I N D E P E N D E N T P U B L I C A T I O N B Y
R A C O N T E U R . N E T
1 0 / 1 2 / 20 20
# 07 07
AI FOR BUSINESS
IS BRITAIN STILL AN AI
LEADER?
SIX WAYS AI CAN HELP
SAVE THE PLANET
10
03
WHEN TO TRUST ROBOT
RECOMMENDATIONS
12
R A C O N T E U R . N E T
03
0 0/ 0 0/ 2 0 2 0
I N D E P E N D E N T P U B L I C A T I O N B Y
# 0 0 0 0
R A C O N T E U R . N E T
/ai-business-2020-dec
he UK has been at the cut-
ting edge of artificial intel-
ligence (AI)
innovation,
from Alan Turing, the pioneer-
ing mathematician and computer
visionary, who launched the field,
to DeepMind’s AlphaGo, the first
computer program to defeat a pro-
fessional Go player in 2015.
Several pioneering AI companies
were founded in the UK, including
DeepMind, SwiftKey and Magic
Pony, all of which were acquired by
US companies – Google, Microsoft
and Twitter – for $500 million, $250
million and $150 million, respec-
tively. Over the last few years, the
UK government has launched its
Office for AI and Centre for Data
Ethics and Innovation. But is the
UK still an AI leader?
In 2019, McKinsey Global Institute
placed the UK in the top quartile for
“AI readiness”. How is the UK main-
taining this position in a competi-
tive landscape, both in a business
sense and a governmental one?
No country can hold a candle to
the United States and China when
it comes to AI, but the UK is one of
Europe’s leaders, according to the
McKinsey report. The UK is glob-
ally in the top quartile for research,
startup investment, digital absorp-
tion, innovation foundation and
ICT connectedness. It does, how-
ever, rank lower on automation
potential and human capital.
The UK has many leading research-
ers, who are published in the top
academic journals. Christine Foster,
chief commercial officer at The Alan
Turing Institute, says: “The UK
has eminent researchers, such as
Christina Pagel, who works on math-
ematical tools to support delivery
of health services; Mark Girolami,
who is developing and applying
advanced statistical and computa-
tional techniques to engineering
challenges; Maxine Mackintosh,
who has founded One HealthTech,”
which supports under-represented
groups in health tech innovation.
There are many others.
Lee Harland, founder and chief
scientific officer at SciBite, an
Elsevier company, says: “We’re very
good at the basic science; a strength
of the UK has always been our intel-
lectual output. The Cambridge-
Oxford-London triangle is a hub
for talent. Because AI is a broad
skill that can fit just as much into
gaming as it does into healthcare,
within the triangle there is a lot
of opportunity for people to move
around, even into different indus-
tries, without trekking halfway
across the world.”
The question is whether this skill-
set filters down to a broader pop-
ulation. “Recruiting talent from
outside the UK will always be
important, but we need to bring the
AI skills closer to our schools and
universities,” says Harland.
“You don’t need a degree in
mechanical engineering to drive
a car and you don’t need a degree
in statistics to use machine-learn-
ing.  There are some great initia-
tives for data science and AI-centric
courses appearing in our univer-
sities; this needs to be accelerated
and cascaded down, at least con-
ceptually, to school age.” Like most
countries, the UK faces a shortage
of people with advanced technolog-
ical skills. Wider education could
remedy that.
Foster adds: “In the UK, we are
fluidly connecting and conven-
ing across the public sector, pri-
vate sector and third sector. Look
at the AI Council [an independent
expert committee that advises the
government]; it’s a great example
of what can happen when people
from industry, public sector and
academia come together, sharing
their broad range of background
and expertise to the AI ecosystem.”
Despite slightly higher invest-
ment in AI, the UK lags behind
France, Germany,
Japan and
South Korea when it comes to AI
patents, according to McKinsey.
What’s more, an
independent
review commissioned by the gov-
ernment noted that “universities
should promote standardisation
in transfer of intellectual prop-
erty”. This would make it easier to
create spin-out businesses.
Taking an idea and turning it into
a business takes a combination
of factors, says Harland. “There
are a lot of institutions out there
to advise – Innovate UK, Digital
Catapult – but it’s often very obtuse
in terms of what they can do and
how they help.” He says other
European countries are better at
being explicit about which agencies
do what for startups. “I think it’s
very hard to understand that in the
UK landscape,” says Harland.
There is something of a “space
race” in the AI realm, says Dr
Michael Feindt, strategic adviser of
Blue Yonder. America is investing
fifty times more in AI than the UK,
and China is investing eight times
more. “We are increasingly see-
ing promising UK startups being
acquired by large US companies
before they can mature, limiting the
UK’s ability to make up ground on
other countries,” says Feindt.
Historically, many innovations in
the computer industry have been pio-
neered by women. The first computer
programmer was Lady Ada Lovelace,
while actress Hedy Lamarr invented
the technology that enabled wifi, GPS
and Bluetooth. In the mid-80s, almost
40 per cent of US computer graduates
were women. But the AI industry now
faces what Bill Gates called the “sea of
dudes problem”. A greater diversity
of people and data would counteract
some of the bias that algorithms have
ingested so far.
To stay at the forefront of AI,
the UK needs a long-term strat-
egy spanning ten to fifteen years,
rather than just one or three, argues
Foster at The Alan Turing Institute.
This strategy needs to ensure data
is more accessible to AI companies,
that innovative pilots can be scaled
and ethical frameworks applied.
“We have a long history in AI. Our
researchers know they’re standing
on the shoulders of giants and that
we have the ability to move the whole
field forward,” she concludes.
Is the UK still
an AI leader?
AI FOR BUSINESS
@raconteur
/raconteur.net
@raconteur_london
It may be one of Europe’s major players when it comes to artificial
intelligence, but a lack of skills and strategic investment may be
holding the UK back from its full potential
MaryLou Costa
Business writer and
editor specialising
in marketing, tech
and startups, with
work published in The
Guardian, The Observer
and Marketing Week.
Marina Gerner
Award-winning arts,
philosophy and finance
writer, contributing to
The Economist's 1843,
The Times Literary
Supplement and
Standpoint.
Sam Haddad
Journalist specialising
in travel, with work
published in The
Guardian, 1843 Magazine
and The Times.
James Lawrence
Freelance journalist
specialising in business
and technology. Senior
Contributing Editor for
I-Global Intelligence
for Digital Leaders and
former Editorial Director
at Redwood Publishing.
Chris Stokel-Walker
Technology and culture
journalist and author,
with bylines in The New
York Times, The Guardian
and Wired.
Jonathan Weinberg
Journalist, writer and
media consultant/
trainer specialising in
technology, business,
social impact and the
future of work and society.
Distributed in
Marina Gerner
Contributors
Although this publication is funded through advertising and
sponsorship, all editorial is without bias and sponsored features
are clearly labelled. For an upcoming schedule, partnership
inquiries or feedback, please call +44 (0)20 3877 3800 or
email info@raconteur.net
Raconteur is a leading publisher of special-interest content and
research. Its publications and articles cover a wide range of topics,
including business, finance, sustainability, healthcare, lifestyle and
technology. Raconteur special reports are published exclusively in
The Times and The Sunday Times as well as online at raconteur.net
The information contained in this publication has been obtained
from sources the Proprietors believe to be correct. However,
no legal liability can be accepted for any errors. No part of this
publication may be reproduced without the prior consent of the
Publisher. © Raconteur Media
raconteur.net
T
C3.ai is solving previously
unsolvable problems
This is Enterprise AI.
© 2020 C3.ai, Inc. All Rights Reserved.
is a mark of C3.ai, Inc.
