When AI Frees Up Capacity, Where Should It Go?

AI has the potential to free up capacity across every function in financial markets, but is the aspiration to do more with considerably fewer people really the best outcome to aspire to? Maybe a better, but harder question to answer is what that freed-up capacity should be used for next. Drawing on an interesting case from outside the industry, Martina Doherty discusses why the real test of an AI rollout isn’t whether the technology works, but what leadership chooses to do with the people it frees up.

A large majority of conversations about AI deployment in financial markets start with the same question: where can we achieve more efficiencies?  Client onboarding, trade confirmations, KYC checks, building presentation decks…the list for potential automation keeps growing, supported by a business case for each – faster processing, lower costs, more efficiencies.

But a question that few are asking, and one that could determine whether AI will make an organisation successful in this AI revolution, is this: once the technology is doing its job, where does the freed-up human capacity go?

I spoke at a capital markets event recently about the human edge in an AI world, talking about the collective intelligence that emerges when humans and machines work together better or in completely different ways to what each can do individually. Since then, I have come across a recent case from outside markets that makes the point really well.

Reuters recently reported how IKEA trained a chatbot to handle close to half its customer service queries, and it did so brilliantly, with wait times dropping, customer satisfaction rising, and the company saving more than EUR 13 million. On paper, this meant roughly 8,500 people were surplus to requirement and, at this point, most companies would count the freed-up headcount as operational savings and move on. IKEA didn’t.

And even though its business model is a world away from the FX markets, the impact of what the company did is worth noting. 

The insights were in what the technology couldn’t resolve

The chatbot handled routine queries well. What it consistently couldn’t resolve was where real market intelligence lay: customers weren’t just asking about their order status, they were asking for help designing their living spaces – a service IKEA didn’t yet offer at scale. That unresolved demand, sitting inside the chatbot’s “failure” data, was a market signal – one that a human had to actually notice and act on more intelligently than simply logging it as a “bot gap” to patch.

The company invested in the people – not for emotional or altruistic reasons, but for strategic ones – long-term strategic reasons that built a completely new revenue stream

Human judgement, interrogation skills, curiosity and communication – all those uniquely human skills came into play here, as well as the ability to ask a different question: the issue that the tech or AI can’t resolve, is it a bug, a valuable piece of information or a gap in our product/service offering that hasn’t been named yet?

This is what the human edge looks like in an AI world – adding value in a way that only a human can.

What IKEA built with the freed-up capacity

Equally interesting is that the leadership team at IKEA didn’t hire a team of external designers to capitalise on the new opportunity. Instead, they retrained the 8,500 people who already knew their products, their customers and their pain points, and turned them into a premium interior design consultancy – one that generated more than EUR 1.3 billion in 2025, accounting for 3.3% of the firm’s revenue. That is projected to grow to 10% of total revenue by 2028.

The human reasoning and judgment demonstrated by IKEA’s leadership team here is noteworthy. They didn’t treat their customer service agents as a cost to offload once the technology delivered, instead they saw the value of their accumulated knowledge of the IKEA customer base and their products as an asset worth building on. So, they invested in them – not for emotional or altruistic reasons, but for strategic ones – long-term strategic reasons that built a completely new revenue stream.

The underlying principle here is one that the FX industry understands – even if it doesn’t always appear to apply. In an industry that is becoming more tech-driven by the day, what continues to differentiate one FX franchise from another is rarely the technology. It’s relationship depth, market knowledge and the experience of their people that allow them to offer a genuinely personal service and innovative products.  That can never be replaced by technology, regardless of how revolutionary it appears to be – and those human elements need to be an integral part of discussions about AI deployment and spend.

The real question that needs to be answered

Every FX business currently deploying AI into their trading, onboarding processes, client communications, or trade support will eventually hit the same fork IKEA did. Some volume of work will disappear and many jobs as they exist today will no longer be viable. The natural instinct is to see these as operational savings, cut the headcount and move on.

But for every leader, here’s a different question to contemplate: is that freed-up capacity really a cost to eliminate, or an asset to redeploy? The answer could determine what your business looks like in three years’ time. 

IKEA didn’t just build a revolutionary chatbot with its AI capabilities. It responded intelligently to a question that hadn’t been obvious before the data became available. In FX and markets, where relationship depth and markets experience are still impossible to automate and easy to under-invest in, what possible questions are worth asking across your teams and organisation before the next AI investment or round of redundancies?

An organisation’s competitive edge in an AI world will never be based on its technology. It will be created by people who dare to think differently and by leaders who question where the skills and the expertise of those freed-up people are best deployed.