Appen

Highlighting best practice as a representative in The Parliamentary Review

The ability to listen and learn from one another has always been vital in parliament, in business and in most aspects of daily life. But at this particular moment in time, as national and global events continue to reiterate, it is uncommonly crucial that we forge new channels of communication and reinforce existing ones. The following article from Appen is an attempt to do just that. We would welcome your thoughts on this or any other Parliamentary Review article.

www.appen.com

33APPEN |
BEST PRACTICE REPRESENTATIVE 2019
VP Language Resources UK
Sarah Lowe
Supervisory and
management team, Exeter
Many worry about the potential effects of AI on the
economy and society. Appen, however, feels that
the focus should be on the benefits of AI, as firms
adopting this technology can revolutionise the way they do
business. Machine learning – an important component of AI – is
used by leading companies around the world to create more
personalised online experiences, streamline operations and
improve products. Successfully using machine learning, Appen
says, requires more than good algorithms; it needs data both
high in quality and great in number. This is what differentiates
Appen, which manages large-scale data collection and
annotation projects for many industries, operating in over 180
languages and dialects, working with a global “crowd” of over
one million people. Sarah Lowe, Vice President of Language
Resources UK, says more about this vast undertaking.
Interpreting language
Here at Appen, our business is all about data. From providing data that is used
to improve AI and machine learning-based technologies, to producing data for
evidential, official record transcripts, Appen works with major international and UK
clients who require both types of data sets.
In practice, the management of data might mean turning an audio recording of an
investigative or police witness interview into a readable text transcript – something
FACTS ABOUT
APPEN
»VP Language Resources UK:
Sarah Lowe
»Established in 1996
»UK operations based in Exeter,
Devon
»Service: Provision of data
sets for artificial intelligence,
machine learning and official
record transcripts
»No. of employees: Over 50 in
the UK, over 500 globally
»ISO 27001, ISO 9001 and
Cyber Essentials accreditation
Appen
THE PARLIAMENTARY REVIEW
Highlighting best practice
34 | APPEN
particularly useful in a legal or formal
hearing context. It’s for this reason
that we hold many government and
public sector contracts, including
with the Crown Prosecution Service,
police forces and local authorities.
This removes a significant amount
of unnecessary labour from their
processes, meaning that justice can
be pursued efficiently and cost-
effectively through a professionally
managed, outsourced and secure
transcriptionsolution.
The characteristics that make us a
preferred provider for many branches
of the British justice system include:
more than 25 years of experience;
thorough screening, security vetting
and supervision of resources; and our
ISO 27001, Cyber Essentials and ISO
9001 certifications, which guarantee
highly secure, quality transcription.
In terms of Appen’s AI services, our
strength lies in our ability to handle
large-scale data collection and
annotation projects with quick ramp-
up times to produce high-quality data
for our clients. Our services underpin
a variety of machine learning-based
technologies, including chatbots and
virtual assistants, in-car infotainment
systems, search engines, social media
platforms and more.
Support for machine learning
Our rapid growth has been fuelled
by investments in machine learning
programs that have been made by
companies across all industries to
gain competitive advantage. The
process requires close collaboration
with our clients, ensuring that their
specific goals and projects are fully
understood. We have partnered with
leading technology companies across
the board, from search engines to
social media platforms. As with our
transcription and annotation services,
we once again distinguish ourselves
in this area by having a strong and
provable commitment to data security.
Data with a human touch
It’s worth emphasising once more
that the data we provide is not
indiscriminate – for machine learning is
only as good as the data that is used to
train it. Our business model hinges on
custom engagements to understand
the problem the clients are trying to
solve with machine learning, their
data needs and their quality goals. We
then design a data program specific
to meeting their objectives, and our
project managers work hand in hand
Appen Exeter team
Data with a human
touch
Our strength
lies in our ability
to handle large-
scale data
collection and
annotation
projects with
quick ramp-up
times
35APPEN |
BEST PRACTICE REPRESENTATIVE 2019
with clients to ensure that quality
metrics are met. Our approach has
allowed us to build strong relationships
with our clients, who continue to
renew with us year after year.
Meeting market demands
In so restless and effervescent an
industry as ours, one of the main
difficulties is simply keeping up with
the rapid pace of development. As
a recognisable company in this area,
we feel it’s important that we not
only stay ahead of the curve, but also
remain able to cope with the high
levels of demand. We pride ourselves
on providing “data with a human
touch”, so it’s doubly necessary that
the data we provide is appropriate for
the situation at hand. That is, we do
not simply churn things out at scale;
we instead offer a truly customised
solution to our clients. This is a more
difficult undertaking, but it’s an
aspect of our service that our clients
deeplyappreciate.
Another challenge – albeit one that
we recognise as wholly necessary –
is the regulatory and data security
environment through which we must
navigate. This is especially the case
when we want to win government
contracts or work for clients who
have highly confidential, cutting-edge
technology applications. These all
require the utmost security and trust.
To give an example of how this plays
out, we have to organise our company
in such a way that only those who
are sufficiently vetted can handle
certain types of data. To meet all these
prescribed thresholds is a challenging
and ongoing task, but excelling in
this area has won us a much more
esteemed position in the market.
Thoughts for the future
Servicing and supporting the ever-
expanding needs of AI is an exciting
place to be for Appen. The future for
us doesn’t have the distant quality
it might have for other businesses,
because we ourselves are part of
the process of defining the future.
For us, then, the future consists of
perpetuating what we are already
doing, continually seeking to improve
and refine our client services. It also
means growing our “crowd” of
specialist data workers and keeping
abreast of the competition. It’s for
these reasons among many others that
we are optimistic as we look ahead.
Our project
managers
work hand in
hand with
clients to
ensure that
quality metrics
are met
Secure transcription area

www.appen.com

This article was sponsored by Appen. The Parliamentary Review is wholly funded by the representatives who write for it.