If AI can improve every forecast in your business, why is FX forecasting still largely judgemental?

“If AI can improve every forecast in your business, why is FX forecasting still largely judgemental?”

Every CFO I speak to is investing in AI.

a> Sales forecasting.
b> Demand planning.
c> Inventory optimisation.
d> Working capital.
e> Cash forecasting.

Yet one area is still surprisingly reliant on instinct, opinion and whatever the bank happened to say last week.

Foreign exchange!

The irony is that FX isn’t usually a forecasting problem.

It’s a decision problem.

Nobody consistently knows where EUR/USD or GBP/USD will trade six months from now. If they did, they wouldn’t be running a treasury department.

But that’s often how companies behave.

- Someone has a view.
 -The board becomes optimistic or pessimistic.
 - The market moves.
 - A hedge is delayed, accelerated or abandoned.

The result isn’t better forecasting.

It’s simply emotion dressed up as strategy.

I think AI can genuinely transform corporate treasury. Not because it will magically predict exchange rates. It can dramatically improve the quality and consistency of the decisions surrounding them.

Imagine an AI system that continuously asks questions such as:

• How exposed are we today compared with last month?
 • What would a 10% currency move do to EBITDA?
 • Are we operating within our board-approved policy?
 • Have our budget assumptions materially changed?
 • Are we over- or under-hedged relative to our risk appetite?
 • Which decisions are genuinely new, and which are simply emotional reactions to market moves?

Notice what’s missing.

“Where will EUR/USD be in December?”

That’s often the least useful question.

The companies that manage currency risk best aren’t necessarily better forecasters.

They’re better decision-makers.
They have clearer frameworks.
More consistent processes.
Better governance.
Less emotion.

I suspect AI’s biggest contribution to treasury won’t be replacing judgement. It will be removing the noise that too often masquerades as judgement. Perhaps the future of FX isn’t an AI that predicts the market. It’s an AI that helps finance teams make better decisions regardless of what the market does. That’s a much more achievable and valuable goal.

What do you think?

Will AI eventually become a decision-support tool for treasury, or do you believe currency management will always remain primarily a human judgement?