Energy
Management
Precision Health
Inventory
Optimization
Predictive
Maintenance
Anti-money
Laundering
T:264 mm
T:338 mmC O M P E T I T I O N
+22%
TOP25%
TUC/BritainThinks 2020
The UK sits in the upper quartile
for global AI readiness
Boost to the UK economy that
could be provided by AI by 2030
+20%
Increase in the AI gap between
Europe and the US in the past 3 years
Diego Barbieri/ShutterstockPublishing manager
Jamie Oglesby
Acting managing editor
Francesca Cassidy
Associate editor
Peter Archer
Digital content executive
Taryn Brickner
Production manager
Hannah Smallman
Art director
Joanna Bird
Design director
Tim Whitlock
Design
Sara Gelfgren
Kellie Jerrard
Colm McDermott
Samuele Motta
Nita Saroglou
Jack Woolrich
Sean Wyatt-Livesley
R A C O N T E U R . N E T
A I F O R B U S I N E S S
05
04
There are often concerns that artificial
intelligence will replace people’s jobs, but in
the case of forward-thinking multinational
Schneider Electric, the opposite is true
rolling out the platform would have
on some of its people, particularly
those in middle-management roles.
Frequently, they have felt their staff
are more open to being “poached”
internally, while being unable to see
the broader business benefits of a
more dynamic workforce.
“What I don't think we did incred-
ibly well was the change man-
agement around mindsets,” says
Pelletier. “Do not underestimate
the people part of it and the fact
you have to be open to shifting
and rethinking, not only your HR
department, but how managers
and employees are equipped to deal
with this.”
However, once these human ele-
ments are addressed Schneider’s
team leaders are usually able to see
the bigger picture. “Most progressive
managers get it,” she says.
As for the future, Pelletier’s boss,
chief
human
resources
officer
Charise Le, is clear about how
Schneider needs to double down
on the outcomes the Open Talent
Market
is delivering. “When it
comes to talent, we need to achieve
empowerment for all,” she says.
“Expectations of employees may
change, but the need to make your
own career choices will not.”
Pelletier is excited about the pos-
sibilities of using AI to unlock fur-
ther value, particularly when it
comes to operating at pace. “Speed
is the key to winning in the market,
whether it’s with talent or with our
business,” she says. “AI has brought
speed to us that we’ve never had
before. And that's why we continue
to keep looking at it.”
employees to vacant roles, helping
them find a mentor and connect-
ing them to side projects. Crucially,
it puts employees in control of their
own careers. People are free to share
whatever personal information they
feel is relevant, such as skills and
goals, which is then matched by the
system’s algorithms to the compa-
ny’s requirements.
“Previously, we would humanly
try to make matches, but we weren’t
able to bring the supply and demand
together,” says Jean Pelletier, vice
president of digital talent transforma-
tion at Schneider, who played a lead-
ing role in launching the project. “We’d
been asking to do this for years, as we
outgrew our ability to operate without
it. The spirit was there, but the technol-
ogy was missing, and that's where AI is
the game-changer.”
ntil
early
this
year,
Schneider Electric was liv-
ing with an uncomfortable
truth. Some 47 per cent of employ-
ees who left the global energy man-
agement and automation business
said they were leaving because they
couldn’t see any future career oppor-
tunities. It was clear that, in this
company of 140,000 people, the tra-
ditional means of internal talent
recruitment and career progression
were not working.
To solve the problem, Schneider
launched its Open Talent Market, an
innovative application of artificial
intelligence (AI) in human resources
that is helping to place the compa-
ny’s considerable internal expertise
where it is most needed.
The
platform
currently
per-
forms three functions: matching
U
James Lawrence
What’s more, applying technolo-
gy-driven solutions like this is par-
ticularly crucial in organisations
that are undergoing a digital trans-
formation, says Josh Bersin, a lead-
ing HR industry analyst who special-
ises in HR technology.
“The more ‘digital’ your company
becomes the more project-based it
needs to be,” he argues. “So we need
tools and systems to facilitate this
new world of work and I’m excited
to see them here at last. Creating a
talent network in your company will
greatly improve your retention. And
when your people feel safe to try
new things, contribute to other pro-
jects and share their expertise, they
can innovate and solve problems
faster than ever.”
A further benefit is the way it
helps to enhance the company’s
diversity and inclusion initiatives,
says Pelletier. “We can’t help but be
human and have unconscious bias.
But AI looks at hard facts, it looks at
skills, it’s making things agnostic,”
she says.
However, she is also aware of
the possibility of in-built prej-
udices lurking within the
system’s algorithms. “We’re
super
vigilant
towards
that,” she says, but is con-
fident when the technol-
ogy
is combined with
human skills, the result
is far superior to where
Schneider was before. “It’s
brought science to where
we used to only have art
and now we’ve found a good
balance between the two.”
Of course, rapidly imple-
menting and scaling a system
like this in a 184-year-old global
enterprise is always likely to throw
up challenges. “This is by far the
most disruptive technology we have
brought into Schneider. It’s a com-
plete rewrite of HR,” says Pelletier.
She explains the company failed
to predict the effect that rapidly
Transforming
the workforce
with AI in talent
management
Despite this only being rolled out
globally in April, Schneider is already
seeing the business benefits. Although
it’s too early to say exactly how the sys-
tem has affected the employee attrition
rate, the early signs are encouraging.
Some 38,000 of the company’s 75,000
white-collar workers have already
enrolled and there’s a plan to make it
available to blue-collar employees via
on-site kiosks. Meanwhile, an imme-
diately visible upside is that manag-
ers looking for suitable internal can-
didates for vacant roles have been able
to reduce the time taken for sourcing
“from months or weeks to seconds”,
says Pelletier.
Helping people find
suitable
side projects is also transforming
employee experience at Schneider,
whose workforce are encouraged
to spend 10 to 15 per cent of their
time on areas that fall outside
their usual role. “We’re measur-
ing that as ‘unlocked hours’,” says
Pelletier. “Those hours are not only
the employee making discretionary
effort for development, it’s us actu-
ally sourcing internally for the skills
we don't have resident on our teams.”
But the overarching business bene-
fit of using AI in HR in this way is the
visible increase in dynamism the tal-
ent market is fostering. “We have cre-
ated an internal gig economy within
Schneider that is delivering exactly
the agility we need,” she says.
Don’t underestimate the
people part of it and the fact
you have to rethink how
managers and employees are
equipped to deal with this
C A S E S T U D Y
Commercial feature
AI comes into
its own in the
fight against
financial crime
With the financial crime landscape
constantly evolving, AI is now providing
banks with a faster, smarter way to
reduce false positives, gain a more
holistic view of customer behaviour
and reduce costs in the process
oney
laundering
tech-
niques have evolved sig-
nificantly as criminals have
leveraged
technological
advances.
With business and society becom-
ing more connected, financial crimi-
nals have adapted quickly. Meanwhile,
increasingly stringent regulations and
increased numbers of fines levied
against banks mean financial services
organisations are spending $180.9 bil-
lion annually on financial crime com-
pliance, 62 per cent of which goes
on labour expenditure in the Europe,
Middle East and Africa region, the
LexisNexis Risk Solutions Global Study
found. All of this to recover less than 1
per cent of all the criminal proceeds,
according to the United Nations.
Identifying this activity using tradi-
tional methods is extremely difficult.
While banks have been seeking to
adopt technology that does this more
efficiently by simply doubling down on
existing systems, anti-money launder-
ing (AML) processes tend to remain sig-
nificantly siloed, with a lack of cohesion
between systems and departments.
Legacy technology is a great inhibitor,
slowing down banks at a time when
open-source technology is enabling
criminals to adapt and evolve more
quickly. It’s imperative that banks find
ways to look for vulnerabilities in sys-
tems more generally, and intelligently,
while promoting closer alignment
between departments.
“Banks have always kept infor-
mation restricted, sharing it on a
need-to-know basis between depart-
ments. Consequently, they’re unable
to get a holistic view of their clients,”
says Dr Janet Bastiman, head of ana-
lytics at regtech company Napier,
whose intelligent compliance plat-
form helps banks increase efficiency
and minimise risk. “They tend to have
multiple teams for onboarding, client
life-cycle management, transaction
monitoring and sales. Each of these
teams can also be split geographi-
cally, so they don’t interact well or
share data and insights. Things are
quite literally falling through the gaps.
“The cybersecurity sector is very
good at communicating new threats
quickly, so everybody can immedi-
ately start patching. We need to have
the same approach with money laun-
dering. If somebody knows there has
been suspicious activity, that infor-
mation and how to recognise the new
patterns needs to spread quickly to
prevent recurrence. This isn’t hap-
pening because banks don’t have the
systems, processes or technology to
share the information and get a truly
holistic view.”
