December 13, 2023 • 25 mins
Are you curious about how AI can create a more human experience in the workplace? Look no further. Join William Tincup and Jack Berkowitz as they delve into the incredible ways ADP harnesses AI to enhance the employee experience. With Jack's expertise in data governance, privacy, and security, this conversation unveils how AI is revolutionizing HR practices and empowering managers to build better relationships with their teams.
But, what truly sets ADP apart is their focus on the human side of AI. Jack reveals how AI-powered nudges are transforming the employee-manager relationship. By analyzing data and providing personalized recommendations, AI empowers managers to support their employees' growth, encourage better work-life balance, and address potential issues proactively. It's a game-changer, ensuring happier employees and more effective management.
With a keen understanding of HR complexities, Jack highlights that AI is not meant to replace HR professionals, but to empower them. By automating transactional tasks, AI gives HR leaders more time to focus on strategic initiatives and the human aspects of their roles. Jack emphasizes that AI should be seamless, trusted, and built with a human touch in order to truly enhance the employee experience and foster positive workplace relationships.
As the conversation unfolds, Jack's insights and vision for the future of AI in HR make it evident that ADP is at the forefront of revolutionizing how businesses approach employee experience. Together with William, he offers a glimpse into how AI is redefining the way we work, creating a more human-centric workplace where employees, managers, and organizations thrive.
Listen & Subscribe on your favorite platform
Apple | Spotify | Google | Amazon
Visit us at RecruitingDaily for all of your recruiting, sourcing, and HR content.
Follow on Twitter @RecruitingDaily
Attend one of our #HRTX Events
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
William Tincup (00:32):
This is William Tincup, and you're listening to
the Recruiting Daily podcast.
Today we have Jack on from ADP,and our topic today is how ADP
is using AI to create a morehuman experience.
So let's just jump right intoit.
Jack, would you, uh, introduceyourself and what you do for
ADP?
Jack Berkowitz (00:49):
Yeah, my name's Jack Berkowitz.
I'm the Chief Data Officer hereat ADP.
Um, sort of got two jobs.
One, um, help build products,uh, whether it's people
analytics or reporting, uh, orAI products.
And then the second thing is I'mthe Corporate Chief Data
Officer.
So everything about datagovernance and data privacy and
(01:09):
data security and all thosetypes of things.
Oh, that's
William Tincup (01:13):
cool.
That's the harder part of thejob.
I'm assuming the compliance and,uh, and especially just meeting
all the, I mean, you've got,you've just, you just mentioned
privacy and security.
Those two things are massive.
Jack Berkowitz (01:28):
Yeah, you know, and I think, but they're super
important.
Oh, yeah, critical.
And it helps you build betterproducts if you do the proper
security, if you do the properprivacy.
If you think about ethics upfront, particularly in our
space, right, in HR, then youbuild better products.
And so it's a nice loop.
It does feed each other.
I
William Tincup (01:46):
can see that.
I can see that.
So, you know, I've probably beenin a lab working with AI for
years now, but, um, let's talk alittle bit about, again, how
you're using it to create morehuman experience.
Yeah.
Jack Berkowitz (02:00):
You know, we've had a number of different things
out there.
We have a skills graph all basedon AI technologies.
We have counter relevancyalgorithms, all this type of
stuff.
Right.
Um, which is super interestingfor practitioners about pooling
information, but where we'regoing next is really this notion
about how do I nudge?
How do I nudge?
Whether it's a practitioner sothey can do, you know, have more
(02:23):
automated payroll, or how can Inudge a manager to say, you
know, hey, for example, you'vegot an employee who's not been
taking vacation.
Maybe you should help them finda way to get a vacation before
they, Oh, that's cool.
Um, that type of thing.
Yeah.
William Tincup (02:38):
And so it's kind of the, uh, it looks at the
data.
I say it, uh, but basically it'slooking at data to then find
where the business can getbetter.
But instead of going direct tothat, uh, employee per se, you
nudge the manager so that themanager can be a good manager.
Jack Berkowitz (02:58):
That's right.
That's right.
Right.
Um.
At the end of the day, we knowfrom all of our data, we know
that, yeah, people leave forsalary sometimes and they'll
leave for, for careeropportunities, but normally
they're leaving because of theirmanager or their manager's
manager.
Right.
And so if we can create a betterrelationship or if we can help
support, better that, supportthe relationship between the
(03:22):
manager and the employee,everybody benefits.
Right.
You get happy employees, you geta better, uh, management
experience, you get betterrelationships.
Better results for your company.
Yeah,
William Tincup (03:31):
what I love about it is it's, it does fall
in line with employeeexperience.
It's a different take onemployee experience, actually,
because most of the EX folks isfocused on the employee, not the
manager helping the employee.
Jack Berkowitz (03:45):
Yeah, and we're not, we're not neglecting the
employee.
No, no, no, no, nudges actuallywith the employees.
Right.
So, you know, we started a fewyears ago.
