Fdy2eusin0rzf PDF
Fdy2eusin0rzf PDF
Fdy2eusin0rzf PDF
Artificial intelligence
and the future of
financial services
COMMISSIONED BY
The road ahead:
2
Artificial intelligence and the future of financial services
Contents
3 About the research
4 Executive summary
8 Main benefits
13 A transformational journey
18 Conclusion
Our thanks are due to the following individuals for their time and insight:
This report was authored by Dewi John and edited by Katya Kocourek.
Executive summary
The financial services industry has long been an early adopter of
technology. The telegraph system was still a novel idea when Western
Union began using it for money transfers in the mid-19th century.
Online banking emerged in the mid-1990s, half a decade after the
internet and well before most people had an email account.
In order to gauge the effect AI is already having Within the category of heavy adopters, virtual
among those making most use of it, the survey assistants, machine learning (ML) and predictive
looks at the specific technologies being used by analytics are making the running followed by natural
“heavy adopters” (those who indicated that language processing (NLP) and image analysis.
70 71
63
60 62 61
60
58 58
57 56
54 55
52 52 52
50
50
46 46
45
44
40 42
39 39
35
33
30
20
10
0
Predictive analytics Machine learning Virtual assistants Natural language Image analysis Robotic process
(e.g. chatbots) processing automation
Investment banks are taking the lead in respectively), which mainly reflects the level
implementation of most AI applications, including of investment available to big firms for a
NLP and ML, while retail banking has the edge multitude of AI technologies. This puts the
in predictive analytics (71% adoption), which larger firms in a good position to deal with
reflects the significant usage of data science the burden of overcoming legacy systems.
tools in customer retention. However, insurance
Of the heavy adopters, the main perceived
lagged in all fields. A recurring theme throughout
benefit of AI for around 40% is increased
the research, this is probably due to the fewer
employee capacity to handle volume of general
and relatively simpler products in the insurance
work. In stark contrast, light adopters do not
industry compared with the banking sector.
consider this a main benefit (at just 27%). It
Overall, larger organisations (with 5,000+ appears that in order to reap this benefit there
employees) have higher AI penetration than is a hurdle of a certain level of investment that is
their smaller counterparts (54% and 49% simply unattainable for many light adopters.
61 %
of organisations in APAC
are significantly more
likely than others to be
heavy adopters of AI
Main benefits
The benefits of AI are many and often vary between Regarding such capacity benefits, Cary Krosinsky,
sectors and regions. Overall, companies see AI a sustainable finance lecturer at the Yale School
as an important lever to innovate, launch new of Management, says this is, in effect, using
products and services and enter new markets. new tools to achieve an old objective: “what the
industry has always attempted to do—maximise
In the survey round, lower operational costs
returns”. Some 36% of heavy adopters also saw
emerged as the top benefit of AI, as cited by 37%
more efficient product and marketing services
of respondents. Around a third said the same
as a significant benefit, a view shared by just 23%
about facilitating data-driven decisions through
of light adopters. This is probably because these
greater use of predictive analytics and increasing
benefits derive from market monitoring that
employee capacity to handle larger volumes of work.
can only come into effect when more “core” AI
systems are in place for companies.
0 5 10 15 20 25 30 35 40 45
34
Greater use of predictive analytics 36
(eg, for data-driven decisions) 33
31
33
Increased employee capacity 30
to handle volume 30
39
32
Enhanced customer personalised 26
service and customer satisfaction 35
36
31
Reduced employee workloads 44
30
25
*The above chart includes respondent answers in the five strongest categories for this particular question.
Source: The Economist Intelligence Unit
Improved risk management, such as fraud can invest heavily to reap the benefits. Smaller
prevention, was the main perceived benefit operations that don’t have the scale face an
for APAC respondents (46%), while reduced increased risk of going to the wall. Arguably, large
operational costs and reduced employee operations should be larger, leaving niche players
workloads were the other two predominant to service more specialised needs.” He speculates
perceived benefits (44%). It’s possible that these that second-tier firms may make easier merger
two factors tie in with the fact that APAC is and acquisition (M&A) targets, leading to further
the location of many employee-heavy service consolidation across the financial services sector.
centres where these technologies are already While similar proportions of heavy and light
having an impact. adopters selected enhanced customer service
For Alaa Saeed, MD and global head of as a benefit of AI implementation, varying
institutional eSales and Client eCom products proportions (66% of heavy adopters and 43% of
at Citibank in London, the benefit of the AI light adopters) selected customer/stakeholder
technologies underpinning many of these satisfaction as a measurement of success.
