ARTIFICIAL
INTELLIGENCE
There is a broad array of
areas within logistics where
AI is coming into play, from
back offi ce automation and
predictive operations to
intelligent logistics assets,
revenue optimisation and
new customer experience
models. A survey of supply
chain analytics published in
January by APQC indicates
that industry executives look
to AI primarily for supply
chain optimisation and cost
reduction at this point, less for
risk reduction and improving
productivity.
Gesing sees it in four
major sectors of the
logistics fi eld – predictive
analytics, voice recognition,
visual AI applications and
autonomous units.
Predictive analytics is
head and shoulders above the
rest. APQC found that 60%
of respondents were using
machine learning in this arena.
The accuracy of delivery
predictions is a prominent
concern for many players.
Drawing in external data like
weather and traffi c reports, AI
can be used to predict delivery
times and rates. Along the
same lines, airports can use
AI to make projections about
peak times and adequate
staffi ng levels for these.
Demand anticipation and asset
allocation is another favourite
on the airline side, along with
route optimisation.
Security is yet another area
of interest. AI can be used
to identify deviations from
payment patterns that indicate
potential fraud. Messages from
fi nancial fi rms about possible
credit card fraud are usually
triggered by patterns detected
through machine learning,
remarks Gesing.
A picking cell in Beringe. These robotic arms can
process up to 600 packages an hour
A capacity management tool
Air France/KLM/Martinair Cargo has concentrated to a large degree
on capacity management, using AI in aspects like allotment and
contract processing. The cargo division is applying predictive and
optimisation techniques to examine client behaviour in order to
determine which tenders for capacity are preferable over others.
As it rolls out its fi rst AI application, Air Canada is setting
its sights on spot pricing. The objective is to determine the best
possible price that can drive business to the airline’s domain, says
Strauss.
“I think we leave a lot of cargo behind because we haven’t
got the technology to get it right. I lose business because I was
too high, too low or too slow,” he refl ects. “We can use AI to
tell us in some lanes at what level we should be.” He reckons
that the potential savings are going to be less than from the fi rst
implementation, but still considerable.
The potential cost savings that AI can generate overall are
massive. According to one estimate, about US$2bn is wasted every
year in air cargo because of missed revenue, aircraft ineffi ciencies,
container costs and other factors.
Whereas predictive analytics have garnered a lot of attention,
voice recognition has not really taken off in this industry, observes
Gesing. He reckons that adoption would accelerate if it were to
become the new paradigm in computer interaction – but at this
point, this is not on the horizon.
The hype and
fear around it... is
ending
Ben Gesing, Senior Innovation
Manager, Trend Research, DHL
34 April 2020 www.airlogisticsinternational.com
/www.airlogisticsinternational.com