SEVEN QUESTIONS FOR HENNA KARNA BY DAN REYNOLDS
As XL Catlin’s chief data officer, Henna Karna is responsible for its digital/data innovation
strategy. Within that realm, she develops data-driven risk solutions comprising genetic
cognitive learning, behavior modeling, proprietary algorithms and deep neural networks.
R& I: What is XL Catlin’s data
H K: XL Catlin’s data strategy
has three parts: It’s about building
sustainable business impact; speed
of converting data into insights; and
extending our culture towards digital
Our businesses and our industry
are guardians of sorts. Our goal is to
help other businesses, and other people,
become whole after a devastating loss.
We do that by examining risk
through the lens of data — structured,
unstructured, bespoke, complex,
generalized and precise. Building a
sustainable data ecosystem helps us to
better understand not just historical
risks but those that are imminent
and upcoming, from cyber to natural
When it comes to empowering
our underwriters, actuaries, claims
and risk management colleagues, we
measure ourselves by speed to value.
Prior to the digitization of data,
our industry had disjointed pictures
of risk. It is likely that our industry
spends more time connecting the data
than leveraging the connected data
that offers valuable insights.
A year or more ago, we had a fair
explanation as to why that was the
case — we couldn’t connect the dots
because not all of them were available
to us. Now, with digital capabilities
in hand, we have the ability and
capability to acquire a more holistic,
real-time view of risk.
This gets to the third part of our
data strategy, digital innovation.
As a niche B2B entity, we deal with
very complex risks. To develop this
360-degree holistic perspective of risk,
we need to engage with clients and
brokers in very close, collaborative
The digital world can help forge
these closer engagements, not only
the ecosystem we’re operating in but
also the sub-tier ecosystems across
the network, both internally and
externally. Our strategy calls for
tearing down internal and external
silos that impede collaboration and
creating stronger bridges across the
network to encourage it.
R&I: How long have you been in
HK: Over a year now.
R&I: What have you found to be
your biggest challenge?
HK: One simple answer is priorities.
It is a thoughtful, iterative activity that
we do. We prioritize all the time; we
calibrate our learnings, changes and
validate our approach constantly —
that’s where we spend most of our
To be clear, prioritization must be
calibrated by our constant learnings.
Sometimes the initiatives we prioritize
are continuing efforts or continuous
improvements or strategic. Each
type of initiative warrants diffident
criterion for success and for milestones.
Hence our way to prioritize must be
transparent and flexible. If you want to
be in a world that is changing, then the
traditional priorities may no longer be
as important as before.
Depending on the type of initiative,
what you measure as a metric for
success may be vastly different, and
so the dimensions for prioritization
would also be quite different.
Human nature is to often abide by
current rules and traditions and only
consider new ones if the current ones
are no longer valid or useful. Now, we
want to maintain what is working and
simultaneously drum up new ideas.
R&I: How do you marry analytics
to the professional experience of
HK: We’re very much an
organization, but we’re an organization
made out of people. Collectively, our
people are our biggest asset. So the
goal becomes how we can further
empower our people.
They already have strong
intuitions — more often than not
they’ve assessed diverse risk-oriented
trends and developed their own
analysis. If we can give our skilled
talent another set of lenses to do what
they already do well, they will do their
work even better.
For example, when we speak to
our underwriters, it is apparent their
accuracy and precision in certain areas
of risk is off the charts; they know
the subject so well. We also have
to balance holistic knowledge and
expertise in a domain. That’s a tough
balance to constantly manage.
If we specialize and get focused on
one thing, we may miss other pieces
of the insurance value chain. One
analogy would be a heart doctor who
knows everything there is to know
about the heart but not nearly as much
as the lungs. Yet both organs function
as parts of a whole.
If we look at things holistically,
though, we risk not seeing all the
symptoms. Much like an internist
who looks for overall symptoms of
health and only “goes deep” if there’s a
Digitalization not only is good
for specialization, but it also opens
our eyes to other parts of the whole.
It compels underwriters to consider
more than what is obvious about a
risk, without having to juggle so many
disparate pieces. In effect, digitization
does the juggling.
R & I : You’re working to create
models that will help underwriters
prioritize which risks to focus on. Is
that process gaining traction?
HK: We rely heavily on our team’s
expertise. Their domain expertise,
breadth of knowledge and intellectual
rigor prepares them to look ahead
at what is in store for a particular
geographic region or in a book of
We’ve now asked them to highlight
the areas where they have achieved
particular success, looking backwards
at the historical to unearth the best
situations, risk-wise. We’re now asking
them to find analogous situations
across their books.
This concept will be ongoing,
despite the predictive power of
machine learning and artificial
intelligence. They’re basically
avatars; we are not replacing the
underwriter. We’re just empowering
the underwriter with vastly better eyes
R&I: How do you share strategies
between books of business?
HK: One way is crowdsourcing — for
example, getting input from across the
enterprise to collectively balance our
book regionally, which helps prioritize
certain resources within the company
and with our partnering brokers. If we
find successful practices in a particular
book, we may recommend our
colleagues to leverage these practices
in their books.
R&I: Is this work helping XL
Catlin get to market faster or get
products to market faster?
HK: In days past, with a new product
or service, typically there will be a
group of people who go off and get
the data, transpose it and think about
which market is affected. All this
happened iteratively from the ground
up, consuming weeks if not months
of time. That’s what slowed down
What we are doing with our
digital transformation is tethering
all of our data and information to
a tree trunk. Each time new data
is discovered, it becomes a bespoke
branch of the tree. Over time, the
tree grows into a resplendent array
of branches, each one a source of
intelligence. But it’s the tree here that
is important, as it represents the whole
— an ecosystem and multi-sided
Henna Karna, chief data
officer, XL Catlin