This article is provided by Alt 21 Limited, trading as Alt21, an FCA authorised payments and FX provider, and is intended for UK businesses considering FX hedging as part of their treasury activity. This article is presented as an interview with Pritesh Ruparel, CEO of Alt21, and reflects his personal views and experience working with Alt21’s clients, rather than independent market commentary or the views of Alt 21 Limited as a firm. All examples provided are illustrative only and not a current or guaranteed rate. Nothing in this article is a personal recommendation or regulated financial advice, and Alt 21 Limited does not provide investment advice. Please read the full disclaimer at the bottom of the page.
AI FX hedging often starts with the same question.
Can AI tell you when to hedge?
Probably not. But that might be the least interesting thing it can do.
Traders have spent decades trying to predict where currency markets will move next. Machine learning has made it possible to process far more market data, but more data doesn’t turn an uncertain outcome into a certain one.
Where AI gets interesting is in what it can help you understand. It can process information quickly, make complex market data more accessible and, increasingly, help you make sense of what’s happening inside your own business.
The future of AI in FX may not be a machine telling you whether sterling will rise or fall tomorrow. It may be giving you better information about your exposure so you can decide what to do about it.
We sat down with Pritesh Ruparel, founder and CEO of Alt21, to discuss the future of FX and where he believes AI will have the greatest impact.
Can AI tell you when to hedge?
Not in the way the question suggests.
There’s an important distinction between machine learning, artificial intelligence and market prediction.
Machine learning has been used extensively in trading for years. Its ability to process huge numbers of data points can help sophisticated investors analyse information that would be difficult for a person to process manually.
That doesn’t make the future predictable.
As Prit puts it:
“People say, ‘Do you think now is a good time to buy or sell FX?’ I always go, well, it’s a coin flip. I don’t know that. What I can tell you is that the data says XYZ.”
Market data can help you understand what’s currently priced in and the probabilities the market is assigning to different outcomes. Those probabilities can change as new information emerges.
What the data can’t give you is certainty about which outcome will actually happen.
So asking AI whether GBP/EUR will rise or fall next month may be less useful than asking a completely different set of questions.
What exchange rate has your business budgeted for?
What currency exposure do you have?
When will you need to exchange the money?
What outcome are you trying to achieve?
That’s where AI FX hedging starts to become much more interesting.
Where is AI genuinely useful in currency management?
If the biggest opportunity isn’t predicting the next exchange rate, where is it?
For Prit, the answer starts with information.
Finance teams already have access to enormous amounts of data about their business and the market. The problem is that understanding it, bringing it together and working out what’s relevant can still require significant time or specialist knowledge.
AI has the potential to change that.
Not by deciding for you, but by making the information surrounding that decision more accessible, straightforward to interrogate and simpler to understand.
1. Making market intelligence easier to use
The idea of using machines to understand financial markets isn’t new.
Quantitative investors have used machine learning for years because computers can consume far more information than a person could reasonably process. They can analyse large numbers of data points and continuously update models as new information becomes available.
Human judgement hasn’t disappeared either.
We must treat any AI-generated explanation as a starting point and carefully verify anything material before relying on it.
Some traders apply what Prit calls a “discretionary overlay”. They absorb the information in front of them, draw on experience and make a judgement that may be difficult to reduce to a neat formula.
AI doesn’t suddenly make either approach obsolete. What it could change is how easily a finance team can interact with the intelligence those systems produce.
You don’t necessarily need to understand the mathematics behind a model to ask a useful question about its output. Increasingly, natural-language interfaces could allow you to interrogate complex information in much the same way you’d ask a person.
That could make sophisticated market information considerably more accessible.
But accessibility shouldn’t be confused with prediction.
The system might help you understand what the market is currently pricing or what probabilities are attached to different outcomes. That gives you more information with which to assess your position. It still doesn’t tell you what will happen next.
2. Bringing market intelligence and your own data together
Market information is only one side of an FX decision. The other side is your business:
- What currencies are you exposed to?
- What payments and receipts are coming up?
- What have you already hedged?
- How certain are those underlying cash flows?
- What exchange rate did you use when setting your budget?
