Operating model redesign will come slowly in regulated firms.
That does not make it minor. In regulated organisations, redesign is more likely to arrive through a series of small workflow decisions than through one dramatic before-and-after transformation.
A lot of AI discussion still assumes change will arrive as a formal programme: a new platform, a major redesign, a clear moment where the old model gives way to the new one. In regulated firms, it usually does not happen like that.
It comes in pieces. Through procurement cycles, software upgrades, policy changes, local process reviews, staff turnover, and small decisions about what can be delegated, what can be supported, and what still needs a person to judge, check, or own.
Because that pace is slower, it can create false reassurance. Redesign does not look dramatic, so it is easy to treat it as marginal. But small changes have a habit of accumulating. Over time, they amount to a different operating model whether anyone formally announced one or not.
That is the more realistic path now.
The question will keep coming back
As tools improve, organisations will keep asking the same practical question: where can AI assist here, and how can it do so safely?
That is a sensible question. In many cases, the answer will become clearer over time. Some tasks will be easier to draft, summarise, triage, or structure. Some forms of administrative work will become more machine-assistable. Some workflows will shift from human production at every step towards human review, validation, and decision at the points that matter.
So yes, operating models are likely to change. That does not automatically mean work gets better.
Time saved is rarely left empty
One of the weaker assumptions in AI discussion is that time saved becomes space. In practice, it often becomes capacity for more work.
More throughput. More parallel tasks. More oversight. More expectation that people can handle several streams of activity at once. In some settings, managers may end up supervising not just people, but systems, outputs, and machine-supported processes as well. The role becomes less about doing everything directly and more about directing, checking, escalating, and carrying accountability.
That may improve productivity. It may also increase work intensity. Those are not the same thing.
In regulated firms, that distinction matters, because the question is not only whether work can be done faster. It is whether the redesigned model remains governable, reviewable, and workable for the people inside it.
Productivity and work intensity are not interchangeable.
Better redesign needs restraint as well as enthusiasm
Over the next few years, the harder task will not be spotting where AI can help. Organisations will get better at that. The harder task will be deciding what should happen to the capacity created, what new risks are being introduced, and what part of the human role still needs protecting.
If every efficiency gain is immediately converted into more load, organisations may end up with a model that looks more productive on paper but is less sustainable in practice.
That is why operating model redesign in regulated environments needs to be deliberate. Not because change can be avoided. It probably cannot. But because a model that is more efficient, harder to govern, and harsher to work inside is not obviously progress.
The redesign is coming slowly. That is exactly why it needs attention now.