AI, RECRUITING,
SELECTION AND TRIANGLES
John Wentworth, the founder of Wentworth Recruiting, and I
were having lunch at Beach City Grill. Obviously
a regular at this local San Pedro establishment, John didn’t even pick up the
menu. While I was focused on what would best feed my midday growling stomach, he
was having one of his nutty professor days. “It’s all about two triangles,” he
muttered.
“What are you talking about?” I asked, still thinking about
my lunch.
“Oh, sorry,” he said. “I’ve been thinking about how
recruiting automation has missed the mark.”
He started scribbling on a piece of paper. “Look at
this. Most of what’s automated with applicant tracking systems is
administrative: when did the candidate apply, contact info, first screening,
salary expectations, etc. That’s all important for managing the function and
for compliance, but none of it has to do with delivering the very best possible
candidates to the hiring manager. The actions that deliver the best
possible candidates are either not automated or automated inadequately. They got lost in the rush to make the
recruiters’ lives easier.
“They are missing the most important part of recruiting.”
“Which is delivering the very best candidates possible?” I
asked.
“Yes. For many recruiters, their secret goal is to put
butts in the seats, not to deliver the very best candidates possible. It’s
how they get measured, sadly. It’s how
they get raises.”
“The employment of a sub-par workforce starts with the
recruiter who lets adequate but marginal candidates in the front door and lets
them go forward to the hiring manager.”
“I get that,” I said. “Not knowing how to select the
best candidates, hiring managers pick the candidates they like, or the best of
what’s provided to them even if what's provided to them is lousy.”
“It’s on the recruiters,” John said again, repeating
himself. He was getting wound up. “It’s
their obligation to put forward only the best possible candidates so the hiring
manager can hire any of the pool and still get a great employee. Recruiters rarely can do it on their
own. They need help. Artificial Intelligence is supposed to save
them. But Artificial Intelligence can
only help so much.”
“It shows the limits of Artificial Intelligence. AI does a great job of hitting the problem
with a 2x4, but good selection needs a scalpel, not a big piece of wood.”
“Help me understand.”
“AI can look at information and patterns, but only makes
sense of some patterns, not others. Let’s say that AI sees that a candidate
had the right titles and that the titles have grown over the years through
promotions. That means one thing if her employers were all about the same
size as your company is. She probably has the skills for your job.
However, it means something else if her early career employers were
bigger than your company and had gotten smaller than your company over the
years. Has she been out of her depth on her jobs, needing to go to
successively smaller companies with less complex jobs? Is she too light
for your open job? Probably. Can AI recognize that? Probably
not today. Maybe in the future when it
can tap into a directory of every company in the world and is told to compare
the size of the employer to the companies on the resume, and how to interpret
those data.
Here’s another example; what if your candidate inherited
some money and was only interested in working for a few more years, but had not
mentioned that? Do you think AI will be able to pick that up? A personality
instrument that measured ambition could give you a clue, but AI probably would
not. Do you think the candidate’s ‘I want to cruise to retirement’ attitude
would be dysfunctional in an aggressive, get-it-done employer?
Probably. Would it lead to termination? Probably. A really good recruiter could catch
that. I’m not sure how a machine would.”
“Wow. That’s a dismal picture,” I responded glumly.
“Not so dismal. You can figure out your job’s key
success factors, and to build a measurement scales for them. We need to
find behaviors that correlates with how much someone wants challenge, for
instance, interview for that behavior and rate it on a 1-5 scale. We need
to know how much the job needs and then how much the candidate has and then compare
them on the scale.”
“What would automation do?”
“Running these types of comparisons creates a lot of data
that can be difficult to understand and can add to the ‘fog of recruiting.’
Automation can make complex data into simple and vivid displays that show
how candidates compare to the requirements and to each other. Automation
helps to lift the fog of recruiting, making the weak and strong candidates
obvious.”
“You talked before about AI hitting the problem with a 2x4
vs. a scalpel. Is this the scalpel part?”
“Yes it is,” John said.
“And no machine does this,” I said, not sure if I was
asking a question or making a statement.
“They can but most just record, count and report or
recognize gross patterns or look or key words. But I just ran into
one that does add real value. It expands
your pool of good candidates by automating the employee referral process.
This is valuable because it enriches the supply of qualified candidates.
It reduces the dependence on recruiters managing the employee referral program. Recruiters forget stuff; the machine does
not. It sets up innovative and personalized rewards and lets the employee
pick; the recruiter does not need to manage that process. It finds
potential candidates through employees voluntarily sharing access to their
social media, discovers the candidates’ contact information, invites the
employee to select and invite candidates to apply, and then processes the
candidates throughout their journey. All this frees the recruiter up to
talk to people, to do the scalpel part. It’s
the best use of automation and AI I’ve seen in the talent space. It’s
called Boon.”
We were finished with lunch. And there was a silence.
I said, “That was a lot.”
“Sorry,” John said, chuckling. And then he pulled out
his usual quiz of what was just said to see if I was listening. “Tell me what you think the point is?”
“I think the point is that if you want to hire the very best
people available to you, you can’t cut corners. There is no silver
bullet. Automation can help, but finding the best employees, selling them
on your job and selecting the right ones is a ‘contact sport’ and takes skilled
people applying themselves in a disciplined way. Machines can’t do that
yet.”
“All true. Once we start acting like hiring the very
best people possible is the real goal, the methodologies will emerge and the
automation will follow. They actually
exist now. They just have not become
widespread yet.”
“Who did you say was buying lunch?” he asked with a smile.
I reached for my
credit card.
John Wentworth has been in HR since 1972, a recruiter
since 1979 and run his own company since 1984. John is a Vietnam War combat
veteran and The Wentworth Company is a State of California Certified Small
Business & DVBE.
The Wentworth Company has helped 550+ employers hire over 20,000 new
employees. The firm consults regarding talent, designs talent systems and
offers high volume, RPO and single search services. 88% of their hires
exceed the hiring manager's’ definition of qualified. He was interviewed by an
imaginary companion.
No comments:
Post a Comment