The Intermediary – October 2026 - Flipbook - Page 82
T E C H N O L O GY
Opinion
The battle
for trust
T
here is a danger when
we talk about artificial
intelligence (AI) in
financial services
that we assume the
revolution has already
happened. The language certainly
suggests it has, with agentic AI,
autonomous systems and intelligent
decisioning rapidly becoming part
of the industry’s vocabulary, yet the
consumer reality remains rather
more prosaic.
Research published alongside
the Mills Review suggests that only
16% of consumers currently use
AI for a personal finance task and,
among those who do, its principal
role is still to summarise, explain or
simplify information.
We are, in other words, largely
using AI to help us understand
financial services rather than asking it
to make financial decisions for us – a
distinction that maers, although I
suspect it may not last very long.
Consumers are already becoming
accustomed to using AI elsewhere
in their lives. It seems unlikely
that financial services will remain
an exception. Over the next 12 to
24 months, we should expect the
conversation to move quickly from
whether consumers will use AI in
managing their finances towards what
they are prepared to allow it to do.
There is an important question
within that transition which goes
beyond technology, as the Mills
Review research also suggests that
willingness to use AI-enabled financial
tools tends to increase with income.
That creates an interesting dilemma
for an industry already wrestling with
an advice gap.
AI should, theoretically, help to
close it. The economics are compelling
because technology can potentially
make information, guidance and
more personalised support available to
people for whom traditional financial
advice has been inaccessible.
80
The Intermediary | October 2026
But technology does not necessarily
distribute itself evenly. If more
affluent consumers are the first to
adopt increasingly sophisticated AI
tools, then AI could initially reinforce
an existing information advantage.
Those who are already beer served
could gain access to beer analysis,
more personalised information and
increasingly capable digital assistance
while those who might benefit most
remain outside the system. What if
the advice gap becomes an AI gap?
For lenders, however, there is
another dimension to this which I
think is potentially more significant,
and that is trust. Consumers may
increasingly become comfortable
asking AI about money, but there is
a considerable difference between
asking a system to explain a mortgage
and trusting it to choose one. Accuracy
maers, but so do accountability,
data security and the knowledge that
somebody ultimately stands behind
the decision.
Lenders have spent decades building
the infrastructure of trust, sometimes
voluntarily and sometimes because
regulation has required them to.
They operate within established
rules, hold enormous quantities
of customer data and provide
routes to redress when things go
wrong. Historically some of that
infrastructure has been regarded as
friction, particularly when compared
with the ‘seamless’ experience offered
by technology companies.
In an AI-enabled financial system,
it may start to look rather more
valuable because the longer-term
question is not simply whether AI
helps somebody complete a mortgage
application more quickly, but what
happens when AI begins to sit between
the customer and the financial
institution itself?
Today, a consumer searches for a
mortgage, compares products, speaks
to an intermediary or approaches a
lender. Tomorrow they may simply
ANDREW LLOYD
is CEO at Fignum
ask an AI agent to find the most
appropriate way of financing their
home. It’s at that point the institution
that owns the customer relationship
maers enormously.
If the AI interface belongs to
somebody else, lenders risk becoming
increasingly invisible manufacturers
of financial products, selected and
exchanged by technology siing above
them. The threat of disintermediation
is therefore not that AI removes
financial services, but that it changes
the relationship through which
services are accessed.
There is another possible future.
Lenders can use the trust they
already possess to become part of that
interface themselves, combining the
capability of AI with the governance,
and accountability of a regulated
institution. Rather than simply
deploying AI for efficiency, they can
begin thinking about how it changes
the relationship with the customer.
That requires technology capable
of evolving with those expectations.
Legacy architecture designed around
individual products, processes and
channels will struggle in a world in
which customers increasingly expect
information and decisions to move
intelligently between them. It’s why I
think the AI debate for lenders should
be so much broader than efficiency.
The technology will become
increasingly ubiquitous but trust will
need to be earned. In the next phase of
financial services, that may turn out to
be the more valuable asset. ●