developments “is huge because it standardises North Americans have the greatest ambitions
things”. Such standardisation in areas such as here with 33% believing AI will change how they
NLP and ML can be followed by better controls, innovate and 31% saying that it will allow them to
governance and efficiencies of scale. This is a release new products and services. Those figures
“relatively new scenario,” he says, made possible are lower for APAC and Europe (see Figure 3).
by software platforms integrating chatbots and
Despite this, respondents from APAC and North
automating ever-more complex requests that
America see the greatest opportunity to enter
were previously resource- and people-intensive.
new markets (at 30% and 27% respectively). This
In addition, these technologies could lead to a reflects the higher rates of economic growth in
much-needed shaking out of financial services, both regions overall compared with the rest of
reckons Mr Krosinsky. “Large operations such the world as well as the level of AI investment
as JP Morgan have the advantage that they from individual firms to support business growth.
0 5 10 15 20 25 30 35 40 45
34
38
Lower our cost base 38
28
31
Increase need for high-value 41
technology skills 21
33
27
Lead us to develop new products 23
and services 24
31
25
Allow us to enter into new 30
markets or industries 19
27
25
Change how we innovate 16
22
33
25
Increase exposure to 25
technology-related regulation 27
20
*The above chart includes respondent answers in the six strongest categories for this particular question.
Source: The Economist Intelligence Unit
Despite their lower overall commitment, it’s the of bankers share this view. This may be because
insurers who predict the greatest impact of AI— insurers’ lower commitment thus far allows for
32% expect to see a significant impact on both a greater base effect, with a similarly notable
their product shelf and manner of innovation effect on the narrower product shelf they have in
over the next five years. Only about a quarter comparison to investment and retail banks.
While most respondents and experts agree in Europe (41%). The discrepancy is largely
that gauging the success of AI applications is attributable to the fact that 6% of European
important for business strategy, there are diverse respondents say metrics had not been in use
views regarding the most reliable metrics. for long enough to make an assessment, while
3% had no established metrics whatsoever.
Customer and stakeholder satisfaction were This contrasts with the view from APAC where
the prime measures of AI success, much more the figures were zero in each case—all APAC
so for APAC respondents (66%) than those respondents have workable metrics in place.
0 10 20 30 40 50 60 70
55
Customer and/or stakeholder 66
satisfaction 41
55
52
Reduction in operational costs 62
48
48
50
Achieving expected return 56
on investment (ROI) 37
53
45
Contribution to strategic goals 49
35
48
39
Lower instances of fraud and 46
other financial crimes 43
33
Mr Saeed notes the importance of customer Reduction in operational cost was the second
satisfaction to gauge success, especially in the key metric, followed by the impact on the
areas of NLP and ML where there is significant expected return on investment (ROI). These
client demand for services such as automated factors scored significantly across all three
sectors, but especially so for retail banking.
chats and request for quotes (RFQs), both of
The impact on ROI was deemed particularly
which rely on such technologies. While this
significant in APAC (56%), closely followed by
may carry “a franchise risk of inadvertently
North America (53%) and in contrast to just 37%
responding incorrectly to your client or a of European respondents. APAC respondents
group of clients,” Mr Saeed says, “it’s less of also report a reduction in operating costs
a high risk than a market impact risk”. But he as the second most important factor (62%).
explains that “the framework [for customer Regardless of the sector, however, these three
service and chats] is becoming more robust”. measures comprised the top three metrics.
A transformational journey
The impact of these technologies on how financial years there will be far fewer financial services
companies are structured will be profound. jobs “and one knock-on effect could be that this
will depress real estate prices in these cities”.
Kerry Peacock, EMEA chief of operations and
international head of operations at MUFG in Greater adoption of AI will nevertheless be
London, highlights one effect on the hitherto gradual, particularly in the banking sector.
ubiquitous call centre, shedding light on why retail “I’ve started to introduce robots into my
banks are leading in virtual assistants. “If you go operation,” says Mr Peacock. “[In] doing
back even as recently as five years, for heavily
that, you have to overcome what’s called
manual functions that are repetitive and process
automation anxiety; the ‘robots are going to
driven you would look to a low-cost geography
take my job’ type of fear.” As such, introducing
such as India to perform those tasks. That was
robots into the workplace is “something that
yesterday’s strategy. Today and tomorrow, you
has to be done very carefully,” he says.
move to a digitised workforce and build robots.”
Mr Krosinsky agrees that a major impact of Similarly, Mr Krosinsky believes that it’s not
these changes will be to “do away with many simply a question of job replacement. In
traditional jobs” in a way that could extend well some areas AI will “supplement and enhance
beyond offshorable manual jobs. This could be actual people”. In this respect, everything
transformative for those cities with high levels leads back to people: taking the strain off
of dependence on financial services, such as them, or at the very least allowing them
London and New York. He believes that in five to do more with the same workload.