Hey, by the way.
You're not saving enough forretirement.
Or, oh, by the way, there arethese types of benefits you can
take advantage of.
And so we started with that typeof thing.
(04:05):
Uh, and we've learned from thatand really decided to step it
up, uh, now.
So
William Tincup (04:11):
I, I read a study the other day, and
basically, I'll paraphrase, butit's 80%, uh, CHROs have played
with in some way or form ai, uh,chat, GPT probably, uh, most of
them and 80% of CHROs areterrified of of ai.
Um, so how do we get'em over thehump?
(04:33):
I mean, it's, first of all, um,even if, let's, I believe that's
true.
'cause I, I, I know the researchfirm, but how do we, if, if we,
if, if, if there are terrified.
If that's true, how do we getthem over the hump and see this
as a friend, not a foe?
And it's
Jack Berkowitz (04:51):
a, it's a really great point.
I've thought about this a longtime.
I've been in the HR industry nowfor years.
And, you know, I remember when Ifirst got into it, I was around
HR analytics.
And HR people would say, well,you know, I don't know about
that.
I'm just, I'm not comfortablewith numbers.
And I would be like, well, I'm apsych major.
(05:13):
I'm somewhat comfortable withthem.
What, what's going on?
And I think it's really justabout the technology that we've
provided to them.
In corporate settings, it's justnot the same as the technology
that they're using in theireveryday lives.
Like if they can use Instagramor they can use Spotify, well,
they certainly can use chat GPTand build that trust
(05:33):
relationship.
So it's really contingent on usas tech providers or in the
industry to help them understandboth what it can do and what it
can't do.
And I don't mean by someresearch study, I mean in the
fingertips of when they're usingit.
So you build the system so thatit says, hey, by the way, I'm an
intelligent agent.
Hey, by the way, I actuallydon't know the answer.
(05:54):
So how do we, there's, thetechnical term is ground it.
How do we ground it so itdoesn't It doesn't tell you
something wrong, but it actuallytells you something right, and
it doesn't overstep its boundson your, you know,
inadvertently, but we can, wecan bring people down that same
path that they did in theconsumer world.
We can do that in enterprisesoftware.
William Tincup (06:13):
Yeah.
I love this because some of itis, is, as you said, it's
building trust and it'sincrementally building trust,
right?
It's just like, you don't buildtrust and all of a sudden you
have it.
Uh, uh, over 24 hours orsomething like that.
It's just small little wins andthey just see it as, and that
the word I wrote down was, wasseamless.
It's, it's almost, it's almost,it becomes a part of their
(06:36):
experience with the softwarethat they're using.
So they don't really think aboutit much.
That's right.
And,
Jack Berkowitz (06:42):
and it's important for us to, to just
remember why are we doing this?
We're doing this so that we canget HR people back to what they
want to do, which is help peopleand work with people.
I don't think any HR person eversigned up to do input
transactions or build reports.
Right.
Right.
Right.
They, they signed up becausethey want to help build a
business or they have empathyfor the employees in their
(07:05):
company.
And so, what this will allowthem to do is to spend more time
there.
And so, it's really importantthat we, we don't lose that
focus.
Why are we doing this?
It's to let them get back tobeing HR people and not, like I
said, you know, report writers.
William Tincup (07:21):
Yeah, that's, it's, it's kind of the classic
trope of, uh, the firefighters.
They come in on a Monday, youknow, if it's in office and it's
just one fire to the next fireto the next fire to the next
fire.
And of course, from seniorleadership, we've got, hey,
y'all need to be more strategic.
There's always so much time ina, in a day given, you know, in
a week, et cetera.
(07:42):
When do they have the time toeven be strategic?
Jack Berkowitz (07:45):
That's right.
So if we can give somebody back,I, I don't believe you're going
to give back a hundred percentof the time, but boy, of course,
right?
If we can give somebody back 30percent of the time, 40 percent
of the time, think about, thinkabout the benefit for, for the
company.
Think about the benefit for theemployees.
Think about the benefit for thatpractitioner.
(08:06):
Just having 40 percent of thetime or 30 percent of the time
to think or to be empathetic.
It's a huge deal.
And so that's the type of thingwe're looking for, 30%, 40%, not
a hundred percent.
William Tincup (08:18):
Right.
Right.
Well, a hundred percent wouldalmost terrify them.
Jack Berkowitz (08:21):
That's right.
Right.
That's right.
Nobody's job is going away.
I think that that's, that's thething.
Well, wait a second.
What will I do?
William Tincup (08:29):
How about think?
Jack Berkowitz (08:30):
How about think?
How about be, how about go for awalk
William Tincup (08:33):
with one of the employees?
I don't know.
Maybe, uh, take somebody out forbreakfast.
Yeah.
Something like that.
We're going to have to train.
I mean, we, we're joking aboutit, but we're going to have to.
Train or retrain HR how to beHR, great HR leaders.