Napier’s award-winning compli-
ance platform provides the compre-
hensive view of customers that banks
need. The company’s
intelligent
approach, which successfully com-
bines big data technologies with arti-
ficial intelligence (AI), robotic process
automation and machine learning, is
applied to underpin policy, process
and procedure. The Napier platform
is fast, scalable and modular, mean-
ing that financial institutions don’t
need to replace their existing sys-
tems immediately and can build their
sophistication incrementally.
The software helps different depart-
ments work together more effec-
tively. As information runs through
the system, it forms a top-level over-
view that then provides alerts to the
appropriate teams at the right time.
Client onboarding and KYC (know
your customer) checks, for example,
can be powered by contextualised
information from numerous sources,
and compared with behaviour from
other customers and entities. A cus-
tomer’s behaviour may look normal
when viewed in isolation, but looking
at it more holistically, next to all other
sources of information, could show
some unusual patterns.
“We want to make compliance
officers sleep easily. With Napier,
banks can better understand the fun-
damental interconnectedness of their
data,” Bastiman adds. “You have the
customers and how they’re connected
to other customers and all their trans-
actions, and being able to see that spi-
der-like view really exposes any incon-
sistencies. But to do this you need
that holistic view, a customer-centric
approach, rather than just looking at
siloed transactions.”
Taking
this more
intelligent
approach also brings other benefits,
including freeing up the time of com-
pliance analysts who no longer have
to sift through reams of transactions
to try to spot patterns. Reducing the
number of people required on these
kinds of investigations, and feeding
people with accurate information
quickly, means humans can focus on
more sophisticated tasks.
The technology is powering better
explainability in a regulatory sense
too. It’s not enough to say an issue
was flagged by AI. Analysts need the
detail in a simple, digestible lan-
guage so they can explain to regula-
tors exactly what caused concern.
Historically, AI has not been suc-
cessful here, with any explainabil-
ity focused only on metrics for data
scientists. Napier’s Client Activity
Review has AI flags that show in plain
English what the unusual transac-
tion was in that period for the client,
and why it was unusual. This enables
anyone in the team to work with the
insights, without requiring a data sci-
entist to interpret the data.
“Criminals will always be trying to
hide their activities and be one step
ahead,” says Luca Primerano, Chief
AI Officer at Napier. “As economics
change in the world, whether that’s
political or through major world events
we’ve seen this year, it’s going to bring
them new opportunities and also new
challenges. The financial industry must
really adapt to those challenges while
stopping the opportunities for crimi-
nals as fast as they can.
“An AI engine can look at the trans-
actional activities of customers much
better than a human. It’s a completely
independent set of lenses that can
go through billions of transactions
across multiple dimensions to detect
anomalies, something that would take
humans years to complete at great
cost. The AI then collates that infor-
mation into a simple summary of the
top suspicious behaviours, includ-
ing why they are unusual, so that a
human analyst can make better deci-
sions. Napier provides a completely
unified solution with these capabil-
ities, eliminating silos and combin-
ing everything together to gain that
holistic view of customer behaviour.
This helps banks work faster and
smarter to fight financial crime, while
at the same time reducing costs.”
For more information, visit napier.ai
M
We want to help compliance
officers sleep easily
Cost of compliance in
Europe makes up
of the total global cost
(LexisNexis Risk Solution Study)
75%
MAKING THE CASE FOR AI: THE HUGE COST OF FINANCIAL CRIME COMPLIANCE
Global compliance fines against Financial Institutions
Overview of global compliance spend of
technology and labour
Financial institutions spend globally
$180.9bn
to catch less than 1% of money laundering
Other
$5bn
Technology
$72bn
Labour
$103bn
$10bn
$8bn
$6bn
$4bn
$2bn
2016
2017
2018
2019
$1.4bn
$0.8bn
$4.5bn
$8.4bn
LexisNexis Risk Solutions True Cost of Financial Crime
Compliance Study Global Report 2020
Fenergo Global Financial Institutions Fines Report
IMPLEMENTING AI IS AN HR
CHALLENGE TOO
1300 HR executives from across the
globe were asked whether preparing
the workforce for AI was the biggest
challenge for their function
(Values rounded)
KPMG 2020
21%
22%
Neutral
Disagree it’s the
biggest challenge
Smith Collection/Gado/Getty Images56%
Agree it’s the
biggest challenge
R A C O N T E U R . N E T
A I F O R B U S I N E S S
07
06
Enables fl exible working (such as home or remote working) Enables more effective communication Reduces commute times/costs for staff if working from homeHelps employees have more control over their work and working patternImproves effi ciency and frees up time to focus on more meaningful tasksEnhances employee voice (such as through an intranet)Enables collection of data to help inform organisation’s wellbeing approachEnables immediate feedback to be given to staffNone - there are no positive effectss remote working becomes
increasingly
common-
place, keeping employees
engaged and interested in their
work, while struggling with the
stresses and strains of life dur-
ing a pandemic, is no easy task.
But AI and employee engagement
can dovetail together to provide
employers with an overview of how
to ensure wellness runs through an
organisation and pick up on issues
before they arise.
All of us are being tested in ways
we have never been before, as we
struggle under the pressure of roll-
ing lockdowns, time away from
family and juggling work-life bal-
ances. Sentiment analysis can
help ensure an engaged employee
remains engaged, and can pick up
on issues with health and wellbeing
from those who feel uncomfortable,
at a time when unemployment is
reaching record highs, about com-
ing forward.
The movement in AI and employee
engagement is being spearheaded
by a range of startups that are work-
ing with major employers, helping
them feel more able to get a grip on
where employees are facing issues,
and off ering solutions to problems
when they arise.
“We’ve built an extension arm,
an anonymised dashboard, which
aggregates this pool of data that
says, ‘It looks like in your popula-
tion of employees in London, 67 per
cent are at risk of stress or anxiety,
43 per cent of diabetes. And liter-
ally 100 per cent of your people are
at risk of musculo-skeletal con-
ditions,’” explains Lorena Puica,
chief executive of iamYiam, a big
data analytics fi rm.
The company takes countless
anonymised data points and, using
machine-learning, translates them
into a predicted cost of whatever the
issues raised will be to an organi-
sation, providing suggestions on
how to support employees from the
top down. “The idea is to have this
integrated end-to-end, from the
employee to the corporate and then
back to the employee,” says Puica,
whose clients include large consult-
ing organisations, law fi rms, insur-
ance companies, and healthcare
in professional services fi rms and by
between 5 and 7 per cent in retail.
The twinned roles of AI and
employee engagement are known
by many people. Bupa, the pri-
vate healthcare provider, uses AI
to monitor health and wellbeing
among its employees worldwide,
with a tool developed by Glint, a
Silicon Valley startup.
“In the past, you had to employ
data scientists to understand what’s
going on in your organisation,” says
Nigel Sullivan, chief people offi cer at
Bupa. “You try and pull out the driv-
ers of engagement. They are things
specifi c to your organisation that
might have a disproportionate eff ect
on engagement. It might be com-
munication or the prospects of the
fi rm. It’ll be diff erent depending on
the circumstances.” But AI enables
Bupa to get to the heart of what’s
troubling employees and off ers sug-
gestions how to fi x it.
“It’s like skittles: you hit one
and get the whole shebang,” says
Sullivan. “Your bang for your buck
is a lot better if you can fi nd out
what the drivers are. AI helps you
get that.” Bupa uses natural lan-
guage processing to fi lter through
free text responses, in eight lan-
guages worldwide, to its survey of
83,000 workers and pinpoint what
are each of their concerns. Three
quarters of Bupa’s employees com-
pleted the most recent survey, con-
ducted in late-November, providing
68,500 comments.
“We can really analyse that and
fi nd out what it is people are think-
ing about and what’s on their mind,”
services and enterprises worldwide.