- What outcome is your business trying to achieve?
Prit sees one of the biggest opportunities for AI in bringing these two sets of information closer together.
“It’s giving all that intelligence that people don’t have, which is the intelligence of the market and then the intelligence of their own data.”
At Alt21, this is still an early area of exploration. We’re looking at use cases around analysing customer behaviour and creating data experiences that help customers interact with information about their currency management.
Prit is deliberately cautious about where the technology is today.
But consider where it could go.
You ask:
“Is now a good time to hedge?”
Rather than producing a yes-or-no answer, such a system may have enough context to ask better questions:
- What’s your budget rate?
- What exposure do you actually have?
- How much of it is already hedged?
- When is the underlying payment expected?
- What are you trying to achieve?
It can then help bring the relevant business and market information into the same conversation.
“Is now a good time to hedge?” stops being an attempt to predict the market and becomes a way into understanding your own position.
3. Making FX knowledge more accessible
Prit believes one of AI’s most useful contributions to finance is already happening – It’s lowering the information barrier.
FX has traditionally come with a knowledge gap:
- Products have their own terminology.
- Proposals can be difficult to interpret if you don’t work with them regularly.
- Relatively straightforward market concepts, once explained, can sound considerably more complicated when you encounter them for the first time.
Generative AI gives you another way into that information.
“Knowledge access, that’s the best use case, I think. It’s free. It’s out there. Just ask questions.”
Say you’ve received a hedging proposal and you don’t fully understand one of the terms. You can ask an AI tool to explain it in plain English.
Or perhaps you’re preparing for a meeting with an FX provider.
“You can actually tell it, ‘I’m going into a meeting with counterparty X. I’m looking to hedge this, or I want to improve my currency management. Tell me what I should be asking.’”
You can also challenge the answer you receive. Prit sometimes puts the output from one AI model into another and asks it to critique the response.
It’s not a guarantee of accuracy.
AI can misunderstand context, omit important details or confidently generate information that isn’t accurate.
When you’re dealing with financial decisions, its output shouldn’t be treated as authoritative simply because it sounds convincing.
But it can give you a starting point for understanding a subject that might previously have required hours of research or a conversation with someone who already understood it.
4. Changing the value of FX expertise
Making knowledge more accessible has consequences for an industry that has traditionally placed a high value on possessing that knowledge.
If you don’t understand a product, market convention or piece of terminology, historically you’ve needed somebody else to explain it.
That person holds the information. You depend on them to interpret it.
Prit has never been particularly convinced that access to information alone should command such a premium.
“Breaking down the information barrier is huge, because a lot of value in this industry has historically come from having access to information other people didn’t.”
AI puts pressure on that model because increasingly, the information itself is becoming easier to access.
You can ask what a term means or how a product works. You can prepare questions before a meeting and ask for an explanation of something you’ve been sent.
That doesn’t take away from genuine expertise. If anything, it makes it more obvious where genuine expertise begins.
Learning something new through AI isn’t the same as having years of experience applying that knowledge in the real world.
As basic information becomes less of a barrier, the people sitting between you and that information have to offer something more.
Prit expects that to change the role of intermediaries considerably.
The people who continue to add value will be those bringing genuine expertise, context or judgement that technology can’t simply reproduce.
The rest of the information gap gets harder to defend.
Enjoying this article? You might be interested in what Prit has to say about currency hedging for SMEs.
5. Removing work that never needed to be human
There’s another, less glamorous opportunity for AI.
It can remove work. Not the strategic decisions, but the repetitive work surrounding them.
Finance teams still spend time finding information, pulling data together, checking systems, chasing documents and answering questions that technology could increasingly handle without human intervention.
That could extend to FX.
A finance team shouldn’t necessarily need another person to retrieve basic information, explain where a transaction is or help them understand data that already exists somewhere in a system.
If technology can handle those tasks, people can spend more of their time on work where human judgement actually contributes something.
That doesn’t mean finance teams will disappear. They may become smaller, with highly skilled people able to do considerably more.
The advantage will increasingly belong to people who know how to use AI alongside their own expertise.
6. Creating an FX experience that starts with the question, not the product
Perhaps the most interesting opportunity is what happens when all of these capabilities come together.
