Software in HR. It is an introduction to
both the technology and the market. If
you’re looking to get a running start in
AI (or a quick look at the state-of-the-art tech), this is where to begin. Expect
a simplified view of the underlying
technologies, a deep look at the shape
of the market and a grounding in the
most important ethical issues. The
session will investigate the question of
whether AI techniques that are useful
for managing things should also be
applied to managing people.
Questions to Ask When Evaluating
HR AI Products— This will be a
conversation with a single customer
about the questions asked during the
initial purchase of intelligent software.
At the heart of this conversation is
the need to cut through the noise,
evaluate security issues, understand
the total cost of ownership and
explore the quality of data sources. A
big part of the value delivered by AI
systems depends on the vendor’s data
model and sources. The discussion
will revolve around comparing and
contrasting different approaches.
HR Tech AI Best-in-Class Showcase—
Four companies, picked as the result
of HRExaminer market research, will
deliver disciplined, well-oiled, seven-minute presentations about their
business models, intelligent services,
implementation processes, data models
and customer support. The companies
represent how emerging industry
leaders can solve different aspects of
the HR-tech problem set.
Is Tech Eliminating or Amplifying
Bias?—Heather Bussing of HRExaminer
and Kate Bischoff of tHRiveLaw and
Consulting are the two brightest legal
minds on the topic of technology and
bias in the workplace. They are also
steeped in technical expertise. In
this conversation, they will illuminate
approaches that effectively address
workplace bias and those that don’t.
Many AI vendors claim to be able
to eliminate bias, but Bussing and
Bischoff will help the audience see the
risk in that approach, how to mitigate it
and how to define your company’s needs
when purchasing AI technologies.
At HR Tech, you can expect to be
deluged with multiple (often conflicting)
points of view about what AI is and
how to harness it. These five sessions
are intended to give you ways to think
about the issues, an understanding of
the research and a look at the best the
industry has to offer.
Surveying the Landscape
Expect the topics of AI, analytics,
predictive tools and data to dominate
the discussion and the vendor exhibit
hall at the conference. Here are some
of the themes you will encounter:
Although there is a growing sense that
cannot simply be jettisoned, their utility
remains a vibrant topic of conversation.
The offerings in this category
include machine-assisted coaching
of supervisors to help them develop
a team; deep interaction among
employee-development, learning-content and performance-management
systems; hyper-scheduled check-ins
coupled with pulse surveys; and
There is a clear connection among
engagement scores, company morale
and performance management. Several
of the vendors use AI to blur those
distinctions in the pursuit of performance
improvement for the overall organization.
Improvements—Recent advances in
recruiting technology have included
an increased emphasis on the quality
of a candidate’s interaction with a
prospective employer. Many of the
same variables apply to questions of
internal mobility. There are several
solutions coming to the market that use
statistical techniques to help external
and internal candidates navigate their
careers and the job-hunting process.
Offerings include programmatic control
of advertising for budget and quality;
for candidates; career discovery that
shows the historical linkages among
specific jobs; and skills analysis that
allows recruiters to understand job
requirements more broadly and helps
candidates see additional opportunities.
are layers upon layers of inefficiency in
recruiting processes. Solutions in this
category range from comprehensive
toolkits as part of an overall
recruiting platform to individual point
solutions. There is a resurgence of
emphasis on matching because of the
improvements gained from natural-language processing. Multiple vendors
promise to predict the likelihood
that a candidate will be open to new
job opportunities, and there are a
few integrated interview-scheduling
tools that should save lots of time and
money. Candidate-discovery tools help
users sift through massive amounts of
Changing Business Models—The
explosion in AI tools is driven by a
radical increase in the numbers of
things we measure and the data we
gain from those measurements. Old
views of what belongs in which HR
silo under which circumstances are
under assault. Crowd-sourced labor
markets with a single invoice are
being delivered in a variety of forms.
Training can now be purchased as a
Netflix-style subscription, and it’s now
possible to buy tools that will illuminate
the real organization (not the one in
the organizational chart). There is
even a company that will predict (with
accuracy) the likelihood that a given
team will meet its objectives.
Suite Providers—There are 12 to 15
major HR Tech suite providers making
the case that their way of integrating
data is superior in the aggregate to
any single-point solution. Each of the
“usual suspects” brings a unique point
of view to the arena. For instance,
Cornerstone focuses on learning
first while Ultimate Software builds
employee-sentiment analysis into the
core of its offering. Workday is focused
on data integration and planning at its
core, while Ceridian’s efforts flow
from its workforce data, and Kronos
uses historical data to make schedule
recommendations that are tied to KPIs.
The Disruption of HR Itself—If you
look carefully, you’ll see products and
services that bypass traditional ways
of thinking: performance-management
tools that act like learning-management
systems; scheduling tools that become
full-scale workforce-planning tools and
others that improve the supervisor’s
plan; and behavioral analytics that allow
operations professionals to make better
And that’s just for starters.
The advent of intelligent software
also creates an array of ethical issues.
As machines take a greater role in
understanding and managing people,
there will be a constant clarification
of the different roles people and
computers play. Expect to see some
unintended consequences while we
explore the difference between a
recommendation and a decision. The
difference, of course, is that a decision
involves liability, and a recommendation
doesn’t. The underlying question as to
when a recommendation becomes a
decision will occupy lots of conversation
in the coming years.
This year’s HR Tech Conference
promises to be a gateway for learning
about the future of the profession. As
sci-fi author Bruce Sterling says, “The
future is already here. It’s just not
evenly distributed.” You can get to see
the future firsthand at this conference.
John Sumser is the principal analyst
at HRExaminer. He researches the
workplace impacts of data, analytics and
AI and the associated ethical issues. He
works with vendors and HR departments
to identify problems, define solutions and
clarify the narrative. HRExaminer’s
Second Annual Index of Intelligent
Software in HR will publish Sept. 1,
2018. The HR Technology Conference &
Exposition® will be held Sept. 14 through
Sept. 16 in Las Vegas: http://www.
4How do you define the difference between a
recommendation and a
decision? Does your
company ever have liability?
7What are our options if we want to switch vendors or
use different data models?
5What do we do if the system is producing results that are not
productive for our company?
1Describe the basic data model. How do I know that it accurately reflects today’s circumstances?
What do you do to test its real-world validity?
2How do you handle security? Is the system GDPR-compliant?
How is the actual data protected?
3How would a malicious person hack the data model? How do you protect
the integrity of core algorithms?
John Sumser’s Guide to
Evaluating a Vendor
8 Can you reproduce the results of your data model(s), and how do you
maintain them when a change is required?
6When the machine gives a wrong answer, how
do we recover? How do we
prevent it in the future?