So some of this is actuallyinteracting while, while this
kind of low hanging, uh, stuffhappens for them, uh, in a real
(08:57):
seamless way.
It's just happening for them.
They're going to get their timeback.
They're probably, we joke aboutit, but they probably won't know
what to do with their time.
Jack Berkowitz (09:06):
Yeah.
But isn't it wonderful with theability to have this
conversation?
Right.
Yes.
And isn't it wonderful that,that we can be part of, and when
I say we, I don't mean you andI, I mean, you know, the
industry, be part of sort ofthis, this reawakening as to why
we do what we do.
William Tincup (09:25):
It is a brand new era, uh, for them.
And especially if they look atit like that, because I think
one of the things that y'all arechanging and that I really like
is.
HR, right now, you talk to anypeople later, and it's at least
80 percent of their job isreactive.
They're just reacting towhatever's going on, whatever
the next, again, whatever thatnext thing is, and with nudges
(09:48):
and other types of technologylike that, they can become
proactive.
Jack Berkowitz (09:53):
That's right.
That's right.
And you also think about, I was,right before this phone call, I
was looking at the number ofnested regulations.
That exist for, um, familyleave, right?
There's something like 331nested regulations for somebody
in New York or New Jersey tohave to deal with.
(10:15):
If we can just help them withthe complexity of that.
How do they keep up with that?
How do they keep up with it?
And each one's probablychanging.
So we can help them with thecomplexity of that so that we
can just get them.
A quick answer or at least aview?
Yeah.
And just think about that.
Just cutting down the time theyhave to go deal with that would
(10:35):
be a huge advantage.
And that's the type of problems.
William Tincup (10:38):
Well, that's, that's part of y'all's kind of
ethos because y'all do that withpayroll libraries.
So you understand, you know,both domestically and
internationally what goes on ina payroll library.
And if there's a change inminimum wage in the city of
Dallas, uh, y'all, y'all knowthat.
Y'all, y'all probably know itbefore they know it.
Y'all know it.
And then you, if anybody has anhourly employee in Dallas,
(11:01):
that's just seamless.
It's just hap, it's just.
So, they don't have to thinkabout it, but it's more than
that they don't have to think ofit.
They can't keep track of all thechanges.
We're just on the payroll side,but you're dealing with all the
other things that are tertiaryto payroll.
How do they keep up with, uh,you know, the compensation
changes that are happening?
(11:23):
Well, if again, if thetechnology can nudge them and
educate them in a way that justsays, oh, by the way, this is
happening, it's already beingtaken care of.
That's
Jack Berkowitz (11:33):
right.
That's right.
Or, hey, by the way, you shouldknow about this, right?
Sometimes it'll take care of,sometimes it'll be like, you
just should know.
Yeah.
Yeah.
Like one of the, one of myfavorite nudges, and we haven't,
we haven't finished this oneyet.
We did the vacation one, by theway.
Right.
Uh, one of my favorite nudgesis, you know, Hey, it's great.
You just hired somebody for thatrole.
But by the way, somebody else inyour company has applied to
(11:55):
three jobs and they haven'tgotten any of them.
Yeah.
You should have a talk withthem.
Oh, I
William Tincup (11:59):
love it.
Oh, I love it.
Cause that's, that's stillmorale.
You know, somebody, somebodyclearly thinks that they should
be somewhere else, which is,which is, which is great.
But if we don't give them thetools and mechanisms to then get
to that next place and, andworse than that, if we don't
even know, which is what you'retalking about is uncovering the,
the technology to uncover,connect the dots in a way and
(12:23):
uncover these things and thenrender it to somebody so they
have insight and then they cantake an action.
That's right.
That's right.
And if you don't, you lose theperson.
I mean, we, we both know what'sgoing on here, because if that
person applies for a fourth anddoesn't get it, they're just
going to say, okay, the companyobviously doesn't want me to be
here, whether or not that's trueor not, they just don't want me
(12:43):
to be here.
And then they'll get a jobelsewhere.
And you've
Jack Berkowitz (12:46):
lost that person.
That's right.
And particularly in largercompanies, right, the fact that
a person applied to those jobsmight, nobody might ever
actually know because things aredistributed, right?
Right.
And so, so that's really thepromise of all of this.
You know, we've been working onthe data, we've been working on
the machine learning, we've beenworking on all of that for quite
(13:06):
a while.
But we're really, at this point,focused on the person side of
it, the human side of it now,and super exciting time.
Do you
William Tincup (13:14):
think that it gets to a point where, in that
scenario, that proactivelyreaches out to the employee and
tells them that there's aposition that probably fits
their skills and that theyshould apply to?
Yeah,
Jack Berkowitz (13:30):
so we We, that's exactly where this goes.
And so we have that, we, we useit a bit internally, um, but we
have this great skills graphthat's all data driven.
It's not built by hand.
It's all built on the data thatflows through our systems.
And so we, we already can tellpeople, Hey, by the way, you
know, if this is a career pathyou want to go on, here's
(13:51):
training that.
