Digital Innovation
should be automated by the end of next
“For recruiting teams, it will mean they get to do more of year. We take it client by client to make
what they are good at; their USP will be more pronounced” sure they’re happy.
Does it save money?
There’s an initial cost in setting up the
infrastructure and working with your
IT team to implement it; but there are
cost savings and efficiencies, which are
part of our business case in driving
this forward. However, in terms of the
client experience, we’re finding that
the biggest change is that sleekness of
process end-to-end.
How long have you been
using this tech?
We’ve had our toe dipped in the
water for around 18 months. That’s
ramped up to a wide-ranging, back-
end enterprise project.
I’d say we’re an early adopter.
I do a lot of talking to clients and
consultancies; they have a good view
of where this might be useful, but they
haven’t done it yet. There’s a lead time
of about 12 months to learn how to use
it and then you speed up enormously
and it becomes a day-to-day thing.
The ‘mother and father’ of the robots
are myself and Jolanta Gantkowska
who heads our Polish office; she’s very
interested in how we can use robots to
provide consumer-quality services. She
drove us to blueprint the technologies,
then I worked to pull those together.
Could you introduce us to
your robot offspring?
DORIS was our first robot; she was
really good at organising documents and
putting them into folders. She worked
with 70,000 documents in 48 hours,
which would have taken two months
with 10 people. ISAAC is an interview-
scheduling robot; he sends emails, tries
to find dates for when candidates are
available and posts details into their
diaries. Then there’s RALPH, who
tracks the activities of all the people in
our shared service centres and reports,
in real time, on what they’re doing.
We also have STAN, who logs into a
number of external systems, downloads
reports on a daily basis and consolidates
those. HANA is in development: she’s a
help-desk and chat robot, able to take
requests from people externally and
trigger services.
We actually have 20 different STANs,
so he’s a whole family now and the
ISAAC family is growing: we will have
one ISAAC per client. There are lots of
other robots in development, including
an onboarding robot in design.
We have a wide range of core
processes we’re looking to automate
entirely; scheduling is one: we arrange
150,000 interviews a year, so all of that
Meanwhile, I’m trying to use
‘contagious implementation’ to make
robots part of a team. I envisage sending
out a robot to a team that delivers
reporting across the whole organisation
so they can learn to incorporate it into
their day-to-day work. That team
will then share their knowledge with
other teams.
How do you identify tasks
to automate?
Initially, people pitched in with
suggestions. We looked for those that
best fit a robot and gave us the biggest
opportunity to enhance the experience
and provide efficiencies. We started by
gut feel and built little business cases for
internal investment. Then we started
getting better at identifying which robots
you can create quickly and achieve the
most. We’ve put a governance model
around that now.
At the beginning, people would say
“could you write a robot to do this?”
and you’d look at the process and it
was really complicated; it would take
somebody three months to write a robot
programme that saved 10 minutes a
day. But there were other processes
where someone could write the robot
in half-an-hour to do a piece of work
that could run in the background and
saved somebody three-to-four hours a
day, so they could focus on other things.
Is this tech a threat to
human employees?
There’s a war for talent outside in the
marketplace; we don’t have enough
people to meet our gross needs, and
this is how we’re going to fill the gap.
When laptops and email were first
implemented, everyone immediately
thought “we’ll need less people because
we’ll be so much more efficient than
previously”, but it actually just meant
people could do more.
Initially we’ll automate back-end
processes, and everyone will start
to use this sort of thing. Then I think
the sidekick – the ‘personal assistant’
Issue 1 - 2017
11