People & Strategy Summer 2017 Vol. 40 Issue 3 - 44
profitability and HR system structures etc.), continue to
evolve their capabilities Further, this has enabled the implementation of outcome-direction thinking.
Outcome-direction has already produced results. "Hitachi
AI Technology/H" (AT/H), developed and provided by a
A key trait of AI is that it is borne from
diversity both on the human and
group of experts, has applied this outcome-direction methodology to a wide range of industries and organizational
settings including banking, securities, commerce, distribution, water, railroads, business, HR, and others (14 in all).
Proofs of concepts has been successful in 57 separate cases in
these various domains. This technology is demonstrating that
it can revolutionize business and the key characteristic of this
technology its ability to improve the outcomes in a variety of
fields without changing the underlying software that recorded and captured the data.
Misconception 2: AI Is a New
Technology or Machine
AI is neither a new technology nor a machine. AI is a realization of the above "outcome-direction" and the software tools
that allow it to be realized.
A Theoretical Method for Problem Solving
AI's theoretical method is simple and people lacking a technical background can understand its essential form. It can be
broken down into three simple steps:
First, determine numerical values that indicate the target
to be improved and the range of that improvement. This is
known as the outcome.
Second, collect past data related to the outcome, and seek
to understand what conditions and behaviors increase the
outcome. Here, we are trying to understand what trends we
can see with this data. For example, "taking this action [defined as anything seen to be connected to the outcome] when
discussing business with this client increases the outcome
of the client placing an order by 20 percent." Many factors
may influence an outcome. Thereafter, measure the weight
of each of these factors' ultimate effect. Changes in past outcomes, depending on the prevalence of a given factor, allow
the AIs to use the relative weighting of each factor to create a
prediction (evaluation) formula.
Third, apply the formula created from the weighted
factors to business or management decisions. Specifically,
prepare multiple options at the time of the business or management decision, and anticipate the outcome of each option
estimated based on the evaluation formula mentioned above.
This is how option strength is evaluated. Such a process
allows for a flexible determination to be made based on the
accumulated past data. That is what AI does.
PEOPLE + STRATEGY
This is the most direct theoretical method for the
systematic learning of past results. Software AI includes
software that employs this theoretical method, such as deep
This theoretical method is particularly important in the
HR field because HR has access and possession of most of an
organization's data sets and data fields related to behavior
and people activity. That is why this method is so effective
when applied to data decisions.
Historically, HR as a department within a company gave
the impression of being far removed from the digital world
and front line business decisions. However, the introduction
of AI completely changes that. Since HR has conventionally
been a field that relies on experience, to make decisions in
complex and ' judgement based' circumstances, there is a
large, white space ripe for development between the confluence of HR, AI, and business. The application of AI on HR
will make a huge impact to the future of organizations.
Misconception 3: AI Requires Large Amounts of
Data to Be Useful
If data is not applied, AI is nothing more than a box. Therefore, there are many arguments that emphasize the importance of the data volume. Yet, more important is the setting
of the problem that we must solve. Specifically, selecting
target outcomes, defining potential actions that create those
outcomes, and defining probable conditions that affect them
is critical. Only human beings can do this.
Furthermore, along with the target, outcomes, actions,
and conditions are three types of data that constitute required items. Together, these three items bring about the
power of AI.
Outcomes of HR Research: Achieving Happiness
Outcomes are the most important items in the problem settings. If you set a simple, non-valuable outcome, only simple,
non-valuable outcomes will emerge. Accordingly, if you assign
a poorly defined abstract outcome, the achievement of the
outcome will not be obtained.
So what should we set as the most significant outcome?
What is the best outcome for society? We'd argue happiness
is. If you set happiness as the outcome, you can achieve this
by looking at a wide variety of data to achieve happiness.
Many may think that the happiness or well-being of employees as an outcome may not have an impact on business.
In fact, there have been advances in measuring and analyzing human behavior using wearable sensors in addition to a
series of recent breakthrough findings. By using an accelerometer sensor and gathering the associated data on the
employee during activity, we can now measure and quantify
human happiness. This human activity which is also recorded even while sitting (supposedly at rest), is composed of a
constant series of small, unconscious patterns between rest
and movement transitions. It has nothing to do with the
amount of what is considered obvious physical movement.
This unconscious movement can be measured and it is this
measurement that can determine human happiness.
Moreover, the key to the HR field is that people with high