You want to take, or take thisclass or read this book and,
and, and we have the ability totell people, by the way, there's
this other job that you mightwant to think about that it's
not because of job title, butyou have these interesting
skills because everybody hasgreat skills.
Have you ever thought aboutthis?
Almost like a career counseloras opposed to just a job
(14:14):
matching thing, right?
Which,
William Tincup (14:16):
which, which everyone needs, everyone needs
somebody to be an advocate forthem.
And if the technology can be,uh, lead to this advocacy and
just, even if they don't apply,I mean, even if they're like,
no, man, I really like my gig.
I like my boss, I like the team,you know, first of all, that's
cool.
But it's someone's, I saysomeone, the technology is
advocating for them in a waythat's, that's soft, but also I
(14:40):
see really powerful.
Yeah.
Jack Berkowitz (14:42):
I mean, think about it, right?
Did you ever think you'd bedoing what you're doing for a
living 15 years ago?
I, I certainly didn't even, myjob title never even existed.
And so, yeah, it's, it's, it's,things are always changing.
And so for us to be able to helppeople on that journey is the
most important thing.
William Tincup (15:01):
So we've made it about 20 minutes and we haven't
said generative AI.
So I applaud both of us.
Uh, Now, we're about to go to HRTech, and I believe every other
booth will talk about generativeAI?
Oh, I'm gonna,
Jack Berkowitz (15:15):
I'm gonna talk about it.
There's, there's no doubt, I'llsay.
William Tincup (15:19):
Well, you know, we have to talk about it,
because it is actually reallycool, and I can see people Using
it in, in a lot of really,really cool ways.
So, so as we talk about it, whatare you thinking about
generative AI right now, justkind of on the onset?
Jack Berkowitz (15:35):
So, so we gave some really nice visions last
week and I'll do it again at HRTech, but we've got it in the
hands of people for three,three, three cases right now.
So we've got it in the hands ofpeople.
Um, for doing that law, right?
Interpreting all these policiesand things.
And it's, we've shipped it toabout 2, 000 clients at this
(15:55):
moment and learning from that.
So they're able to ask for, youknow, about legal issues or, or,
or HR policy stuff.
Um, we also have built a, um,natural language interface to
our reporting tool.
Remember I said, nobody wants towrite reports, you don't need
to, you can just ask it what youwant.
And we have this little trickwhere, that I'll show next week,
(16:16):
where you can actually copy anemail and the system will just
build the report off that.
Right.
It's really cool.
And then, and then we've got itin the hands of our associates
that support people.
Cause one of the things aboutADP is, you know, we have
thousands of people whose wholejob is to help HR people do
their job.
And, you know, we're not just asoftware vendor or SAS provider,
but we also have all thesepeople, this expertise.
(16:39):
And so to help them interpreteverything.
On behalf of clients, we've,we've got, you know, thousands
of people using generative AI tocome up with summaries of
problems and give the rightanswers.
Yeah.
And it
William Tincup (16:53):
builds, what I love about that one, uh, is it
builds a library.
It builds a library of, of, of,okay, this is, you know,
customer A has this problem.
Okay.
Well, listen, there's threedifferent ways that we've seen
this being tackled.
You know, let's, let's talkabout your company and what's
going to be the mostappropriate, et cetera, like
that.
I could see that really helping.
What I love about what you aredoing is you're not, you know,
(17:15):
pushing all your chips in andsaying generative AI, blah,
blah, blah.
You're like, let's go find aspecific use case for these
things.
And you know, let's build themslowly, build them
appropriately, and then deploythem.
And then make sure that peopleunderstand how to use them, et
cetera.
Jack Berkowitz (17:31):
That's right.
That's right.
You know, we know that there'llbe a new technology down the
road, but we're, we're here withclients for a long time.
Right.
And we want to be there withclients for a long time.
That's part of our journey withthem.
And so we want to make sure thatwhatever we field and whatever
we provide, Either to apractitioner or to, uh, to an
(17:54):
employee is the right thing forthem at the right time.
And so we, we, we do it slowly,uh, we do it with, with a
cadence, but it's impactful inthe long run.
William Tincup (18:05):
Right.
I ran sales a hundred years agoand, uh, one of the things that,
uh, talked to my sales teamabout was FUD, fear, uncertainty
and doubt.
So how do you, how do you, howdo you get around FUD or how do
you displace it with a prospect?
And so I'm thinking about HR.
And I wonder, is it, is, do weneed to have explicit
(18:25):
conversations about AI?
Does it just need to be bakedinto what we do?
Jack Berkowitz (18:33):
Uh, you know, I wish I, I wish I knew the
answer.
I, I like, I like it.
Damn it,
William Tincup (18:38):
Jack, that's why I had you on the podcast.
Jack Berkowitz (18:39):
Yeah, I know.
Uh, I like the, I like thelatter, right?
I like, it's just baked intowhat you do.
