Here’s a number worth sitting with: 69% of financial services firms report that AI is improving productivity in front-office and client-facing roles. Yet only 40% report increased profitability, and 55% say measuring AI’s value remains challenging. Those findings come from the Cambridge Centre for Alternative Finance at Cambridge Judge Business School, and I think they’re worth sitting with for a second.
Because that’s a strange combination. Many firms perceive better front-office productivity. Far fewer report a profit increase, and more than half still find the value difficult to measure. So which is it? Is AI genuinely saving advisors time, or does it just feel that way because the work is new and a little exciting?
I don’t think the honest answer is one or the other. I think it depends entirely on the task, and most of us have never actually stopped to check. That’s really the question behind AI productivity for financial advisors right now: not whether the models work, but whether you can prove they’re working for you, specifically, on your desk. So here’s what I want to offer: not another argument for or against AI, but five questions you can run against anything you’re using it for, before you decide it’s earned a permanent place in your week.
Feeling faster isn’t the same as being faster
There’s a reason 55% of firms in that study say AI’s value is hard to measure. Time savings are sneaky. A task that used to take forty minutes might take fifteen minutes to draft with AI, but if you then spend another twenty minutes double-checking the output, the net gain is only five minutes. That win is much smaller than it first appears.
I get why that’s easy to miss. Anything new can feel efficient at first. The first few times you use a new AI model, the excitement of trying something different can look a lot like real productivity, even when it isn’t yet. The only way to know the difference is to ask a few plain questions instead of trusting the feeling.
Feeling faster and being faster are not the same thing.
Five questions to ask before you trust the time savings
I’d start with whether the task even repeats. A one-off research question you’ll never ask again isn’t where AI earns its keep. It’s the Tuesday and Thursday tasks, the meeting notes, the first drafts of client emails, the same three spreadsheet cleanups you do every month, where repetition lets a small saving compound into something real.
Then I’d ask whether the result is usable, or close to it. If you’re rewriting most of what the model gives you, you haven’t saved time. You’ve just moved the work from a blank page to an edit, which is a different job, not necessarily a shorter one.
That leads straight into the third question: how much review does this take? Be honest here. If a task still needs a careful read, a fact-check, and a tone pass before it goes anywhere near a client, count that time. It’s real. It belongs in the ledger.
Fourth, is this the kind of work AI is appropriate for? Summarizing a call, drafting a first pass at a follow up email, organizing scattered notes into something coherent. Those play to what these models are good at. Interpreting a client’s specific tax situation or making a judgment call about their risk tolerance does not, and no amount of clever prompting changes that. Your judgment stays the part that can’t be automated and shouldn’t be.
And fifth, has the improvement lasted? Not for a day. Not for the first excited week when everything feels new. For several weeks, once the novelty wears off and the task is just another Tuesday again.
The real test isn’t day one. It’s week six.
What AI productivity for financial advisors looks like
Picture two advisors, same firm, same AI models. One uses AI mainly for meeting notes and follow up drafts, things she does five times a week in more or less the same shape. Within a month she’s not thinking about it much anymore. It’s just faster, quietly, in the background.
The other is reaching for AI on something different every time. A one-off market question here, a complicated planning scenario there, nothing repeats long enough to build a rhythm. He’s not wrong to experiment. But if you asked him in month three whether AI had saved him real time, he’d probably pause before answering. That pause is the answer.
Neither advisor is doing anything wrong. They’re just discovering, one task at a time, where the actual value lives for them. That’s the work. Not adopting AI in general but figuring out specifically where it holds up under your own five questions.
The point isn’t to prove AI wrong
I don’t read that 40% profitability number as bad news, and don’t think you should either. I read it as evidence of a gap between perceived productivity and measurable financial return. The report suggests that firms are still working out how to translate AI use into enterprise value. That’s not a knock on AI. It’s an invitation to get more precise about where it earns its place.
It’s not humans versus AI. It’s humans with AI versus humans without it, and the advisors who come out ahead won’t be the ones who used the most AI models. They’ll be the ones who figured out, task by task, where it genuinely helped and where it just added motion.
If you’re trying to build that kind of habit, one task at a time, the Optimal AdvisorAI Learning Center is built for exactly this. Not a promise of automatic time savings, just a structured way to get sharper about where AI really helps in your day and where your own judgment still does the heavier lifting.
References: Cambridge Centre for Alternative Finance. The 2026 Global AI in Financial Services Report: Adoption, impact and risks. University of Cambridge Judge Business School, April 2026.
