The clients most worried about deepfake scams want the same phone call as everyone else

Fraud or a data breach has already reached 45% of the 500 ultra high net worth respondents in WealthLens™ 2026, against 18% of the 900 high net worth respondents: two and a half times the rate. Neither group wants to meet its bank any more often than the other.

A large share of affluent and high net worth respondents want to run their private banking relationship by voice: 42% name the phone call among their preferred channels, alongside the website, the mobile app, the branch, and text message.

The same respondents put AI-generated deepfake scams, aimed at them personally, level at the top of the digital risk list with data breaches at their own institution, both at 41%.

No other risk stands apart from those two.

Five of the seven risks sit within four percentage points of each other, between 41% and 37%, and only digital reputation management and lack of human judgment score lower.

When so many concerns are rated almost equally, a bank has no clear order of priority to build a protection program around.

Concern sits at much the same level whichever risk is named.

Actual experience of fraud is not level at all: it rises sharply with the size of the balance.

The largest balances carry the most reported fraud

Among the 500 ultra high net worth respondents, 45% say fraud or a data breach has already reached their financial accounts, against 18% of the 900 high net worth respondents.

The gap holds in all four markets, widest in China and narrowest in Australia.

These are accounts of what respondents say has happened to them, and part of the difference may be in who notices and who gets told.

For a coverage team the effect is the same either way: the highest-value clients are also the ones reporting the most incidents.

The same group is not the one most worried about the institution.

Data breaches at their financial institution were named by 38% of ultra high net worth respondents, against 43% of the high net worth group.

Half want one meeting, then a screen

Offered a choice of how to run the relationship, 50% would meet a consultant once to build it and use digital channels after that.

Another 23% would not meet a consultant in person at all.

Meeting face to face every time a question comes up is the standing preference of 18%.

Naming deepfakes as a risk does not change the channel

The respondents who name deepfakes ask for the same arrangement as everyone else.

Among the 571 respondents who put AI-generated deepfake scams in their top three, 18% want to meet face to face every time, against 19% of the 829 who did not name them.

Both groups name the phone call for the relationship, at 42% each.

Experience makes no difference either.

Among the 389 respondents who say they have experienced fraud or a breach, 18% want to meet face to face every time, and so do 18% of the 1,011 who have not.

Whatever a client has been through, and whatever they say they fear, the arrangement they ask for is the same one.

Put the check where the client already is

A bank that hardens verification this year will be doing it on its own authority, and will be judged on how well the new check works in practice, not on whether anyone asked for it.

That leaves two decisions. The first is where the check goes.

The phone call and the mobile app are the channels clients have already chosen, and authentication built into them asks nothing of a client that the client has not already accepted. Routing an exception into an in-person meeting sends the client to the one channel they are least likely to want: only 18% ask to meet face to face every time.

The second is the order. Sequencing by reported exposure puts the 500 ultra high net worth respondents first.

They report fraud or a breach at two and a half times the rate of the 900 high net worth respondents, so a new verification standard will meet the most incidents in that group, and a clumsy one will be noticed there fastest. What the survey cannot show in advance is how clients react the first time the check blocks a transfer that was genuine. That is the test worth building into the pilot.

Respondents were also asked what they would do if their relationship manager moved to another institution: 66% say they would move too.

We would read that as a statement about where authority sits in these relationships. The credential a client recognizes is a person, and that person is portable. That also makes the relationship manager the obvious person for a fraudster to imitate. The survey did not ask what a client would do if someone posing as their relationship manager called with instructions, so these figures do not measure how clients handle impersonation.

The deepfake figures measure concern.

The experience question does not record whether a deepfake was involved in any of the cases respondents reported.

The phone call is where these relationships are run, by the clients’ own choice.

It is where the verification standard has to work.

Source note

Source: Agility Research & Strategy, WealthLens 2026. Online survey of 1,400 affluent and high net worth respondents, fielded 29 May to 9 July 2026: Australia n=300, Singapore n=300, China n=500, Hong Kong n=300.