Too often, treasury is still about operating day-to-day activities and then trying to focus on key strategic priorities as and when there is time for it.

The reality check: A highly experienced treasurer, reflecting on more than 20 years in the profession, still sees many treasury teams spending the majority of their capacity on manual operational tasks in 2026. This is not an unusual situation – it is the baseline from which the AI conversation actually begins for most corporates.

‍

Q2.  The technology inflection point

How do you see AI-embedded solutions changing the way treasury teams evaluate technology? What does that distinction mean to you in practice? Does it change the questions treasury should be asking, or what teams should be willing to pay for?

Over the years, I have worked with or evaluated a number of treasury systems. Having worked in treasury for more than 20 years, my experience is that many of them are quite similar. They follow the same types of process flows, with different strengths and weaknesses, but are broadly based on traditional treasury operations.

AI-native newcomers in the space approach treasury from a very different perspective. They are really focused on harnessing the data that can be available to treasurers – to make better, more informed decisions, in a more timely and proactive manner. And hopefully they will also support treasury teams in creating their views through agents that run in the background and generate a reality rather than waiting to be asked.

The value drivers will pretty much be the same – managing debt and cash positions, making sure the company is appropriately funded, managing risks. But how proactive treasury can be in its decision-making will dramatically change. Some solutions are now being described less as treasury systems and more as treasury assistants. As someone who has worked in very much the same way for 20 years, that prospect is genuinely compelling. That said, I also think about what it means to shake someone’s hand at the end of a deal. At the end of the day, someone gets the credit – or takes the blame – and that will never be delegated to a system, however intelligent it becomes.

Many companies now have broader AI strategies, and moving in that direction on the treasury side can support future ambitions for finance and for the business as a whole. Large corporates often sit on significant data sets – on seasonality patterns, customer behavior, market dynamics, funding flows, and currency exposure – that treasury can harness if the right AI capabilities are in place. Treasury aligning with that direction makes complete sense.

Did it change what I believe treasury teams should be willing to pay? Yes, because if it becomes an assistant, that is something you can count against FTE rather than just system cost. Many treasury teams are lean. They may have plans to increase headcount. Maybe this avoids that increase – if not fully, then partially, over time.

“The system historically gives us what we ask them to. AI-native will automatically start generating the information it believes you should have access to – to make the decisions based on the criteria you have set. It is a very different way of working.”
Christian Gjerde

‍

Q3.  Fear or excitement?

When AI is raised in the context of the treasury function, is the subtext usually ‘this is an opportunity’ or ‘we can do more with fewer people’? And how do you personally see your cooperation with AI?
It’s a great question, and something I have discussed extensively in the context of treasury technology and AI. When senior management teams consider whether to explore AI-native approaches alongside more established players, the important point is to assess the opportunity properly before deciding.
The reasoning is clear: if you go straight to one of the established players, you know it will do the job – but it may do it the way it has been done for 20 or 30 years. You could be looking back in five years thinking about the opportunity you missed. For treasury teams that are not locked into a fully mature system landscape, there can be space to take a bit more time, run a proper proof of concept, and see what the next generation of treasury technology can genuinely deliver before committing. That curiosity is important.

For treasury, it is an opportunity to do more with less – but that does not necessarily mean cutting headcount. It means the function can contribute and create more value with better systems. My to-do list never seems to get shorter. This is about creating more value, not about going the other way.
In a volatile world – geopolitical conflicts, a move away from globalization, more protective trade environments, more volatile financial markets – the importance of treasury and having good systems and genuine insight will only increase over the coming years. AI will become a part of everyday life in treasury. I think it will make us even more strategic, even more relevant, and even more widely focused.

The signal:  No fear of replacement here – but something more nuanced: an experienced treasurer who sees AI as the tool that may finally let the function to better deliver on what it has always promised to the CFO.

‍

Q4.  The blind spot

Many treasuries operate in markets with real currency exposure and geopolitical complexity. In periods of interest rate divergence, trade uncertainty, and FX volatility, where would you draw the line between AI-autonomous processes and human-led decisions?
I think you can have several layers of agents doing different steps of an analysis-to-execution process. Agents generating hedging ideas, proposing them to other agents running sanity checks. You can create extra layers to avoid having AI make decisions entirely on its own. There is large complexity in many of the decisions we make, and they are not always purely data-driven. There are other aspects that need to be considered.