0 5 10 15 20 25 30 35 40 45
39
46
Cost of technology
35
33
29
Insufficient infrastructure to 31
accommodate new AI technologies 25
31
28
Insufficient data quality to test 44
and validate AI outcomes 22
23
27
28
Lack of appropriately skilled staff
29
20
23
Lack of awareness of AI use cases 25
among senior management 24
23
⅓
AI may offer an alternative to such frequent
replacement of big and hugely expensive core
legacy systems that are deeply embedded
into companies, argues Mr Peacock. “You
can put new technologies around the legacy
systems which means you don’t need to Only one third of larger firms
necessarily change that core system.” This (5,000+) saw the cost of
should allow businesses to be “more nimble technology as a major barrier to
around those core technologies,” he explains. the adoption or use of AI.
Training and reskilling will be vital for financial The level of value-add to the business assumes a
services firms to implement innovative greater degree of investment into technological
products and services in the future. In terms infrastructure that should make AI applications
of AI specifically, the workforce will require more compatible with existing systems.
different and more complex skills as time
Europe was marginally ahead of APAC in asserting
progresses. This is recognised by respondents
the need for retraining, but 11 percentage
and experts whose focus is not only on how
points behind when asked if such training had
AI changes the quantitative nature of what
been implemented (APAC: 54% vs Europe:
employees will be doing but also the qualitative
43%). This may simply reflect the fact that
aspects of their job: in short, upskilling.
Asia increasingly leads the field in technology,
“There is an expertise and staffing that you have as Mr Krosinsky notes. “With Asia heading to
to build,” says Mr Saeed. “But we’re seeing the become half the world’s economy, a lot of these
skillsets of our people change. And so our people developments will happen there. Within Asia,
are becoming more technical, more quantitative. given that Hong Kong is now less attractive
And our technology team and front office as a financial centre, Singapore has a massive
team are getting closer and closer aligned.” opportunity to take the lead, although this is
something that China might sensibly resist.”
The importance of technological skills is
emphasised by our survey respondents.
0 10 20 30 40 50
49
Total
42
54
APAC
44
Europe 43
49
North America 39
41
Investment banks are most advanced in the this.” Clearly, a sea-change in reskilling will
implementation of training schemes—54% necessitate greater investment in people.
of respondents say they have already been
implemented compared with 46% in insurance Diverging regional views were also seen in the
and 48% at retail banks. This probably explains expectations of technology training between
why only 17% of all respondents see a lack of respondents performing an IT function and those
specialised training as a risk to AI adoption. performing other roles. Whereas 33% of the
“Increasingly, we’re running computer science former and 29% of the latter see an increasing
or coding training courses for our folks,” says need for high-value technology skills (broadly
Mr Saeed. “There’s a ton of investment into the same), only 17% of tech-focused respondents
this space which tells you what we think see this as resulting in retraining and reskilling
about, where we’re going and the benefit of as opposed to 30% of non-tech respondents.
Conclusion
AI is at the forefront of a major shift within changed,” Mr Peacock explains. “Coming back to
the financial services industry, but periods of basics, you can either buy or sell, borrow or lend.
rapid change are not without their risks. That’s literally all you can do. It’s as simple as that.”
There is nevertheless an awareness of the risks Businesses that are able to get ahead of the curve
associated with AI technologies within businesses, in AI adoption appear to be those carrying less
and in some cases there are clear strategies technological baggage, making legacy systems
to navigate them. However, coming to terms simpler to deal with. The benefits of greater
with some of these—notably the technological AI adoption are widely recognised across the
and associated regulatory risks—may yet take financial services industry, including reduced
a while. This is especially pertinent for banks cost base and better predictive analytics. Such
whose business has not fundamentally changed innovation, and its costs, will inevitably drive
and is unlikely to do so in future. “If we look at consolidation. And, ultimately, the focus on
the business, if you look at the products that we customer satisfaction as a crucial measure of
generate as financial institutions, they haven’t success will drive more optimal market outcomes.
0 5 10 15 20 25 30 35 40 45
40
Security considerations 38
41
45
29
Technology risk 33
18
34
26
Amount of investment required 28
25
30
22
Regulatory challenges 16
32
23
21
Maturity of technology 25
(eg, legacy systems) 18
20
LONDON GENEVA
20 Cabot Square Rue de l’Athénée 32
London, E14 4QW 1206 Geneva
United Kingdom Switzerland
Tel: (44.20) 7576 8000 Tel: (41) 22 566 2470
Fax: (44.20) 7576 8500 Fax: (41) 22 346 93 47
Email: london@eiu.com Email: geneva@eiu.com