However, I think you do need tohave a bit of a discussion.
Right.
Because we.
Um, we have an ethical principlehere, like we published our
ethics about AI back in 2019.
And one of our ethicalprinciples is that people need
to be aware that they're workingwith an accelerator or augmented
(19:04):
intelligence, right?
They need to know it's a bot,that it's not a person.
So we do have to have thatconversation at some point.
How we do it is different.
I don't mean putting people in aclassroom for eight hours, maybe
the agent just pops up and says,Hey, I'm an agent.
I'm here to help.
And if you, if you need to bailout of me, click this button,
you'll talk to a human.
(19:24):
We can do that.
I
William Tincup (19:25):
like that.
I like that a lot.
Then it gives the person thepower to then say, I'd want to
talk to a person.
I mean, I do this all the timewhen I call, you know, a credit
card or something like that, orI just want to talk to a human.
And, uh, and then sometimes Ijust want to talk to a bot.
Like, hey, I have this problem.
Oh, okay.
Um, so I like the, I like theputting the power in the, in the
(19:47):
hands of the user, if you will,to then decide what, how that
plays out.
Um, and I think you, you nailedit at the beginning, privacy,
security, compliance, ethics,having the discussion about all
of those things, because we'redealing with data, I think is a
good thing.
Uh, I don't think it's probablythe first thing a salesperson
should talk to a prospect about,but before maybe they signed the
(20:10):
contract, it's like, okay,listen, at one point, you're
probably going to be asked aboutthese things.
Let's roll through this, youknow, like, and again, if they
bring it up first, that's finebecause y'all have already kind
of, y'all have already kind ofoutlined those things.
So it's easy.
It's an easy discussion to have.
But if, if, uh, I think.
(20:31):
You know, from my perspective,the less we talk about AI, the
better in the sense of it shouldjust be a part of the product,
like what you do, theirexperience.
Jack Berkowitz (20:44):
Yeah.
I mean, look, look at the greatproducts that we all use, right?
Whether it's Spotify or Uber orsomething else, right?
Those are just data drivensystems.
They're data aware AI systems.
And you know, you, you noticethat thing about you were
saying, uh, sometimes you justwant to talk to the bot, you
know, like if you get a badcharge on Uber or Lyft, you just
(21:05):
go into their bot and you tellit and it takes care of it.
There's nobody on the otherside.
But it gives you that refund andit's fantastic.
Yeah.
And it's like, well, no, I can'tdo it.
No, please look at this again.
Oh, I did.
Thank you.
Right.
Um, and, and those are the typesof experiences that, that we in
the HR industry, you know, wereally should be driving
(21:27):
towards.
Yeah.
William Tincup (21:29):
We should really be driving towards those types
of experiences.
Pun intended.
Uh, so, so.
Why wouldn't an HR leader, whywould, I mean, I'm trying to
figure out the no, like whywould they say no to this?
I
Jack Berkowitz (21:42):
think, I think, I think there's a couple of
things, right?
Um, HR people, justifiably, havebuilt the fact that they're
there to either champion theemployee or champion the
company.
And it is around trust, right?
It is around trust.
And Unless they understand theimplications of everything,
they're rightfully so doubting,right?
(22:05):
Rightfully so doubting.
And, uh, and, and I, and I thinkthat's part of it, right?
So we not, we need to not justgive them the answer, but a
little bit of explanation, or,or at least give them the
ability to ask for theexplanation.
And I think it's that it's, Ithink it's that the HR person
(22:26):
has classically been the one attimes deliver bad news.
Right.
Right.
And so therefore giving them allthe tools that they need to be
able to do that and to, to dothat confidently, uh, it's a
tricky, but we should do that.
William Tincup (22:43):
Yeah, and I think that, as you said, it's a
slow burn with trust.
You just start building trustover time, like Spotify and its
recommendation engine.
You just trust it.
Amazon and its recommendation,you just trust it.
It's like, yeah, okay, uh, I, Itrust
Jack Berkowitz (22:57):
it.
That's right, but, but you alsoknow some of them that go bad.
Yeah.
Uh, one of the video services,it's like, okay, I watched one
video about somebody hikingthrough the Alps.
I have other interests, right?
So, so,
William Tincup (23:10):
in fact, that was, that was not an interest at
all.
I was forced to watch that by mywife.
Exactly.
Jack Berkowitz (23:17):
Exactly.
And so, um, so, so that becomesa really big thing.
is to make sure whoever you'reworking with, whether it's a
vendor or you build it in houseor whatever, that, that the,
that the team supporting isactually looking at this stuff
over time and that it's beingmonitored and that.
(23:40):
You know, it's being checked,not just for bias, but it's
being checked for correctness oreven usefulness over time.
And that's another aspect,accuracy, right?
Otherwise, we're all going to bewatching videos of
William Tincup (23:53):
people walking through the OPS office.
That's a very specific example,by the way, very, very personal,
very personal.