But in the end, we are the ones who will get the praise – or the punishment – for the decisions made. There is something I keep coming back to: at the end of a deal, you want to shake someone’s hand. You want a human being who owns the outcome, who you can hold accountable, and who will feel the weight of getting it wrong. That does not change because an algorithm ran the analysis. Coming from a pre-AI generation, I think we will still want to have a say in the discussions and ultimately in the decisions. AI needs context. Context is important for its ability to make recommendations. We also know that AI can generate hallucinations – so it needs to be vetted and considered carefully. The better it becomes, the more autonomous it can be. But ultimately, we need to make sure that we are running AI. AI is not running us.

Take the foreign exchange market as an example. Very well functioning, traded in trillions a day. Analysts and economists sitting there with their views on where EUR/USD will go. And what we know statistically is that they are more often wrong than right. If you train AI on those types of things, it is somewhat bound to fail because there are no stable patterns – there is a lot of psychology in these markets. A war breaks out. A tariff is announced. These are inputs that always need a human touch, in my view.

The line Christian draws:  Backward-looking, analysis-driven tasks – reconciliation, reporting, cash concentration within treasury accounts – are easier to delegate. Forward-looking decisions with direct economic impact and significant complexity – hedging strategy, funding decisions, anything touching the external world – stay human-led, at least for now.

‍

Q5.  The wishlist

If an AI-native treasury functionality worked as expected today – clean data, trusted outputs, proper guardrails – what is the one task you would hand over immediately, and what is the one you would never delegate?

Something that will be easier to delegate is the reporting side of things – especially internal reporting. Deviation analysis, financial account closings, things that are not directly affecting something economically impacting in real time. They are more about looking back on what has happened. You cannot change the past, but how it is reported you can do incorrectly. With appropriate guardrails and your own sanity checks, I think you can start moving that part into AI relatively soon.

The same goes for certain cash transactions – automating cash concentration processes, moving money between treasury accounts rather than externally. You might sweep the wrong amount on a given day, but ultimately you are not sending money to anyone external. It is recoverable. Everything that does not directly impact economically will always be easier to trust to AI.

When it comes to transactional execution and strategic long-term decisions that will impact your financials in a more significant and longer-term way, those are things I typically want a sound check on. But that said – AI will develop. I am sure we will give it more and more responsibility as long as guardrails are in place and we are doing our own testing.

And yes, we will make mistakes because of AI – but humans make mistakes too. I have made mistakes. You trade for an FX transaction and you buy the wrong currency. It happens. What matters is that you have the oversight and control structure to catch it. AI will not blush if it makes a mistake. I will blush if I go to my CFO and say: I made a mistake. And that responsibility will always be mine.

“Cash is king – but data is queen. The data will feed the AI. And as an experienced treasurer, this is opening new opportunities to become an even more valued contributor to the strategic execution of the business.”
Christian Gjerde

‍

Q6.  The message across generations

A treasury professional entering the field today will spend most of their career working alongside AI. What is your advice to them – and what would you say to your more seasoned peers?

Something that has always been important to me, and I think will continue to be: to become a good treasurer – or whatever profession you pursue – you need a fundamental understanding of the value drivers of that function. They need to be ingrained in you.

AI can support your processes, but you ultimately need insight and understanding to know what is right or wrong, optimal or suboptimal, in the decisions being taken. I have always had a lot of pride in this: if I do not understand something, I sit down and learn to understand it. That still needs to be the basis for anyone starting a professional career. You cannot expect AI to take you through your career and make you successful unless you yourself build the knowledge and understanding. There is an 80/20 rule – you do not have to understand everything, but you have to understand it well enough to know, in a qualified way, that the decision you are making is the right one.

Do not let AI become a cushion for not learning yourself. Do not let it do the job for you before you understand the job.
When I talk to more seasoned peers – in treasury and across finance – I see two camps. Some are cautiously curious, asking how AI might impact their future job prospects in a more negative sense. Others are thinking about how AI can support them in becoming better at the job. I see both in myself, honestly. Over time, I have developed my own understanding of how AI can support treasury work and professional development.

And I think this is the moment where you decide whether you want to be part of it or not. If you have 10, 15, or 20 years of treasury career ahead of you, treasury will look very different in how it operates today versus 20 years from now. The value drivers will be the same – but they will be more complex, more integrated, and hopefully more value-creative in their outcomes. This is an important decision to make.

The generational divide:  Christian sees both responses among peers – cautious anxiety about job security and genuine curiosity about what AI enables. His advice is the same in both cases: build the fundamentals first, then let AI amplify them.

Editorial Note
This interview was conducted in June 2026 by Rafael Dominguez as part of the SkySparc thought-leadership series “The Augmented Treasurer.” Christian Gjerde has more than 20 years of experience in corporate treasury and is currently Head of Treasury and Investor Relations at Elopak. The transcript has been edited for length and clarity; all views are Christian’s own and have been shared with his approval prior to publication.