The real challenge is tackling the
productivity crisis in workforces
and ensuring workers feel supported
at a time when things are highly
uncertain and a number of diff erent
aspects of life tug and pull at their
time. The UK has some of the worst
rates of absenteeism and presentee-
ism in the world, according to the
Chartered Institute of Personnel and
Development, which has a knock-on
eff ect on productivity.
iamYiam has managed to reduce
absenteeism in the companies with
which it works by between one and
two days per person a year. But pre-
senteeism, where people turn up
but aren’t engaged with their work,
is a bigger drag on businesses’ bot-
tom lines. Here iamYiam claims to
improve presenteeism by between
ten and twenty days a year.
“Productivity is that elusive term
everyone talks about, but no one
can grasp,” says Puica. But iamY-
iam’s analysis of key performance
indicators in a company, and sug-
gestions on how to improve it, can
increase productivity by 10 per cent
says Sullivan. “What’s important to
people working in our hospitals in
Spain or insurance companies in
Hong Kong? What do they think?”
Glint enables Bupa’s team managers
to identify the drivers of employee
engagement and provides advice on
how to maintain or improve them.
Other companies rely on bots to
communicate with workers and col-
late their responses. Moneypenny,
which manages call centres and live
chat environments for 21,000 cli-
ents in the UK and United States,
has rolled out the use of bots on
Workplace from Facebook to keep
in touch with workers, identify their
issues and communicate changes.
“For our people, it helped that true
human interaction continued as we
embraced this new normal, recre-
ating those water-cooler moments,
which are the lifeline for a peo-
ple-focused business like ours,” says
Joanna Swash, Moneypenny’s chief
executive. “We have used it pro-
actively to distribute positive and
uplifting news and messages.
“We try to not impose too many
top-down
initiatives,
but
use
Workplace as a tool to get feedback
and ask questions about how the
management teams can better sup-
port frontline staff .”
And this is the concern, that the
shift to AI and employee engagement
could backfi re as already stressed
workers begin to worry about support
turning into surveillance. Some have
expressed concerns with the rollout of
what detractors say is “employee sur-
veillance” software.
Demand for such tools is up 51 per
cent since the start of the corona-
virus pandemic, according to data
compiled by Top10VPN. Search
traffi c for “employee monitoring
software” has risen 65 per cent
between March and September,
while searches
for “work-from-
home monitoring tools” are 2,000
per cent higher than they were
pre-pandemic.
Some companies, struggling to
keep tabs on their employees and
worrying about a decline in pro-
ductivity as the pandemic bites, are
changing their approach to using
AI and employee engagement from
one that benefi ts employees to ben-
efi ting bosses.
It’s being exacerbated by the
unprecedented situation in which
we fi nd ourselves during the pan-
demic and the sheer newness of the
technology. “The speed of change
in this space is truly unprece-
dented,” says Puica at iamYiam.
“When you have something that
changes so fast, the challenge is
you’re not catching downsides or
mistakes fast enough.”
Caution is required and clear
thinking about why you’re rolling
out the use of AI. Employees may
be discomfi ted by the immense
changes going on in their workplace
and need reassurance and stability.
“We need to create a value set that
drives policies,” says Puica, before
we jump into the unknown.
Alistair Berg/Getty ImagesAs employees’ wellbeing is tested to its limits, caring
employers are using a range of AI tools to ensure
concerns are being heard and properly addressed
How to check in on
your distributed
workforce
A
Chris Stokel-Walker
CIPD 2019
WHAT DO EMPLOYEES REALLY WANT TECH TO DO FOR THEM?
UK employees on which advances in technology have had a positive effect on
their workplace wellbeing
E M P L O Y E E W E L L B E I N G
When you have something that
changes so fast, the challenge is
you’re not catching downsides
or mistakes fast enough
What’s important to people working in our
hospitals in Spain, or insurance companies in
Hong Kong? What do they think?
Top10VPN 2020
51%
jump in demand for employee
surveillance software since the start
of the coronavirus pandemic
74%52%47%43%29%27%26%17%11%
R A C O N T E U R . N E T
A I F O R B U S I N E S S
09
08
77% 56%
74% 53%
70% 48%
69% 45%
65% 42%
60% 33%
58% 32%
From chatbots and digital assistants to facial recognition
or biometric scanners, our daily interactions with artifi cial
intelligence have surged over the past few years, most of
them without us even realising it. This infographic explores
some of the ways that AI has infi ltrated our day-to-day
lives and how consumers generally feel about it
Number of times a day
that Gen-Z consumers
unlock their phones
Number of Netfl ix paid subscribers in the third
quarter of 2020, up 37 million year-on-year
Verto Analytics 2019
Netfl ix 2020
Capgemini 2020
79
Unlocking your phone
Netfl ix recommendations
Share of global consumers who have AI-enabled interactions
with organisations over the following frequencies
Daily
Weekly
Fortnightly
Once a
month or less
2018
2020
It will be the fi rst thing many do as soon as they wake
up, but some may be surprised to know that the simple
act of unlocking a smartphone by looking at it relies on
AI. Apple’s TrueDepth camera, for example, projects
30,000 invisible dots on to a user’s face to create a
so-called ‘depth map’, and compares that to the saved
data to allow access. It can even automatically adapt to
changes in appearance, such as facial hair or make-up.
Netfl ix says its recommendation system “strives to
help you fi nd a show or movie to enjoy with minimal
effort”. It assesses a variety of factors, such as your
viewing history, how you rate titles, what others with
similar tastes have watched, which actors or
genres you like to watch and things like the
time of the day you use the service. These
all feed into Netfl ix’s algorithm, which is
improved every time you watch something new.
Number of digital voice
assistants in use worldwide
Juniper Research 2020
4.2bn
Speaking to
smart assistants
Virtual assistants such as Alexa and Siri rely on
voice recognition software and natural language
processing. They break down questions or
phrases into individual sounds, then run those
sounds through a database, using sophisticated
algorithms to fi nd the right answer. As more
people use the assistants, the database of sounds
expands and the algorithm learns as it goes.
A D A Y
I N T H E
L I F E O F
AI INTERACTION FREQUENCY
A I
of spam, phishing and malware
is blocked on Gmail
Google 2020
99.9%
Blocking unwanted emails
Sophisticated spam fi lters such as those used by Gmail
rely on deep learning, where the algorithms learn
from users clicking ‘report spam’ and ‘not spam’, and
adapt accordingly. It tailors inboxes to users’ habits,
for example learning to fi lter out emails that individuals
tend to quickly delete or ignore. Gmail also uses a so-
called artifi cial neural network, which recognises and
fi lters out certain kinds of messages, such as sneaky
phishing attempts.
21%
31%
14%
33%
54%
27%
11%
6%
Salesforce 2019
Consumer and business buyer attitudes towards AI worldwide
PUBLIC ATTITUDES TO AI
Capgemini 2020
Percentage of global customers who are satisfi ed with AI
interactions by industry
SATISFACTION WITH AI INTERACTIONS
195m
I`m open to the use
of AI to improve
my experiences
AI will play as big of
a role in my life as
smartphones
AI will revolutionise
how I interact
with companies
AI is the most
signifi cant technology
of my lifetime
I trust companies
to use AI in a way
that benefi ts me
I can think of an
example of AI I use
every day
Companies are
transparent enough
about how they use AI
61%
58%
58%
54%
53%
increase in chatbot
usage by B2B customers
from 2019 to 2020
Drift/Heinz Marketing 2020
Chatbots
Designed to simulate human conversation, chatbots
operate via chat interfaces on customer service
portals, interpreting written words inputted by
customers to provide a pre-set answer. Their ability
to respond to complex questions is limited, but they
have come a long way over recent years.
93%
people use Grammarly
to improve their writing
Grammarly 2020
30m
Spell check
Doing something as simple as composing an
email can call in the use of AI. Grammarly is
an AI-powered writing assistant that suggests
improvements to grammar or spots errors in
users’ writing. The company says its AI also listens
to feedback from humans – for example if several
users choose to ignore a certain suggestion,
adjustments are made to the algorithms to
make them more accurate.
Banking and insurance
Automotive
Public sector
Consumer products
and retail
Utilities
Consumers
Business buyers
R A C O N T E U R . N E T
A I F O R B U S I N E S S
11
10
The Living Planet Index produced by
WWF estimates that wildlife popu-
lation sizes have dropped by 68 per
cent since 1970. The charity advo-
cates the use of artifi cial intelligence
(AI) as a tool of conservation technol-
ogy to monitor and curb this alarm-
ing rate of decline.
