By Dimitri Limata, COO at MoiraCorp. Three sides of the same problem, and the pattern that turns up in all of them.
Most financial services firms do not have an AI problem. They have an operational layer that was never built to scale.
I have seen it from three sides.
Six years inside investment banking
At Crédit Agricole, working on FX derivatives. Complex systems, sophisticated talent, real money on the table — and an operational layer underneath that created as much friction as it removed. The trading floor moved at speed. The plumbing did not.
Running a business end to end
Owning and operating a company with £3M in revenue. Every hour lost to a broken process came directly off the bottom line. You understand operational cost differently when it is your name on the lease — when a delay in invoicing is not a process metric, it is your cashflow. That experience reframes how you look at “efficiency”.
Two decades advising regulated firms
Across financial services, asset management, and other regulated industries, helping companies structure and improve their operations. Different firms, different sectors, different regulatory pressures.
Three completely different environments. The same pattern in all of them.
The pattern
Capable people losing time every week to workflows that were never designed for the business they became. The team grew. The product evolved. The regulation got heavier. But the underlying operational layer — the way work actually flows from one person to the next, the way information moves, the way decisions get logged — kept the shape it had when the company was half the size.
That shape is where most of the time and cost is now sitting. And most firms cannot see it, because everyone is busy doing their job inside the friction rather than looking at it from the outside.
Why most firms automate the wrong things first
This is where the AI story usually goes wrong. A vendor turns up with a tool that solves something. The firm buys the tool. The tool gets deployed onto the same operational layer that was already creating the problem. It speeds up one step. It does not change the system. Six months later, the bottleneck has moved, and the same conversation starts again with a different vendor.
Firms select tools before they understand where the time is actually going. That is the mistake. Not the AI choice — the sequence.
What we built MoiraCorp to do
MoiraCorp starts with understanding how the work actually flows, and where capacity is being lost, before anything new is introduced. It is unglamorous work. It is also the only work that makes the next automation decision pay back.
Once that layer is visible, the right automation choices become obvious. Some of them are AI. Many of them are not. Some are organisational, some are technical, some are simply about removing a duplicate step that everyone has accepted as inevitable. The order matters more than the technology.
The capacity problem in plain language
The financial services firms I speak to are not short of demand. They are short of the operational infrastructure to handle it. Clients want to engage. Advisers want to advise. Compliance officers want to do compliance properly rather than fight a queue. The work is there. The friction is what is in the way.
If you fix the operational layer, AI starts paying back almost immediately because it has somewhere to land. If you do not, no AI tool will save you. It will just give you faster ways to do the wrong things.
