The Intermediary – August 2026 - Flipbook - Page 95
T E C H N O L O GY
Opinion
AI is becoming
infrastructure,
not innovation
T
here are moments when
a regulator publishes
a report that changes
the rules. Then there
are moments when it
simply acknowledges
that the market has already begun
moving. The Financial Conduct
Authority’s (FCA) Mills Review feels
much closer to the laer.
Its significance lies less in
announcing a new artificial
intelligence (AI) strategy than in
recognising that AI is no longer
a future technology. It is rapidly
becoming part of the operational
fabric of the industry.
The question is therefore shiing
from whether firms should use AI, to
how they implement it responsibly,
transparently and in a way that
genuinely improves outcomes.
The Mills Review does not
recommend an entirely new AI
rulebook. Instead, it argues that the
existing principles-based framework,
centred on Consumer Duty and the
Senior Managers Regime, remains
broadly fit for purpose. The emphasis
is therefore on explainability and
accountability rather than technology
for technology’s sake. AI should help
firms make beer decisions, not
obscure them.
That aligns remarkably closely
with the findings from our own
Mortgage Pulse Report. Rather than
predicting wholesale automation,
lenders increasingly view AI as a way
of removing repetitive manual activity
so that experienced people can spend
more time exercising judgement.
Whether that means validating
information more efficiently,
reducing unnecessary administration,
supporting underwriting decisions or
managing post-contract variations,
the technology is being viewed as an
enabler of beer decision-making
rather than a replacement for it.
Consumers remain comfortable
with AI supporting decisions,
organising information and
improving efficiency. They become
significantly more cautious when AI
begins making autonomous decisions
without appropriate oversight. In our
conversations, the lending industry
appears to understand this.
Across the Pulse research, there
was lile appetite for replacing
experienced professionals with
algorithms. Respondents consistently
identified opportunities where AI
could improve speed, consistency
and operational resilience while
leaving accountability firmly with
the business. This mirrors the FCA’s
vision of AI acting as an enhancement
to expertise rather than a substitute.
Exceeding expectations
Operational excellence will become a
competitive differentiator. For years,
digital transformation has been
discussed as though it were simply
about replacing paper with portals.
The Mills Review looks considerably
further ahead. It envisages an
industry where AI becomes embedded
throughout customer journeys,
operational processes, fraud
prevention, regulatory reporting and
supervisory oversight.
AI is no longer another project
siing alongside the business. It
increasingly becomes part of how the
business operates. That inevitably
places greater emphasis on the quality
of the underlying data.
An AI model cannot compensate
for fragmented processes, incomplete
customer records or disconnected
systems. In fact, poor data simply
allows mistakes to happen
more quickly.
STEVE CARRUTHERS
is growth director at Fignum
In the Pulse report, lenders
consistently highlighted data quality,
integration and process consistency
as the foundations for meaningful
AI adoption. Competitive advantage
will come from interoperable
infrastructures. Organisations with
connected systems, structured data
and repeatable processes will be able
to deploy AI far more effectively
than those aempting to automate
inefficient workflows.
The Mills Review also reflects on
the use of agentic AI to increasingly
undertake actions on behalf of
consumers and firms, within
agreed parameters. That could
fundamentally reshape servicing,
payments, lending and customer
interactions over the coming decade.
For lenders, however, the immediate
priority is not autonomous agents. It is
preparing the operational foundations
that make responsible AI possible.
This means investing in data quality
as well as deploying sophisticated
models. It means digitising processes
that still rely on manual intervention.
It means ensuring decisions remain
transparent and explainable. Most
importantly, it means recognising
that AI is becoming part of business
infrastructure rather than a separate
innovation programme.
The industry, if the findings from
the Fignum Pulse Report are any
guide, is already beginning to move
in exactly the right direction. The
organisations likely to benefit most
will not necessarily be those with the
most advanced models, but those that
have quietly built the right operational
foundations. ●
August 2026 | The Intermediary
93