Today, financial software is largely built around interfaces.
You log in. You navigate to the relevant section. You find the data. You interpret it. You decide what to do next.
AI creates the possibility of reversing that interaction.
You start with the question:
- What currency exposure do I have next quarter?
- How has my forecast changed since last month?
- What am I already hedged for?
- What does this term in my FX proposal mean?
- What information should I have before I think about hedging this exposure?
The technology then works out which information is relevant and helps you make sense of it.
That’s more interesting than adding an AI chat box to an existing platform.
Prit is sceptical of technology built primarily to demonstrate that a company is using AI.
“This isn’t technically hard. This is just doing stuff.”
The real test is whether it does something useful for the customer.
Over the next few years, that could fundamentally change how finance teams interact with currency management technology.
Instead of learning how to operate another financial system, the system starts learning how to respond to the questions you actually need answered.
And that may prove far more transformative than teaching AI to guess where GBP/EUR goes next.
What should finance teams be wary of with AI FX hedging?
For all the opportunities Prit sees in AI, he’s also sceptical about some of the products appearing around it.
His concern isn’t the technology itself. It’s what happens when adding AI, analytics, or a better interface creates the appearance of innovation without necessarily giving the finance team anything more useful.
Imagine you’re a CFO who’s managed your currency exposure in a spreadsheet for years.
You’re offered a new hedging analysis tool. It presents your exposure through a more sophisticated interface, automates some calculations and gives you information you previously had to work out yourself.
That could save you time. But if the underlying information tells you little more than your spreadsheet did, has the technology actually helped you make a more informed decision?
That’s the test Prit believes new AI and FX technology needs to pass.
A better interface can improve how you interact with information. Automation can remove manual work. But neither automatically makes the information itself more useful.
More data doesn’t automatically mean more insight
AI makes it possible to process and present enormous amounts of information.
That doesn’t mean you need all of it.
A finance team may already know its market rate, hedge rate and outstanding exposure. Adding more calculations, charts, alerts and analysis can create the impression that you’re seeing more without necessarily helping you understand what deserves your attention.
The opportunity Prit sees for AI FX hedging is to remove unnecessary information layers, surface what’s relevant and make existing data more useful.
If an AI-powered product adds complexity without helping you remove work from your day or understand something useful, the technology itself isn’t much of a benefit.
If the software is cheap or free, understand the commercial model
There’s another consideration when you’re giving an AI or analysis tool detailed information about your currency exposure.
What happens to that data once you’ve entered it?
Exposure data can reveal a lot about your business, from the currencies you deal in to when future payments or receipts are expected and where you may have unhedged positions.
That makes it worth understanding the commercial model behind the technology you’re using, particularly when a tool is offered free or at a relatively low cost.
Consider:
- How is your data being used?
- Who has access to it?
- Is it shared with third parties?
- What permissions have you agreed to?
- And how does the company providing the software make money?
None of those questions mean there’s necessarily anything problematic happening behind the scenes. They’re simply part of understanding what you’re exchanging for access to the technology.
If AI FX hedging is going to depend on increasingly detailed information about your business, transparency around how that information is used needs to develop alongside it.
Financial technology still needs financial discipline
This leads to a broader concern Prit has about the current wave of AI development.
Building technology and building financial technology aren’t quite the same thing.
An AI product can be technically impressive and still operate in an environment where data handling, customer outcomes and regulatory boundaries need careful consideration.
On the other side, a financial services business can understand regulation extremely well and still fail to build technology that materially improves the customer experience.
The difficult part is bringing the two disciplines together.
It’s something Prit believes Alt21 has had to learn from the beginning:
Building technology while operating within the requirements that come with regulated financial services.
That can sometimes mean moving more carefully than an unregulated software company. But when technology begins influencing how businesses understand their finances, moving quickly isn’t the only measure of progress.
The question is whether what you’re building is genuinely useful, appropriately controlled and clear about what it is and what it isn’t.
Where should we draw the line between AI and the hedging decision?
The more capable AI becomes, the harder question isn’t what it can do. It’s what role we want it to play in financial decisions.