One of the most useful applica-
tions is in acoustic monitoring,
recording the sounds of wildlife
ecosystems on weatherproof sen-
sors. Many animals, from birds
and bats to mammals and even
invertebrates, use sound for com-
munication, navigation and ter-
ritorial defence, providing reams
of rich data on how a species pop-
ulation is doing. AI provides a fast
and cost-eff ective way to analyse
hours of recordings for patterns
of behaviour.
Conserving species
Conservation Metrics, a California-
based company, has used acoustic
listening and machine-learning to
monitor endangered populations of
both red-legged frogs in Santa Cruz,
diverting water to help them mate
successfully, and the forest elephants
of the Central African Republic, help-
ing to protect them from poachers.
Facial recognition technology is
another application of AI that could
help track wildlife populations, when
combined with camera traps in the
wild. BearID, an open-source appli-
cation, which was trained on brown
bears in Canada and the United
States, is a recent AI triumph as,
unlike primates, zebras or giraff es,
bears don’t have distinguishing fea-
tures, so the deep-learning algorithm
had to fi nd patterns in their facial
make-up instead. The researchers
hope this AI will be used to monitor
other species in the future.
From facial recognition technology that monitors brown
bear populations, to intelligent robots sorting recycling, these
initiatives are having a positive impact on the environment
Using AI to save
the planet
1
4
2
More than 2.1 billion tonnes of rub-
bish is generated in the world each
year, yet only 16 per cent of it is
recycled, according to research by
Maplecroft. To make matters worse,
a quarter of waste put into the recy-
cling is not actually recyclable at all,
hindering the whole process.
Several startups are now looking
at how AI and sustainability goals
can be combined to make recycling
more effi cient, even when dealing
with mixed materials. Colorado-
based AMP Robotics uses an
AI-powered robot with optical sen-
sors to quickly identify rubbish as
it passes on a conveyor belt. It then
sorts it with its robotic arms, using
the company’s AMP Neuron AI plat-
form, which can recognise diff erent
textures, colours, shapes, sizes and
even brand labels.
The AI constantly updates itself
and is designed to run 24/7. It has
already been rolled out in the United
States, Canada and Japan, and will
soon be coming to Europe.
In Bali, Gringgo Tech has designed
an image recognition tool to help
informal waste collectors identify
the diff erent monetary values of
various recyclable materials. In a
pilot study, it improved recycling
rates by 35 per cent. They’re now
working with Google to build AI into
the platform to help improve how
quickly and effi ciently the system
can categorise waste.
Improving recycling
Nine in ten of the world’s urban
residents breathe polluted air,
prompting the United Nations to
make access to cycling, walking or
public transportation one of its 17
Sustainable Development Goals.
To meet this challenge, London-
based Vivacity uses AI technology
to capture and classify live trans-
port usage with the goal of enabling
more environmentally sustainable
transport use in cities. The company
has been working with Transport
for London since 2018 to determine
where new cycling infrastructure
should be targeted.
London’s Walking and Cycling
Forests are home to 80 per cent of
the world’s terrestrial biodiver-
sity, and they absorb and store a
third of current carbon emissions.
Halting the loss and degradation
of forest ecosystems is essential to
meeting the objectives of the Paris
Agreement on climate change,
according to the International
Union for Conservation of Nature.
Rainforest
Connection
seeks
to combat illegal logging using
acoustic monitoring
in forests
on hidden solar-powered smart-
phones, which have been recycled
from consumer use. The charity
then uses AI to analyse this sound
data in real time. If the AI detects
the sounds of chainsaws, logging
trucks or gunshots, an alert is sent
Cutting air pollution
Protecting forests
Commissioner Dr Will Norman says:
“By getting more people cycling and
walking, we can help to tackle con-
gestion and pollution in London and
improve our health. Our Healthy
Streets approach is based on evi-
dence and data, and we welcome
new technology that supports this.”
Vivacity’s AI has allowed local
authorities across the UK to assess
the eff ectiveness of their temporary
street layouts to encourage physi-
cally active travel during the coro-
navirus crisis. The company has
also helped Transport for Greater
Manchester roll out smart junc-
tions across the city, which prior-
itise pedestrians and cyclists over
motor-vehicle traffi c.
to rangers. According to Rainforest
Connection,
research
shows
that if illegal loggers are inter-
rupted once or twice, they leave
and don’t return until the next
logging season.
Dryad Networks has secured
seed funding to use the internet of
things and AI to detect wildfires.
Dryad uses AI-based solar-pow-
ered sensors to capture gases emit-
ted at the smouldering stage of
a wildfire which, combined with
real-time analysis of temperature,
humidity, air pressure and wind
data, will alert forest rangers when
a wildfire is imminent. They are
also developing a long-range wire-
less environmental monitoring
sensor network to cover large for-
est areas where there is no mobile-
phone signal.
Sam Haddad
3
5
6
Some 9.5 million tonnes of food is
wasted in the UK every year, accord-
ing to the Waste and Resources
Action Programme, 70 per cent of
which could be avoided. The waste,
which includes food from super-
markets, households and hospital-
ity, generates 25 million tonnes of
greenhouse gas emissions.
Winnow is working with HCL
Technologies to use AI to tackle
the problem in hospitality, where
their data shows up to 15 per cent
of purchased food is being wasted.
Winnow Vision is an AI tool that
takes pictures of food as it’s thrown
into the bin, teaching itself to rec-
ognise what’s
been discarded
Minimising food waste
and tracking the data. IKEA has
deployed Winnow Vision in its UK
stores, cutting food waste by an
average of 50 per cent.
Last year, UK supermarkets signed
up to a government pledge to halve
food waste by 2030. According to
data from Blue Yonder, using AI in
supermarket supply chains could
help the UK’s eight largest retail-
ers cut seven tonnes of food waste a
year, saving £144 million. As Wayne
Snyder, vice president of retail strat-
egy, Europe, Middle East and Africa,
at Blue Yonder says: “AI monitors
goods from farm to fork, resulting in
an increased understanding of the
environmental impacts across the
supply chain and identifi cation of
the areas that need improving.”
Raw sewage was discharged onto
beaches in the UK almost 3,000
times over the last year, according to
a report by Surfers Against Sewage.
The environmental charity advo-
cates stricter monitoring of sea and
river pollution, and operates an app
called the Safer Seas Service, which
warns swimmers, surfers and other
water users when untreated sewage
has been released at their beach.
But the app, which began in 2010
as a text alert system, relies on vol-
untary data provided by water com-
panies, which isn’t always relia-
ble. So, this year, Surfers Against
Sewage added a health report func-
tion to the app, using a citizen sci-
ence approach to warn others about
beach cleanliness issues in real time,
but also to hold water companies
Reducing sewage
pollution
to account. Southern Water, for
example, had released no notifi ca-
tions during 2020 due to reporting
mechanism errors, yet over 20 per
cent of health reports submitted to
Surfers Against Sewage allegedly
came from beaches within Southern
Water’s jurisdiction.
In the future, application of AI
will enable even more precise, live
seawater
quality
assessments.
Scientists working with the National
Research Foundation of Korea have
already shown that artifi cial neu-
ral network models can accurately
predict microbial contamination at
beaches, using variables including
tides, temperatures, wind speed and
direction, rainfall and recent sew-
age discharges. Southern Water has
set a target of zero pollution inci-
dents by 2040 and say they will use
state-of-the-art machine-learning
in that mission.
Commercial feature
AI revolution explodes. As cloud com-
puting takes over the world of data, the
workplace is no longer in one physical
location, it is atomised and scattered.
The puzzle of co-ordinating product
quality, service and operations, and
worker performance in the AI era, will
need a new map of clues, dramatically
changing management styles from tra-
ditional top-down structures to decen-
tralised and remote-working practices.
AI is interweaving into our lives in ways
we would never have imagined; people
need to understand its impact.
A lot of people fear AI will steal
their jobs. Are they right to
have these concerns?
