Page 4 - Logistics News April 2018
P. 4
Thought Leadership
Paving the way
for AI in the warehouse
By Luke Waltz, Supply Chain Quarterly
Recent developments in artifi cial intelligence (AI) are set to revolutionise the way warehouses
operate. But before companies jump into implementations, they must make sure they have the
data and talent they need.
STAYING ABREAST of changes in supply chain Choosing a use case
technology has become almost a full-time As you consider opportunities to apply AI
job. From robotics and automation to data in the supply chain, it may be tempting to
analytics and the industrial Internet of things start with the technology and seek out an
(IoT), new technologies are emerging that have application. However, more useful applications
the potential to further improve how goods are are likely to emerge if you evaluate the
shipped, handled, stored and delivered. With business drivers that represent the greatest
all of these technologies competing for our challenges or opportunities for your company,
attention, it can be diffi cult to know where to and then apply an appropriate understanding
focus. of AI technology’s capabilities to those issues.
One new technology that does deserve In relation to the warehouse, AI applications
a close look is artifi cial intelligence (AI). In should be guided by the key performance
the simplest terms, AI is the development of indicators (KPIs) a particular organisation is
computer systems that can perform tasks that trying to optimise, such as order accuracy,
normally require human intelligence such as safety, productivity, fulfi lment time, facility
visual perception, speech recognition, decision damage or inventory accuracy. Warehouses
making and language translation. AI has been typically already have a wealth of data that is
around since 1956, but humans typically have related to their KPIs and could be used by an
had to explicitly programme intelligence into AI application to automate tasks or decisions.
computers. However, this data typically is in a form that
One type of AI called machine learning is not conducive to using AI techniques and
explores ways to enable computer programmes it often is spread across various warehouse
to improve their output based on learning systems. As a result, many AI applications will
from data inputs. These programmes can be likely require information to be aggregated
embedded in machines or they can operate across various information systems in the
on servers or in the cloud. Large technology warehouse before it can be used.
companies such as Amazon, Google, Facebook, The following examples illustrate the
Microsoft and others are already incorporating potential for AI in the warehouse. Each of them
machine learning into their off erings, creating is focused on a KPI: productivity, equipment
more intuitive web searches, better image and utilisation or effi ciency. While the examples
voice recognition, and smarter devices. may not be applicable to every warehouse,
For many supply chain executives, AI – and they do show how companies can take
particularly machine learning – is an important available data and fi t that data into a form in
technology to consider because it allows tasks which machine-learning techniques can be
to be automated. Organisations that begin applied.
today to develop AI strategies that are relevant Productivity. When it comes to picking
to the supply chain will be positioned to orders, all warehouses experience a range of
increase productivity, speed and effi ciency as productivity, from their highest performing
the technology matures. order pickers to their average performers.
6 April 2018 | Logistics News