[ M A R K E T
R E V I E W
|
A I
I N
F U N D
A D M I N ]
nities and work within the organisation,”
adds Kanwar.
This was backed up in findings by
British Telecom (BT), which said AI
developments will result in enormous
opportunities in emerging employment
fields including programmers, algorithm
designers, software engineers, while
creating new job categories such as AI
trainers, ethicists and lawyers. Millennial
employees at fund administrators may
find themselves becoming beneficiaries
of AI developments. Ernst & Young (EY)
recently said that 30% of graduate work
could be automated, although put a posi-
tive spin on the statement by suggesting it
would result in more staff being promoted
to partner status earlier in their careers.
Despite this, some believe financial
services – including fund administra-
tion – may struggle to attract the best
“The benefits of RPA and AI include client
visible quality improvements, timeliness,
accuracy and risk reductions.”
ROB WARD, HEAD OF CHANGE & INITIATIVES, RBC
INVESTOR & TREASURY SERVICES
56
Global Custodian
Summer 2018
and brightest given the opportunities available to them at the
GAFA (Google, Apple, Facebook, Amazon) companies and the
abundant number of technology start-ups in major markets like
California, New York, London and increasingly Paris and Berlin.
Responsible use
While it is important to evolve with AI and develop systems
around it, human relationships are critical as well. As anyone
who has ever interacted with some of the more rudimentary
chatbots will testify, they can be incredibly frustrating to engage
with. “I would say that the majority of service providers are
still learning what the potential is whilst managing an already
hectic book of business. Traction is gaining now, and digital
is becoming more of a focus area for administrators. From a
forward-looking perspective, the human factor and personal en-
gagement is a key asset in the services’ space. Clients gain com-
fort from knowing their accountant and their service provider,”
says David Thornton, global head of fund services product
management, securities services, Deutsche Bank.
The success of AI is also correlated to the quality of the data
that it works with. If fund administrators deploying AI are in
receipt of flawed or inaccurate data, it will yield all sorts of prob-
lems. Perhaps the best-known example of AI malfunctioning
after receiving bad data was Tay, a machine learning chatbot op-
erated by Microsoft. Tay had to be abruptly pulled by Microsoft
when it spouted a deluge of misogynistic and racist comments
after it was trolled on Twitter by malicious users.
As such, it is crucial administrators shadow and monitor the
work of any AI technology. This in itself creates a problem. AI
is generally being used as a cost reduction tool, but building
the technology is expensive, as is hiring people to monitor and
check that it is working as it should be. At least initially, AI could
have some unfavourable cost-economics at administrators,
particularly if it struggles to interoperate with legacy systems.
Again, rebuilding systems does not come cheap, and it will be
critical that AI tools can co-exist with existing infrastructure or
run parallel to them.
Regulating AI is fraught with difficulty too, because it is a nas-
cent concept but it is also continuously changing. One of the big
challenges for regulators nowadays is to keep up to speed with
the frightening pace of technological innovations. Adopting a
prescriptive approach towards fin tech like AI may be counterin-
tuitive and frustrate progress, yet pursuing a laissez-faire stance
has obvious risks. In de minimis, regulators will want providers
like administrators to explain how their AI tools work and make
judgements.
AI’s processing capabilities will speed up administration at a
time when clients are becoming increasingly digitalised, and de-
manding close to real-time data analytics and portfolio informa-
tion. A future in which large volumes of data is supplied to the
administrator, which uses AI to comb through it, and submit it
to clients or regulators through blockchain or distributed ledger
technology (DLT) enabled platforms is looking like an increas-
ingly realistic prospect.