There’s a meaningful difference between giving you information that helps you understand your currency exposure and telling you what you should do about it.
For Prit, the boundary is clear.
“Our role is the first. Their role is the second.”
For example, Alt21 can give you information about your exposure and the market around it. The decision about whether to hedge, how much to hedge and which approach fits your business remains yours.
AI shouldn’t make that boundary disappear.
Nothing in this article amounts to a personal recommendation or regulated financial advice, and ALT 21 Limited does not provide investment advice.
Better information doesn’t have to become advice
Go back to the question we started with:
“Is now a good time to hedge?”
An AI-powered system could potentially help you explore that question in much more depth than a simple yes or no.
It could show you your exposure. It could surface relevant market information. It could help you understand how your current position compares with your budget rate. It could explain terminology or make complex information easier to interrogate.
What it shouldn’t do is turn that information into an instruction to hedge.
There are regulatory boundaries around financial advice, but Prit’s concern goes beyond regulation.
Once the provider of a financial product becomes involved in deciding whether you should use that product, there’s potential for a conflict of interest.
Prit puts it simply:
“If you’re getting advice from somebody, it needs to be objective, not conflicted by the fact that they make money from it.”
AI doesn’t automatically solve that problem.
In fact, it could make the boundary harder to see if a recommendation appears to come from an algorithm rather than a person.
A system telling you “you have £500,000 of unhedged exposure in the next six months” is providing information.
A system telling you “you should hedge £400,000 of that exposure today using this product” is doing something very different.
The interface may look the same. The role it’s playing isn’t.
What should remain a human decision?
For now, Prit’s answer is simple:
The hedging decision itself.
“The decision needs to be human. If you’re talking about execution timing, that can be automated within parameters.”
That doesn’t mean every action surrounding the decision needs a person behind it.
Imagine your business establishes a systematic hedging programme. You decide which exposures fall within it and agree the parameters the programme should operate within.
Technology can then monitor the market and execute according to those predetermined rules.
As Prit explains:
“They’ve made the decision. We just do the execution.”
The technology isn’t independently deciding that your business needs to hedge or what your objectives should be. It’s carrying out instructions within boundaries you’ve already established.
That may change as the technology matures.
“I think it needs to get a bit more mature before we get there.”
AI models will become more capable. Controls around inaccurate or fabricated outputs should improve, and regulatory frameworks will continue to develop as financial services businesses find new ways to use the technology.
But capability alone shouldn’t determine how quickly responsibility moves from human to machine.
If you’re using AI to explain information and it gets something wrong, you can interrogate the answer, check another source or decide not to rely on it.
The consequences become very different when a system is making financial decisions or taking action autonomously.
For now, human judgement provides an important boundary.
That doesn’t mean roles such as the treasurer are about to become obsolete either.
Large parts of finance, treasury and almost every other business function are likely to become increasingly automated. People may spend less time gathering information, carrying out routine processes or completing administrative work and more time applying judgement where it adds value.
How is AI going to change Alt21?
Alt21 has already tested and built early AI use cases, but Prit is cautious about introducing AI simply because the technology exists.
The starting point is whether it solves a real problem for the customer.
“We’re just being mindful not to be too hyped up and focusing on things where the client goes, ‘Actually, that helps me.’”
That principle shapes how Alt21 is approaching AI.
There are areas where the opportunity is relatively straightforward. AI can help analyse information, understand customer behaviour and create better ways for finance teams to interact with their data.
Other applications move much closer to the boundary between information and advice. That requires a different level of consideration.
Prit has long taken the view that financial services shouldn’t hold back useful applications of AI simply because the technology is developing quickly, while being clear that greater caution is needed once AI begins influencing decisions where an inaccurate output could have financial consequences.
That balance still informs how he thinks about AI at Alt21 today.
The goal isn’t to attach AI to every part of the platform.
It’s to identify where the technology can remove work, make information easier to understand or create a better way for customers to interact with their currency data, while staying clear about where appropriate boundaries need to sit.
That may mean moving more deliberately in some areas than a software company without the same regulatory responsibilities.
For Prit, that’s part of building financial technology properly.