So what the last question I haveis around risk, uh, because HR's
job in large parts is managingrisk, uh, and, and, and they.
You know, they do this job well,they take it seriously, et
cetera.
Do, do you think that we have tohave a risk discussion with them
(24:16):
around AI and kind of mitigatetheir, any concerns that they
might have?
Jack Berkowitz (24:22):
I think so.
I think so.
And, and by the way, that's notjust because of the vendor or
the technology provider.
I mean, the government.
Right?
Yes.
Government and agencies andgroups, you know, whether it's
California or the FTC or theEEOC is, is, is opening these
discussions about risk.
And there is risk, right?
There, there are, there arerightfully, you know, um, bad
(24:45):
implementations that, that weall need to be aware of.
Right.
Yeah.
We're dealing with people'slives.
I think that's the thing.
It's maybe not medical systems.
But it's still people's lives.
And whether it's, it's becausethis is a career thing or simply
equal pay for, for, for, for,for jobs, right?
It's people's lives.
And so understanding the riskand having that discussion,
(25:09):
being upfront about what you aredoing to control for it, to
account for it, to tell peopleyou don't know when you don't
know, um, that's important forus to have that conversation.
William Tincup (25:21):
Jobs Mike walks off stage.
Thank you so much.
I know you're crazy busy.
I can't even, I wouldn't want tolook at your calendar, but, uh,
but thanks again, Jack, for thetime.
Absolutely appreciate it.
Same
Jack Berkowitz (25:33):
here.
Thank you for your time.
William Tincup (25:35):
It's been great.
And thanks, and thanks everyonefor listening.
Until next time.
This is William Tincup, and you're listening to
the Recruiting Daily podcast.
Today we have Jack on from ADP,and our topic today is how ADP
is using AI to create a morehuman experience.
So let's just jump right intoit.
Jack, would you, uh, introduceyourself and what you do for
ADP?
Jack Berkowitz (00:49):
Yeah, my name's Jack Berkowitz.
I'm the Chief Data Officer hereat ADP.
Um, sort of got two jobs.
One, um, help build products,uh, whether it's people
analytics or reporting, uh, orAI products.
And then the second thing is I'mthe Corporate Chief Data
Officer.
So everything about datagovernance and data privacy and
(01:09):
data security and all thosetypes of things.
Oh, that's
William Tincup (01:13):
cool.
That's the harder part of thejob.
I'm assuming the compliance and,uh, and especially just meeting
all the, I mean, you've got,you've just, you just mentioned
privacy and security.
Those two things are massive.
Jack Berkowitz (01:28):
Yeah, you know, and I think, but they're super
important.
Oh, yeah, critical.
And it helps you build betterproducts if you do the proper
security, if you do the properprivacy.
If you think about ethics upfront, particularly in our
space, right, in HR, then youbuild better products.
And so it's a nice loop.
It does feed each other.
I
William Tincup (01:46):
can see that.
I can see that.
So, you know, I've probably beenin a lab working with AI for
years now, but, um, let's talk alittle bit about, again, how
you're using it to create morehuman experience.
Yeah.
Jack Berkowitz (02:00):
You know, we've had a number of different things
out there.
We have a skills graph all basedon AI technologies.
We have counter relevancyalgorithms, all this type of
stuff.
Right.
Um, which is super interestingfor practitioners about pooling
information, but where we'regoing next is really this notion
about how do I nudge?
How do I nudge?
Whether it's a practitioner sothey can do, you know, have more
(02:23):
automated payroll, or how can Inudge a manager to say, you
know, hey, for example, you'vegot an employee who's not been
taking vacation.
Maybe you should help them finda way to get a vacation before
they, Oh, that's cool.
Um, that type of thing.
Yeah.
William Tincup (02:38):
And so it's kind of the, uh, it looks at the
data.
I say it, uh, but basically it'slooking at data to then find
where the business can getbetter.
But instead of going direct tothat, uh, employee per se, you
nudge the manager so that themanager can be a good manager.
Jack Berkowitz (02:58):
That's right.
That's right.
Right.
Um.
At the end of the day, we knowfrom all of our data, we know
that, yeah, people leave forsalary sometimes and they'll
leave for, for careeropportunities, but normally
they're leaving because of theirmanager or their manager's
manager.
Right.
And so if we can create a betterrelationship or if we can help
support, better that, supportthe relationship between the
(03:22):
manager and the employee,everybody benefits.
Right.
You get happy employees, you geta better, uh, management
experience, you get betterrelationships.
Better results for your company.
Yeah,
William Tincup (03:31):
what I love about it is it's, it does fall
in line with employeeexperience.
It's a different take onemployee experience, actually,
because most of the EX folks isfocused on the employee, not the
manager helping the employee.
Jack Berkowitz (03:45):
Yeah, and we're not, we're not neglecting the
employee.
No, no, no, no, nudges actuallywith the employees.
Right.
So, you know, we started a fewyears ago.
