Most of the debate in this industry is still about which tool… Which model, which vendor, which feature list, which integration. Those are real questions, but they are not the ones that decide whether a rollout works. What decides it is what happens on the Monday after go-live, in a business where every hour was already spoken for and nobody’s job description changed.
The friction that stalls AI adoption at advisory firms is mostly human and organizational. It shows up as buy-in that never arrived, workflows nobody redesigned, and training that taught the interface instead of the moment. None of it is anyone’s fault, and all of it is predictable.
The adoption gap is measurable, and it predates AI
The 2026 Connected Wealth Report from Advisor360, based on 300 advisors at firms averaging $548M in client assets, found that 74% say they are not getting full value from the technology they already have, and 72% describe their firm’s technology as outdated or in need of improvement.
Research from Datos Insights, sponsored by Practifi and covering 436 advisors and 71 executives, points at the same thing from another angle. 40% of advisors report struggling with technology adoption, and 48% of RIA advisors name CRM workflow enhancements as a priority. Underneath both figures sits a utilization gap: advisors rate their CRM highly on value while acknowledging they do not use it to anything close to capacity.
That gap existed before generative AI arrived, which is the useful part. A firm that has already bought good software and captured a fraction of its value has an adoption problem, not a procurement problem, and adding AI on top of it will not resolve itself. It will simply produce a second underused system faster.
Three frictions that derail rollouts, whichever tool a firm picks
Buy-in is a question about judgment, not about features
A new tool announced as a way to make advisors faster is heard as a statement about how they currently work. That is not defensiveness. It is a reasonable reading of an ambiguous message, and it is the reading firms most often fail to anticipate.
The rollouts that get past this tend to have answered a narrow question in advance: what does this take off someone’s plate, and who decided that thing was worth taking off. When the answer is specific, the tool reads as help. When the answer is a general appeal to efficiency, the tool reads as scrutiny, and the people whose cooperation matters most are the ones with the most experience to feel scrutinized about.
There is a second-order version of this for firm leadership. AI that summarizes an advisor’s work also creates a record of that work. Nobody says this out loud in a launch meeting, and quite a lot of people are thinking it. Naming it early costs a firm very little and buys back a surprising amount of goodwill.
Workflow redesign rarely has an owner
Software arrives, yet nothing leaves. The tool is added to a week that was already full, and adoption becomes an addition rather than a substitution.
Whoever runs the rollout usually owns the software including the licenses, the configuration, and the training calendar. The week the software lands in belongs to nobody. No one is accountable for deciding which existing step this replaces, which meeting it shortens, or which document it makes redundant, so all those things carry on happening alongside it.
The effect compounds when the tool sits somewhere other than where the work already happens. A separate destination is a separate decision to go there, made dozens of times a week by people who are between client meetings. Firms rarely lose that battle dramatically. They lose it by attrition, over about six weeks.
Training teaches the interface and skips the moment
Most training covers what the features do. The question advisors actually have is narrower and harder: “At what point in an ordinary Tuesday does this happen, and what am I doing instead of it now?”
Training is also badly timed almost everywhere. It runs at go-live, when nobody has a live case in front of them and nothing is at stake, and it does not run again three weeks later when the first real question appears. The knowledge decays before it is ever needed, and the person who forgets how to do something once tends not to try a second time.
Why the usual remedies underperform
Firms reach for three fixes, and each has a known failure mode.
Mandates produce compliance rather than use. A required tool gets opened, and the output gets ignored, which is the outcome that looks like adoption in a usage report and is worth nothing.
Champion programs select for the people least representative of the firm. A pilot staffed by volunteers measures enthusiasm, not fit, because enthusiasts absorb friction that everyone else declines to absorb. The real number appears at general rollout, which is also when it is most expensive to discover.
More training treats a design problem as a knowledge problem. If a capable person has been shown something twice and still does not use it, the issue is usually that using it costs more than it returns at the moment they would have to decide.
None of these are bad instincts. They are just remedies aimed at the people rather than at the thing the people are being asked to do.
The question worth asking before the next purchase
Firms tend to ask how they will drive adoption. The more useful question is how much adoption a given tool requires in the first place.
Any capability that sits behind a door someone has to remember to open will be used by the people who remember, in the weeks when they have time to remember, which is not the same as being used by the firm. Every additional step between a person and the value is a place where a rollout quietly loses another few percent.
That reframes the test of a rollout. The measure is not whether advisors form a new habit. It is whether they notice, when they arrive, that some of the work is already done.
Four questions make that checkable before anything is bought:
- Does anything arrive without being asked for, or does every piece of value require someone to initiate it?
- Does it read the firm’s own records, or only what a person remembers to paste into it?
- Does it remove a step, or add a destination? A tool that lives inside the system people already open is in a different category from one that lives beside it.
- Can a person check the output? Work that cannot be verified will not be relied on, whatever the training said, and rightly so.
A firm that can answer those four honestly about a product will know most of what it needs to know about how the rollout will go.
Where this leaves the change-management conversation
We think the change-management problem in this industry is real, widely underestimated, and mostly a design problem wearing a people problem’s clothes. Buy-in, redesign and training all matter, and firms should invest in them. But the amount of change management a rollout needs is largely set at the point of purchase, by how much the technology asks of the people using it.
Over the next few weeks we will look at how the systems firms actually choose between handle that question, and what to look for when comparing them. The place to start is not with the feature list, but with how much of the work only happens when somebody remembers to make it happen.
Interested in seeing how Sentir’s agent workforce reads your firm’s relationships?
Sentir is Practifi’s AI-native Intelligent CRM for wealth management, and the first of a new category where intelligence is part of the architecture, not bolted on. Built on Practifi’s enterprise-grade foundations, Sentir places a workforce of specialized AI agents inside the system of record, where they continuously interpret client context, prepare work, surface next steps, and help teams act with confidence. The result is a CRM that works before advisors log in, enabling every interaction to start fully informed and allowing every team to spend less time feeding the system and more time applying expertise, strengthening relationships, and staying ahead of client needs, all within the compliance, security, and governance controls the business already trusts.
