The Intermediary – February 2026 - Flipbook - Page 67
T E C H N O L O GY
Opinion
Less automation,
more intelligence
F
or years, the mortgage
industry has talked about
digitisation as progress.
Paper forms became
PDFs. Wet signatures
became e-signatures.
Yet brokers are still spending days
clarifying information that has
already been submied through a
form. Digitising a broken process does
not fix it. It simply makes people hit
friction faster.
Nowhere is this clearer than in
specialist lending. These complex
cases do not fail because brokers
lack information. They fail
because systems were never built
to understand nuance. As a result,
lenders are forced to revert to manual
processes to make sense of it all. As
we move into 2026, lenders need
infrastructure that can interpret
policy and precedent, not just move
data around.
Death of the black box
One of the biggest barriers to adopting
artificial intelligence (AI) in secured
and specialist lending is trust. Brokers
and underwriters are rightly sceptical
of black box decisions, particularly
in a regulatory environment shaped
by Consumer Duty.
Intelligent infrastructure cannot
replace judgement. It must support it.
In a specialist case, technology should
do the analysis while humans verify
and confirm the outcome. It should
handle the mundane but critical
underwriting work that does not
require specialist judgement, such as
calculating usable income for a selfemployed applicant based on trading
history, determining how much
variable income like bonuses can be
relied on, or flagging cases where
rental income does not meet interest
coverage requirements.
The underwriter can then apply
expertise immediately, with evidence
in front of them. Removing the basic
challenges of transforming data for an
underwriter frees up mental capacity
to focus on what actually maers:
assessing the case based on the
evidence provided, while maintaining
a clear audit trail.
Specialist lending, like vanilla
lending, is still weighed down by
documents. Bank statements,
accounts, payslips, company
structures. Most systems treat these as
files to be stored. Intelligent platforms
treat them as structured data.
When a broker uploads a document,
the system should extract, verify, and
cross-check the information against
lender policy instantly. This removes
the endless back-and-forth caused by
missing data or avoidable queries,
ensuring cases are fully and correctly
packaged upfront.
We are seeing lenders process
specialist applications up to three
times faster when structured,
verified data replaces manual checks.
That speed is not about rushing
decisions. It is about eliminating
unnecessary friction.
Agility beats planning
Specialist lenders need to innovate
continuously, whether through
new products, evolving policy, or
more sophisticated affordability
assessments. Yet legacy technology
means even simple changes can take
months to implement. That is not
innovation. It is constraint, dressed
up as progress.
True agility comes from no-code
infrastructure. Credit and policy
teams should be able to update
rules themselves and see them live
immediately.
If a system cannot reflect a policy
change within days, brokers will
feel the gap long before IT catches
up and will move to lenders who
respond faster.
Intelligent technology removes the
administrative burden. This frees
brokers to advise and underwriters to
assess risk properly.
JOY ABISAAB
is CEO at Mast
The next evolution is agentic
underwriting. Data is automatically
collated, structured, and assessed
against underwriting policy, informed
by previous decisions and lenderspecific risk logic. A first layer of
reasoning is applied before the case
ever reaches a human. The system
arrives with a recommended outcome,
the supporting evidence, and the
policy rationale already applied.
Only then does the underwriter
step in. Not to gather information,
but to confirm, challenge, and apply
judgement where it truly maers
When technology handles
execution and pre-analysis,
underwriting timelines that once took
weeks can be reduced to days without
compromising risk. For brokers,
this means faster decisions, fewer
resubmissions, and more reliable
outcomes for clients. For lenders,
it creates a genuine and defensible
competitive edge.
The bottom line
This is not about removing brokers
or underwriters. Final decisions still
sit with people. It is about stripping
away repetitive, low-judgement
manual work that is prone to error
and too oen mistaken for risk
management. When systems apply
policy, learn from precedent, and
present a complete, evidenced
recommendation, humans are freed to
do what machines cannot: exercise
judgement.
In specialist lending, speed without
intelligence creates risk. Intelligence
without human oversight creates
mistrust. Agentic underwriting
delivers both. In specialist lending, the
future belongs to systems that think
first and humans who decide last. ●
February 2026 | The Intermediary
65
