Beyond Hours Saved: How Firms Should Actually Measure AI ROI

If you’ve evaluated an AI tool in the last two years, you’ve seen the pitch. A PowerPoint displays the oh so compelling numbers of hours saved per advisor, per week, per year.

They’re easy numbers to produce and, as a result, easy numbers to compare across vendors.

They are also the wrong numbers to stop at. Because while saving time is beneficial, it’s not the end goal.

The true metric is not just the input of time saved, but also the output of what that tool enables your firm to achieve. And in wealth management, the truest test of a tool’s output is its effect on the client experience.

The client relationship is the growth engine

Ask most firms what drives new business and you’ll hear a mix of answers. Everything from marketing to on the ground prospecting to maybe a niche specialty.

The research, however, tells a much simpler story.

Referrals are, by a wide margin, the dominant source of new clients in wealth management. In a study by Cerulli Associates, it was found that referrals account for 74% of new client acquisition among RIAs. Meanwhile, Charles Schwab’s 2023 RIA Benchmarking Study found a similar pattern. Referrals represented 67% of new clients and new client assets that year.

The same Cerulli research found advisors spend only about 7% of their time on business development, amounting to roughly three hours in a 40-hour week. That’s not because advisors don’t understand referrals matter. It’s because there’s no time left over to work the relationship side of the job once service, prep, and administrative work take their share of the week.

Which is the exact bottleneck efficiency gains are supposed to relieve. And also why the productivity question matters as much as efficiency.

After all, if referrals drive roughly three-quarters of new business, the client relationship isn’t one growth channel among several. It’s the growth channel. Which means an AI ROI framework that only measures time saved is measuring everything except the thing that actually grows the business.

What firms should actually track

Instead of stopping at hours saved, firms should track three layers.

1. Capacity created

This is the old metric of time saved now correctly relabeled as an input rather than a result. It’s still worth tracking of course, just not worth stopping at.

Useful metrics here may include:

• Hours saved per advisor per week.
• The percentage of routine tasks now automated.
• Time-to-prep for a client meeting before and after implementation.

2. Capacity reinvested

Focusing on capacity is the difference between “more of the same” and “more from the same time.” Whether the hours a tool frees up get poured into a fuller calendar or into deeper conversations with the clients already on it, this metric helps track the potential of that time.

Useful metrics include: 

• Meeting cadence versus meeting depth (are advisors having more meetings, or longer and more substantive ones?).
• Households per advisor tracked against headcount, proactive outreach — advisor-initiated touches per client, not client-requested ones — and the number of planning topics that make it into a meeting instead of getting pushed to “next time.”
• How much of the newly freed time actually gets earmarked for relationship-building and referral generation, versus absorbed into a fuller calendar of the same kind of work.

3. Outcomes changed

If capacity created tells you what got freed up and capacity reinvested tells you where it went, outcomes changed tells you whether any of it mattered. In other words, this is the actual ROI that shows up in your business, not just the calendar.

Useful metrics include:

• Client attrition/retention year over year.
• Referral rate, including referrals generated per advisor, or the share of new clients and new assets sourced through referral, benchmarked against an industry baseline.
• Share-of-wallet growth as clients consolidate outside accounts.
• Time-to-detect a life event, being the gap between something happening in a client’s life and the advisor knowing about it.
• How many investment-only clients deepen into full planning relationships.
• Assets under management per advisor without a corresponding increase in headcount.

Consider a scenario where two firms report the identical number of hours saved and land in entirely different places.

One books more meetings with the same depth as before. The other keeps the same number of meetings but uses the extra time to catch a life event early, have the estate-planning conversation that used to get deferred, or simply call a client who wasn’t expecting to hear from them.

Both firms will show the same efficiency number. Only one of them is doing anything that a referral or new business is likely to come from.

What to consider before buying or renewing an AI tool

When the majority of a firm’s new business comes from the strength of its existing relationships rather than its top-of-funnel marketing, the real test of any AI investment isn’t how many hours it saves. It’s whether it makes the relationships that already do most of the growing even stronger.

Therefore, your analysis should follow a simple framework:

• Instead of asking how many hours a tool will save, ask what advisors are expected to do with the hours it saves or if the tool helps increase the productivity of those hours.
• Instead of asking how much faster prep gets, ask whether that faster prep changes what actually happens in the meeting.
• Instead of asking for an efficiency benchmark, ask for a retention or referral benchmark six months after rollout.

None of this makes efficiency irrelevant. But efficiency alone is never the goal. The firms that get the most out of AI won’t be the ones with the biggest hours-saved number on a slide. They’ll be the ones who took that time and gave it back to the relationships that are already doing most of the work of growing their business.