[ I N
D E P T H
|
S E L F - S E R V I C I N G ]
“Serving up data through an application-
type software is definitely an area where
the market will go and something we are
investing in.”
JUSTIN CHAPMAN, GLOBAL HEAD OF MARKET
ADVOCACY AND INNOVATION, NORTHERN TRUST
heavily in new technologies to facilitate
the adoption of this model.
Application Programming Interfaces
(APIs) have been identified as an enabler
for this model, whereby an open-ended
architecture can gather data sets across a
number of environments into one and can
interact with other systems.
“We have started to see client requests
to access APIs for their data pulling
systems that can help them enhance their
own front-office decisions. Serving up
data through an application-type software
is definitely an area where the market
will go and something we are investing
in,” says Justin Chapman, global head of
market advocacy and innovation research,
Northern Trust.
These APIs can be tailored to specific
business needs and requests, but could
mean certain challenges for banks that
have traditionally worked on a ‘one-size-
fits-all’ basis. There would have to be a
change in tack to ensure these APIs are
flexible for individual requests.
“This event-driven architecture helps
30
Global Custodian
Spring 2019
us to move into the middle-office, and data-pulling technologies
can help in those investment decisions. However, the design of
this architecture must allow flexibility around client needs and
for their future demands. The key is to look at the end-to-end
technology stack, see where you can apply self-service applica-
tions, and provide that data as-a-service,” Chapman adds.
Machines talking to machines
Data access tools have become a value-added service, providing
clients the right tools to extract the information themselves in
real-time and in a more personalised manner. To facilitate this,
machine learning could be the technology of choice.
Panjwani explains Citi is currently piloting the use of new
technologies such as chatbots and natural language process-
ing (NLP) for securities services, allowing clients to get query
responses much quicker than calling someone.
While chatbots can be utilised to answer fundamental ques-
tions, NLP can be deployed as a service to utilise search history
and patterns based on the client. In order to do this, custodians
will have to get as much machine learning into that system as
they can and build those future response capabilities within the
API.
BNY Mellon’s Todd also believes machine learning and NLP
would be ideally suited to facilitate self-servicing. “For NLP, ma-
chines are talking to machines, and for the unstructured data it
can become very valuable as you can take this info on scale, look
at the history, real-time info, and begin to automate that. This
could be for targeted info for the client, or translating an individ-
ual request into an active issue to look at. These are becoming
more ubiquitous and more useful for self-servicing.”
As custodians look to build out their data-as-a-service capabili-
ties, it will also help them stand out in a commoditised market.
“The ability to provide an event-driven application to help
clients get that information is something that custodians could
provide to help differentiate themselves. Clients will be looking