Hey, by the way.
You're not saving enough forretirement.
Or, oh, by the way, there arethese types of benefits you can
take advantage of.
And so we started with that typeof thing.
(04:05):
Uh, and we've learned from thatand really decided to step it
up, uh, now.
So
William Tincup (04:11):
I, I read a study the other day, and
basically, I'll paraphrase, butit's 80%, uh, CHROs have played
with in some way or form ai, uh,chat, GPT probably, uh, most of
them and 80% of CHROs areterrified of of ai.
Um, so how do we get'em over thehump?
(04:33):
I mean, it's, first of all, um,even if, let's, I believe that's
true.
'cause I, I, I know the researchfirm, but how do we, if, if we,
if, if, if there are terrified.
If that's true, how do we getthem over the hump and see this
as a friend, not a foe?
And it's
Jack Berkowitz (04:51):
a, it's a really great point.
I've thought about this a longtime.
I've been in the HR industry nowfor years.
And, you know, I remember when Ifirst got into it, I was around
HR analytics.
And HR people would say, well,you know, I don't know about
that.
I'm just, I'm not comfortablewith numbers.
And I would be like, well, I'm apsych major.
(05:13):
I'm somewhat comfortable withthem.
What, what's going on?
And I think it's really justabout the technology that we've
provided to them.
In corporate settings, it's justnot the same as the technology
that they're using in theireveryday lives.
Like if they can use Instagramor they can use Spotify, well,
they certainly can use chat GPTand build that trust
(05:33):
relationship.
So it's really contingent on usas tech providers or in the
industry to help them understandboth what it can do and what it
can't do.
And I don't mean by someresearch study, I mean in the
fingertips of when they're usingit.
So you build the system so thatit says, hey, by the way, I'm an
intelligent agent.
Hey, by the way, I actuallydon't know the answer.
(05:54):
So how do we, there's, thetechnical term is ground it.
How do we ground it so itdoesn't It doesn't tell you
something wrong, but it actuallytells you something right, and
it doesn't overstep its boundson your, you know,
inadvertently, but we can, wecan bring people down that same
path that they did in theconsumer world.
We can do that in enterprisesoftware.
William Tincup (06:13):
Yeah.
I love this because some of itis, is, as you said, it's
building trust and it'sincrementally building trust,
right?
It's just like, you don't buildtrust and all of a sudden you
have it.
Uh, uh, over 24 hours orsomething like that.
It's just small little wins andthey just see it as, and that
the word I wrote down was, wasseamless.
It's, it's almost, it's almost,it becomes a part of their
(06:36):
experience with the softwarethat they're using.
So they don't really think aboutit much.
That's right.
And,
Jack Berkowitz (06:42):
and it's important for us to, to just
remember why are we doing this?
We're doing this so that we canget HR people back to what they
want to do, which is help peopleand work with people.
I don't think any HR person eversigned up to do input
transactions or build reports.
Right.
Right.
Right.
They, they signed up becausethey want to help build a
business or they have empathyfor the employees in their
(07:05):
company.
And so, what this will allowthem to do is to spend more time
there.
And so, it's really importantthat we, we don't lose that
focus.
Why are we doing this?
It's to let them get back tobeing HR people and not, like I
said, you know, report writers.
William Tincup (07:21):
Yeah, that's, it's, it's kind of the classic
trope of, uh, the firefighters.
They come in on a Monday, youknow, if it's in office and it's
just one fire to the next fireto the next fire to the next
fire.
And of course, from seniorleadership, we've got, hey,
y'all need to be more strategic.
There's always so much time ina, in a day given, you know, in
a week, et cetera.
(07:42):
When do they have the time toeven be strategic?
Jack Berkowitz (07:45):
That's right.
So if we can give somebody back,I, I don't believe you're going
to give back a hundred percentof the time, but boy, of course,
right?
If we can give somebody back 30percent of the time, 40 percent
of the time, think about, thinkabout the benefit for, for the
company.
Think about the benefit for theemployees.
Think about the benefit for thatpractitioner.
(08:06):
Just having 40 percent of thetime or 30 percent of the time
to think or to be empathetic.
It's a huge deal.
And so that's the type of thingwe're looking for, 30%, 40%, not
a hundred percent.
William Tincup (08:18):
Right.
Right.
Well, a hundred percent wouldalmost terrify them.
Jack Berkowitz (08:21):
That's right.
Right.
That's right.
Nobody's job is going away.
I think that that's, that's thething.
Well, wait a second.
What will I do?
William Tincup (08:29):
How about think?
Jack Berkowitz (08:30):
How about think?
How about be, how about go for awalk
William Tincup (08:33):
with one of the employees?
I don't know.
Maybe, uh, take somebody out forbreakfast.
Yeah.
Something like that.
We're going to have to train.
I mean, we, we're joking aboutit, but we're going to have to.
Train or retrain HR how to beHR, great HR leaders.