A power struggle is in the process
of erupting as a shift towards
automation and predictive systems will
displace workers. Those carrying out
roles that can largely be automated
may feel they are losing power, while
incomprehensible volumes of power
will be placed in the hands of others.
Our AI journey must be clearly mapped
out. AI cannot be left to drive its own
train of progress. Its flexibility gives
everyone the freedom to steer it in dif-
ferent directions. The destination must
be clear from the onset, with a soci-
etal sat nav of directions. This is why
democratising AI education is so vital.
What films do you have in
the pipeline to achieve that
democratisation?
Visitors to our website can now
rent our first film The Quest
for Super Intelligence. The cinematic
How has artificial intelligence
evolved over the years?
It goes back further than many
people realise. Alan Turing took
the first steps to test a machine’s ability
to model human intelligence in post-
war Britain and Frank Rosenblatt’s
1958 invention of the perceptron
algorithm accelerated the idea that
artificial intelligence (AI) could mimic
human thinking. But the next truly
major milestone didn’t arrive until 1997
when IBM’s Deep Blue defeated chess
grandmaster Garry Kasparov. The new
millennium flared inspiration. Honda
launched ASIMO, a humanoid robot
purposed to offer home assistance for
people with mobility issues. IBM’s new
super computer Watson was victorious
in the American quiz show Jeopardy
and Google’s AlphaGo became the first
machine to beat a human professional
at Go, a complicated board game. AI is
no longer just an abstract sci-fi con-
cept; it is part of our everyday lives,
through smart cars, robotic devices
and virtual assistants such as Siri
and Alexa.
What did you learn about AI in
your time in academia?
I spent 29 years in academia
including 19 years as a lecturer.
What struck me was how little the
average person understands about AI
and what it can and can’t do, as well
as the lack of alignment on the subject
between scientists, academics, tech-
nology professionals, business lead-
ers and wider society. At its simplest
level, AI enables human integration
with intelligent sensors feeding data,
progressing to levels of hyper-connec-
tivity within all economic spheres. But
nobody is explaining this in a way that
is easily digestible and understanda-
ble or dispelling fears. Some people
restrict AI education because they try
to mystify it. They’ll charge thousands
of pounds for courses they deliver over
numerous weeks. We make it main-
stream by delivering the same informa-
tion in a highly engaging, feature-length
cinematic film. We’re democratising
AI education and we’re the only ones
doing it in this format, partly because
nobody else has the courage to do it.
Why is it so important that
people are better educated
on AI?
A big question is how we sup-
port and protect people as the
Democratising
AI education
A new company, DataWorkout, is raising awareness of disruptive
innovation through cinematic filmmaking. Why? Founder and chief
executive Angel Javier Salazar says it’s crucial to democratise
education of technologies as powerful and transformational as
artificial intelligence
educational experience provides a
comprehensive view of the evolution,
risks and challenges to implement AI
in businesses and for the benefit of
society. We explore how intelligence
is evolving to mimic us as humans, the
fears of automation, the new skills
required, the platforms available, and
issues around employment, diversity
and inclusion. Released in January,
our second film The Lord of the Blocks:
Blockchain Unchained, directed by
Dr Christian de Vartavan, takes view-
ers back to a medieval setting of cas-
tles, wizards and knights to explain this
revolutionary new technology. We are
planning several more films with engag-
ing stories to demystify new technol-
ogies. Our mission is to inspire and
help enthusiastic people enjoy learn-
ing from the comfort of their home.
DataWorkout will directly contribute
to developing the skills of the future
workforce through entertaining “edu-
films” that spark innovation.
For more information please visit
dataworkout.com/SuperAI
S U S T A I N A B I L I T Y
Michele Aldeghi/Shutterstock Rawpixel.com/ShutterstockGeran de Klerk /Unsplash Clem Onojeghuo/Unsplash Nito/ShutterstockDebby Hudson/Unsplash
R A C O N T E U R . N E T
A I F O R B U S I N E S S
13
12
Marketing researchers Chiara
Longoni and Luca Cian’s newly
published paper, Artifi cial
Intelligence in Utilitarian versus
Hedonic Contexts: The “Word-of-
Machine” Effect, highlights varying
scenarios where consumers
are more likely to favour an AI
recommendation system and,
conversely, where human input
is preferred.
In real estate, haircare, food
and clothing, the majority
of users chose the human
recommendation over artifi cial
intelligence (AI) when asked to
focus on experiential and sensory
attributes, such as style, taste
or scent. When tasked with
focusing on practical elements,
such as use-case and function,
most people opted for the AI
recommendation.
Yet the researchers note it
doesn’t mean AI should only
be used when it comes to
more utilitarian products, such
as technology or household
appliances, or that companies
offering more hedonic items, such
as fragrances or food, shouldn’t
be using an AI recommendation
system. In an experiment where
they framed AI as supporting
human recommenders rather than
replacing them, the AI-human
hybrid recommender fared as well
as the human-only one.
Who we trust and when?
Commercial feature
As the glow of
back-office
RPA fades,
spotlight
turns to front-
office virtual
assistants
Evolution of virtual assistants, driven
by robust natural-language processing
and ease of use, is allowing businesses
to automate customer support and
improve employee productivity
ow well a company han-
dles business interactions
defines
its performance.
Within most companies, custom-
ers and employees wait too long for
support, from getting answers to
simple questions to executing com-
plex transaction, and support chan-
nels provide limited self-service and
poor personalisation.
The flood of routine tasks into
contact centres drives up opera-
tional cost and reduces agent effi-
ciency, and
limited
IT resources
makes it difficult to automate routine
interactions internally.
Building chatbots
to
automate
front-office interactions using com-
ponents from Big Tech players, such
as Google, Amazon and Microsoft,
requires IT resources, time and big
budgets. All this has driven companies,
particularly banks, to explore a new way
to automate business interactions.
Leading companies are turning to vir-
tual assistants powered by conversa-
tional artificial intelligence (AI). Gartner
predicts that by the end of 2021, 40 per
cent of digital workers will use a virtual
employee assistant daily, up from only
2 per cent in 2019, and virtual assistants
will automate 69 per cent of front-of-
fice workloads by 2024.
The first wave of chatbot offerings
left much to be desired. Chatbot 1.0
technology was expensive and dif-
ficult to use and could automate
only simple FAQ (frequently asked
question)
interactions. The
lim-
ited ability of chatbots to under-
stand, manage and lead customer
conversations compromised cus-
tomer satisfaction and capped the
containment rate, which is the meas-
ure of how many interactions are
automated without escalation to a
live agent.
“Gone are the days of chatbots that
can’t actually have a conversation,”
says Raj Koneru, chief executive at
Kore.ai, which provides a conver-
sational AI platform for configuring
virtual assistants and pre-trained
industry and functional virtual assis-
tant products. “Improvements in
ease of use and the robustness of
natural-language technology have
helped companies start and scale
their front-office automation pro-
grammes faster and deliver a better
customer and employee experience.”
Knowing where to aim virtual assis-
tants and understanding how much
can be automated has been a key
challenge for companies. “If you think
about the types and volumes of inter-
actions within a business as a triangle,
where simple FAQs are at the wide
bottom and strategic debates are the
peak, the bottom 80 per cent of inter-
actions can be automated with virtual
assistants,” says Adam Devine, chief
marketing officer at Kore.
Both customers and employees
follow a similar journey into and within
a business, starting with FAQs during
onboarding, followed by a series of
transactions along with efforts to
retain and develop, whether that’s
upselling customers or making human
resources and IT support effortless
for employees.
The key to automating this contin-
uum of interactions, while ensuring
a great experience, is natural-lan-
guage processing that can identify
both intent, for example “trans-
fer funds”, and entity, for example
“from current account to savings”,
and manage and lead dialogue with
contextual awareness and empathy
across any channel.
Intelligent conversational user expe-
rience gives people instant, personal-
ised responses from a business, and
integrations between virtual assis-
tants, enterprise systems and robotic
process automation (RPA) make these
conversations actionable by automat-
ing a wide range of transactions.