The interesting question isn’t “Where can we add AI?”
It’s “Where does AI make this genuinely more useful for the customer?”
What does the future of AI FX hedging actually look like?
Probably less dramatic than the headlines suggest.
And potentially much more useful.
Prit expects AI to change the way finance teams manage currencies significantly over the next five years. Not because every decision will suddenly be handed to a machine, but because much of the work surrounding those decisions could look very different.
FX knowledge becomes readily available
The first change may already be happening.
Prit expects the information barrier around FX to continue falling.
Concepts that once required specialist knowledge or lengthy research can increasingly be explored by asking questions in natural language. Over time, the same principle could extend much further into the information held within your own business.
Instead of knowing which report to open, which spreadsheet to check or which person to ask, you could increasingly start with the question you need answered.
That doesn’t make expertise irrelevant.
It means access to basic knowledge is less likely to be the thing that makes someone an expert.
Smaller finance teams could do more
There’s plenty of work in finance that doesn’t require financial judgement at all.
Prit gives some very ordinary examples:
“They spend less time doing stuff like chasing a provider for a statement or asking where a payment is.”
Pulling information from different systems, manually assembling data or finding the status of a transaction all take time before you’ve even started doing anything useful with that information.
Prit expects AI and automation to take much more of that work away.
Information could increasingly be available in real time, allowing finance teams to move from finding the data to deciding what to do with it.
Taking repetitive work out of finance could also change the shape of the team itself.
Prit expects some finance teams to become smaller, with highly skilled people able to do considerably more because AI handles more of the work around them.
The advantage will increasingly belong to people who know how to use AI alongside their own expertise.
That doesn’t mean removing people from finance.
“You need a team. You’re going to need some human discretion.”
Less time goes into work that exists because systems can’t talk to one another, information is difficult to retrieve or a process still needs somebody to move it along manually.
More time can go into the work where experience, commercial context and judgement actually contribute something.
Basic intermediation becomes harder to justify
FX has traditionally put people in the middle of many relatively simple interactions.
You need information, so you contact someone.
You want to understand an exposure, so you contact someone.
You need to transact, so you contact someone.
Prit sees parallels with door-to-door sales in the 1970s – a model that worked because information was hard to access, not because a person needed to be in the room.
If technology can retrieve information, explain straightforward concepts and handle routine processes more efficiently, having a person in the middle simply because that’s how the industry has always operated becomes harder to justify.
The human interaction that remains could become more valuable precisely because there’s less of it.
You speak to an expert when you need expertise, not because you need somebody else to operate the process for you.
The biggest change may be what you no longer have to think about
AI doesn’t need to predict tomorrow’s exchange rate to fundamentally change how businesses manage FX.
Five years from now, the biggest change may not be a spectacular new AI capability. It may be all the things finance teams currently spend time doing that simply no longer require their attention.
As Prit puts it:
“If you go forward five years, what finance teams – or any team – spend their time doing is going to look massively different.”
That’s ultimately a much more practical vision for AI FX hedging than building a machine that claims to know where the market goes next.
Be ready for what’s next in FX
The way finance teams manage currency is changing. AI will be part of that shift, but the goal remains the same:
Get better access to information, remove unnecessary work and better understand your currency position.
Alt21 brings international payments and FX hedging together in one currency management platform, with technology designed to evolve alongside the way finance teams work.
Open an Alt21 account today and be ready for what’s next in currency management. Applicants must pass Alt21’s onboarding process and accept our terms and conditions before becoming a client.
ALT21 Limited is authorised and regulated by the Financial Conduct Authority (FRN: 783837) and is a company registered in England and Wales (number 10723112). The registered address is 45 Eagle Street, London WC1R 4FS, United Kingdom. This article has been produced by ALT21 Limited for information purposes only. It does not constitute financial advice or an offer to sell or the solicitation of an offer to buy any products referenced. Hedging products are not suitable for every business. Before entering into any FX product, you should consider whether it is appropriate for your needs and circumstances. ALT21 Limited assumes no liability for errors, inaccuracies or omissions. Eligibility criteria and terms and conditions apply to all products and services offered by ALT21 Limited. Not all applications will be accepted.