Virtual assistants also improve con-
tact centre agent performance by
identifying successful outcomes and
prompting agents with next-best
actions. By managing omnichannel
business interactions on a single uni-
fied platform, conversational AI users
get advanced operational analytics
that not only show containment rates,
but also provide nuanced trends and
insights into customer and employee
behaviour and agent performance.
The inevitable question for busi-
nesses, particularly banks, is build
or buy? Microservices from Big Tech
players allow businesses to create
purpose-built bots with full con-
trol, but it’s cost, time and resource
intensive. On the other side, pure-
play vendors offer prebuilt virtual
assistants that enable organisations
to go to market almost instantly, but
with limited customisation and a
heavy reliance on vendors for sup-
port and product updates.
A newer option that eliminates the
cons and accentuates the pros is a
no-code conversational AI platform.
With this approach, business users get
all the tech components they need to
configure customised virtual assistants
along with the option to buy preb-
uilt industry solutions, such as bank-
ing, and prebuilt functional solutions,
like HR and IT service management.
Neither the platform nor product
paths require coding, which democra-
tises virtual assistants for any size com-
pany and every stakeholder with a use-
case for a virtual assistant.
Kore is one of a few next-generational
virtual assistant software companies
pioneering this approach and giving
customers a faster and more efficient
alternative to Big Tech. It is the con-
versational AI partner to 100 Fortune
500 companies, including the top four
banks and top three health organisa-
tions, and 500,000 employees and 70
million retail consumers interact with
its virtual assistants.
“The benefits speak for them-
selves,” says Devine. “Typically,
businesses that use conversational
AI are able to reduce 30 per cent
from their front-office costs. That’s
a huge win for support teams. Even
a single percentage point for big
contact centres, which spend hun-
dreds of millions of dollars each
year on live agents and technology,
really adds up. The speed of service
increases tenfold, which increases
speed to revenue.
“But probably most importantly,
when you’re talking about financial ser-
vices business in particular, you’re able
to improve your customer satisfaction
or net promoter score by 25 per cent
and fend off competition by born-digi-
tal fintech competitors.”
Customers have chosen Kore to ride
the virtual assistant wave for its unified
conversational AI platform that pro-
vides a single user experience across
all digital channels and superior natu-
ral-language processing capabilities,
combining machine-learning, funda-
mental meaning and industry-focused
knowledge graph to deliver the highest
automation rates and accuracy.
Asked what is the biggest barrier for
Kore and other next-generation vir-
tual assistant providers, Devine had a
one-word answer: “Awareness. If chief
information officers, chief operating
officers and customer service execu-
tives knew how easy and efficient it is
to spin up a virtual assistant that deliv-
ers human-level performance, there
would be no such things as wait times,
dropped calls and customer attrition,
and employees would get more done
and be a lot happier.” This is good
news for customers and a challenge to
expensive, black-box Big Tech.
For more information please visit
kore.ai
H
Gone are the days of
chatbots that can’t actually
have a conversation
THE BEST PERFORMING COMPANIES AUTOMATE THE BOTTOM 80% OF THE INTERACTION PYRAMID
PEOPLE WORK
BOT WORK
Strategic,
philsophical
interactions,
Routine,
binary
interactions
Debate
Open-ended
discussion
Transactions
General enquiries
Business enquiries
EMPLOYEE INTERACTIONS
“How many vacation days left on my credit”
“How many new opportunities in Salesforce this week”
“Schedule a meeting on Friday with Jon and Alex”
CUSTOMER INTERACTIONS
“Add international plan to my wireless”
“I need an appointment with Dr. Smith for dental checkup”
“Transfer five thousand from checking to savings”
When to rely on robots
T R U S T
Luis Alvarez/GettyImagess more of our every-
day life, from shopping
and dating,
to
learn-
ing and exercise, takes place dig-
itally, there’s an opportunity for
artifi cial intelligence (AI) to serve
large online audiences and create
business effi ciencies.
IDC analysts forecast worldwide
spending on AI will double to $110
billion in 2024, while data from
digital assistant company Amelia
reveals 88 per cent of US organisa-
tions have scaled up their use of AI
since the pandemic began. But have
we reached a tipping point where
consumers trust an AI recommen-
dation system more than a human?
Not yet,
according
to new
research published in the Journal
of Marketing, based on data from
more than 3,000 people who took
part in ten experiments. When it
comes to AI and trust, the key factor
is whether consumers are assess-
ing the practical aspects of a prod-
uct – its utilitarian value – or its
experiential, sensory aspects – its
hedonic value.
“When people are
looking for
things that have to do with practical-
ity, functionality, decisions that are
more cognitively driven, that’s where
they tip over to AI,” says Dr Chiara
Longoni, co-author of the study and
assistant professor of marketing at
Boston University’s Questrom School
of Business. “When it’s a question of
anything sensory related, a human is
usually perceived as best.”
Yet these “lay beliefs” don’t “fully
correspond to the facts” about
the competency of both human
and AI recommendation systems,
Longoni adds.
And as Dr Luca Cian, fellow co-au-
thor and assistant professor of mar-
keting at the University of Virginia’s
Darden Business School, elaborates:
“It’s not that humans, in reality, are
always better at making recommen-
dations when it’s something sensory
related. And computers in reality
aren’t always better when it’s some-
thing utilitarian.
“Humans can be as good as com-
puters in establishing something
utilitarian. And there are many
times when AI is good at making
decisions that are sensory related.
For example, spice and drinks com-
panies use algorithms to create new
fl avours and they work well.”
Human biases do mean AI recom-
mendation systems lend themselves
more to certain sectors, says tech
entrepreneur Emma Smith, founder
and chief executive of Envolve Tech,
which has created a virtual shopping
assistant used by brands including
We Buy Any Car.com and BHS, now
an online-only retailer.
In mass-market retail verticals
with huge product variety, such
So if consumers’ perception of
AI isn’t refl ective of its actual rec-
ommendation abilities, with our
beliefs ingrained by portrayals of
robots in popular culture, what does
this mean for businesses looking to
leverage it?
Longoni and Cian recommend
a hybrid approach, over a poten-
tially “creepy” or misleading overt
humanisation of AI recommenda-
tion systems.
“People are more amenable to AI in
cases where there’s a human compo-
nent. It doesn’t make people prefer
AI to a human, it simply equalises
the preference for human or AI
advice,” says Longoni.
Mishandled uses of AI have become
urban legend, from Target’s faux pas
of outing a teenage girl’s pregnancy, to
Amazon’s same-day shipping pricing
calculation inadvertently deprioritis-
ing certain demographics, making it
hard for consumers to make the con-
nection between AI and trust.
Cian thinks with more exposure
to eff ective, unbiased AI, consumer
views will change. But Dr Keith
Grimes, clinical AI and innovation
director at digital healthcare service
Babylon Health, believes it’s also
essential to help consumers under-
stand AI’s decision-making process,
especially in sensitive areas.
“People get concerned about this
‘black box’ phenomenon, the idea that
decisions get made, and they can't
work out why they're made, or they
can't challenge them. When you're
working in healthcare, you have to be
able to explain how automated deci-
sions are made,” he says.
“If we take care with the messag-
ing around how we use AI, we can
help reduce some of that anxiety and
people will feel more comfortable,”
Grimes concludes. It’s sound advice
for businesses across all sectors.
as fashion, cosmetics or garden-
ing, Envolve Tech’s AI performs
well in areas where people don’t
want to speak to a human, such
as an online condom retailer.
Meanwhile, the same AI on a med-
ical device retailer’s site has been
less successful.
“When shoppers need an exact
answer for a complicated situation,
humans still come out on top, at
least for now,” Smith notes.
Two more important distinguish-
ing factors between a human and AI
recommendation system are the vast
amounts of data AI can process and
being free of personal biases, she says.
“Even the best human customer ser-
vice agent can only possibly stay on top
of a fraction of the information AI sys-
tems can, which means human prod-
uct recommendations are always based
on a smaller dataset,” says Smith.
“A human agent will also bring
their own personal biases in. For
highly bespoke, artisanal purchases
this can be desirable, but for most
purchases it’s better to have a more
objective recommendation.”
If you need recommendations or advice, who are you likely to turn
to? New research suggests artifi cial intelligence can help, even in
cases where most of us normally prefer a human response
A
MaryLou Costa
People are more amenable to
AI in cases in which there’s a
human component
When do we trust AI to
suggest something to us,
and would we ever take a
robot’s recommendation
over a human’s?
Harvard Business Review 2020
67%
of test subjects picked an AI-
recommended hair product when
asked to focus on performance,
practicality and chemical composition
58%
picked a human-recommended
product, when asked to focus on
scent, indulgence, and spa-like vibe
R A C O N T E U R . N E T
A I F O R B U S I N E S S
15
14
re you prepared for artifi -
cial intelligence (AI) imple-
mentation? Do you know
what your accompanying data strat-
egy should be? If not, it is likely you
aren't alone. According to research
by Secondmind, 82 per cent of sup-
ply chain managers are frustrated
by AI systems and tools during the
coronavirus pandemic.
In its survey of 500-plus supply
chain planners and managers across
Europe and the United States, 37 per
cent cited a lack of reliable data to feed
into AI systems as a concern, at a time
when accuracy and speed of deci-
sion-making were of the essence.
They don't doubt AI's capabilities;
90 per cent agreed AI will help them
make better choices by 2025, but a
third raised another critical issue in
their leadership’s lack of understand-
ing of what is currently needed to
make faster, data-driven decisions.
So how do chief executives and
the C-suite approach solving this?
Listening to experts, most agree on
the main problems. These include
incomplete, dirty or duplicated
data, siloed data, inherent bias in
data programmed for AI models
and a lack of focus or knowledge at
board-level on what they hope AI
can, and will, achieve.
application programming interface
strategy to easily connect any appli-
cation, data source or device together
over an app network, where data can
fl ow freely.
"Siloed data stores and a lack of
connectivity between enterprise
applications severely restrict AI's
current ability to infl uence the dig-
ital ecosystem around it, rendering
it little more than a rather expensive
brain in a box,” he says.
"Businesses must build a cen-
tral nervous system that enables
AI to plug in and out of any data
source or capability that can pro-
vide or consume the intelligence
it creates. Point-to-point integra-
tions of the past will lead to atro-
phy in the AI-driven world, where
things can change in an instant
and even the near-future is uncer-
tain. Organisations must decou-
ple very complex systems and turn
their data stores and digital capa-
bilities into fl exible, discoverable
building blocks."
Adrian Tam, director of data
science
at
New
York-based
Synechron, offers a similar solu-
tion. "We have a term called 'data
lake'. It means to keep the data in
its natural format in an accessible
form. I think it doesn't matter if the
data is spread across servers and
across geographic locations as long
as we have a single, unified way to
access it. So, if there are data silos,
you just need to build an interface
to use it,” he says.
"Of course, this is easier said than
done because there are issues like
back-up, version control, system
resilience and availability. This is
another engineering problem, but
should not be part of the AI. It is a
bad engineering practice to blend two
problems into one unnecessarily."
Dr Neil Yager, co-founder and chief
scientist of Phrasee, addresses the
cleansing of data. "It is not widely
appreciated how much eff ort goes
into data cleaning and prepara-
tion,” he says. “A model built using
machine-learning is only as good as
the data it was trained on. Poor qual-
ity data leads to poor quality mod-
els. Unfortunately, pristine data sets
are rare in the wild; most datasets
are riddled with problems.
"The sets are often distributed
across
multiple
incompatible
sources and missing or incorrect
entries are common. A recent sur-
vey of data scientists concluded they
spend around 45 per cent of their
time on data preparation."
Combating all these challenges
means having the right skills widely
dispersed across an organisation to
achieve AI implementation. A part-
nership approach between a tra-
ditional data scientist alongside a
data engineer could be the answer,
according to Dr Greg Benson, chief
scientist at SnapLogic and pro-
fessor of computer science at the
University of San Francisco.
He says the former can "determine
how to apply models and derive train-
ing examples from existing data
sources" and the latter "understands
how to navigate existing IT data sys-
tems, understands regulatory compli-
ance considerations and ultimately
knows how to build data pipelines".
Elsewhere, research from Qlik
with
IDC showed
just
16 per
cent of knowledge workers glob-
ally are equipped to do AI and
machine-learning analysis. This fi g-
ure is predicted to rise to 25 per cent
over the next two years, with the
proportion of those with data liter-
acy skills increasing from 45 to 63
per cent. Two-thirds in another Qlik
study believed data literacy training
would make them more productive.
Adam Mayer, senior manager at
Qlik, says: "Many business leaders
are recognising that having these
capabilities siloed in business intel-
ligence teams will prevent them
from generating the greatest value
from their data."
Despite
all
the
complication
though, could the answer to AI
implementation and data strat-
egy be easier than we think? Jamie
Hutton, chief technology offi cer of
Quantexa, says: "There is usually
a simple test: if there is not enough
data for a human to make an accu-
rate decision, then neither will the
machine be able to do so."
No employee can make a good decision without all
the relevant information and neither can artifi cial
intelligence, making a solid data strategy the fi rst
step for any ambitious organisation
A
Jonathan Weinberg
Qlik 2020
EMPLOYEE DATA SKILLS NOT YET WHERE THEY NEED TO BE
A study of 1206 respondents across 10 countries and all sectors rate their current
in-house data skills.
Siloed data stores severely
restrict AI's ability to infl uence
the digital ecosystem around it,
rendering it little more than an
expensive brain in a box
Leila Seith Hassan, head of data
at the UK arm of global marketing
agency Digitas, believes it pays not to
treat AI as a buzzword. She says: "This
leads to a bit of naivety or even igno-
rance. Many don't really understand
what AI is or does and often apply a
futuristic and/or simplifi ed view.
"Too often, expectations of AI are
mandated without consideration of
what's feasible given an organisa-
tion's data maturity. AI requires an
organisation to have infrastructure,
process and people in place before
embarking on any serious project. If
you don't, you need to be prepared for
the time and cost that comes with get-
ting the organisation fi t for purpose.
"Ultimately, AI is making deci-
sions instead of humans. If you're
building your AI on bad data, it's
going to make bad decisions."
True AI implementation with the
right data strategy requires invest-
ment, time, and the best and most
experienced people. Get it right
and the positives are clear, with
increased profi tability, productivity
and reduced fraud among them.
How to
implement AI
successfully
D AT A S T R AT E G Y
Get it wrong, though, and things
can be very diff erent, especially
if the data used fails to represent
society or enforces existing biases.
Trust and consent is also crucial to
the process.
Dr Alan Bourne, chartered occupa-
tional psychologist and founder of
Sova Assessment, explains: "When
embedding AI into any business sys-
tem, it is essential that a real person is
placed front and centre of the process,
so humanity, laws, regulations and
ethics are considered with as much
importance as technological capabil-
ities. The opportunities to do this are
vast, whether it be using an internal
human resource, an advisory board
or using AI to audit other forms of AI
being applied to the business."
Data-hungry
algorithms must
also be continually tested says Alix
Melchy, vice president of AI at Jumio.
"Another process that business must
implement in their AI practices is a
pilot testing phase, to ensure the algo-
rithm is working as expected and to
better understand why an algorithm
is making a certain decision. By run-
ning a test in the early stages, and
before the algorithm is put into the
real-world scenario, feasibility, dura-
tion, cost and adverse events are all
assessed,” he says.
Where the data comes from and
how clean it is will be paramount
in AI implementation. Historic data
silos are often still needed for rea-
sons of privacy and security, but this
can cause problems, while the lack
of a connected cloud solution serv-
ing all parts of the business can be a
huge barrier too.
Paul Crerand, fi eld chief technology
offi cer for Europe, Middle East and
Africa at MuleSoft, recommends an
45%
16%
25%
56%
Proportion of knowledge
workers trained and
equipped to do standard
business data analysis
in two years
Proportion of knowledge workers
trained and equipped to do standard
business data analysis now
Proportion of knowledge
workers trained and equipped
to do AI/ML analysis now
Proportion of knowledge workers
trained and equipped to do AI/ML
analysis in two years
AI requires an
organisation to have
infrastructure, process
and people in place
before embarking on
any serious project