So some of this is actuallyinteracting while, while this
kind of low hanging, uh, stuffhappens for them, uh, in a real
(08:57):
seamless way.
It's just happening for them.
They're going to get their timeback.
They're probably, we joke aboutit, but they probably won't know
what to do with their time.
Jack Berkowitz (09:06):
Yeah.
But isn't it wonderful with theability to have this
conversation?
Right.
Yes.
And isn't it wonderful that,that we can be part of, and when
I say we, I don't mean you andI, I mean, you know, the
industry, be part of sort ofthis, this reawakening as to why
we do what we do.
William Tincup (09:25):
It is a brand new era, uh, for them.
And especially if they look atit like that, because I think
one of the things that y'all arechanging and that I really like
is.
HR, right now, you talk to anypeople later, and it's at least
80 percent of their job isreactive.
They're just reacting towhatever's going on, whatever
the next, again, whatever thatnext thing is, and with nudges
(09:48):
and other types of technologylike that, they can become
proactive.
Jack Berkowitz (09:53):
That's right.
That's right.
And you also think about, I was,right before this phone call, I
was looking at the number ofnested regulations.
That exist for, um, familyleave, right?
There's something like 331nested regulations for somebody
in New York or New Jersey tohave to deal with.
(10:15):
If we can just help them withthe complexity of that.
How do they keep up with that?
How do they keep up with it?
And each one's probablychanging.
So we can help them with thecomplexity of that so that we
can just get them.
A quick answer or at least aview?
Yeah.
And just think about that.
Just cutting down the time theyhave to go deal with that would
(10:35):
be a huge advantage.
And that's the type of problems.
William Tincup (10:38):
Well, that's, that's part of y'all's kind of
ethos because y'all do that withpayroll libraries.
So you understand, you know,both domestically and
internationally what goes on ina payroll library.
And if there's a change inminimum wage in the city of
Dallas, uh, y'all, y'all knowthat.
Y'all, y'all probably know itbefore they know it.
Y'all know it.
And then you, if anybody has anhourly employee in Dallas,
(11:01):
that's just seamless.
It's just hap, it's just.
So, they don't have to thinkabout it, but it's more than
that they don't have to think ofit.
They can't keep track of all thechanges.
We're just on the payroll side,but you're dealing with all the
other things that are tertiaryto payroll.
How do they keep up with, uh,you know, the compensation
changes that are happening?
(11:23):
Well, if again, if thetechnology can nudge them and
educate them in a way that justsays, oh, by the way, this is
happening, it's already beingtaken care of.
That's
Jack Berkowitz (11:33):
right.
That's right.
Or, hey, by the way, you shouldknow about this, right?
Sometimes it'll take care of,sometimes it'll be like, you
just should know.
Yeah.
Yeah.
Like one of the, one of myfavorite nudges, and we haven't,
we haven't finished this oneyet.
We did the vacation one, by theway.
Right.
Uh, one of my favorite nudgesis, you know, Hey, it's great.
You just hired somebody for thatrole.
But by the way, somebody else inyour company has applied to
(11:55):
three jobs and they haven'tgotten any of them.
Yeah.
You should have a talk withthem.
Oh, I
William Tincup (11:59):
love it.
Oh, I love it.
Cause that's, that's stillmorale.
You know, somebody, somebodyclearly thinks that they should
be somewhere else, which is,which is, which is great.
But if we don't give them thetools and mechanisms to then get
to that next place and, andworse than that, if we don't
even know, which is what you'retalking about is uncovering the,
the technology to uncover,connect the dots in a way and
(12:23):
uncover these things and thenrender it to somebody so they
have insight and then they cantake an action.
That's right.
That's right.
And if you don't, you lose theperson.
I mean, we, we both know what'sgoing on here, because if that
person applies for a fourth anddoesn't get it, they're just
going to say, okay, the companyobviously doesn't want me to be
here, whether or not that's trueor not, they just don't want me
(12:43):
to be here.
And then they'll get a jobelsewhere.
And you've
Jack Berkowitz (12:46):
lost that person.
That's right.
And particularly in largercompanies, right, the fact that
a person applied to those jobsmight, nobody might ever
actually know because things aredistributed, right?
Right.
And so, so that's really thepromise of all of this.
You know, we've been working onthe data, we've been working on
the machine learning, we've beenworking on all of that for quite
(13:06):
a while.
But we're really, at this point,focused on the person side of
it, the human side of it now,and super exciting time.
Do you
William Tincup (13:14):
think that it gets to a point where, in that
scenario, that proactivelyreaches out to the employee and
tells them that there's aposition that probably fits
their skills and that theyshould apply to?
Yeah,
Jack Berkowitz (13:30):
so we We, that's exactly where this goes.
And so we have that, we, we useit a bit internally, um, but we
have this great skills graphthat's all data driven.
It's not built by hand.
It's all built on the data thatflows through our systems.
And so we, we already can tellpeople, Hey, by the way, you
know, if this is a career pathyou want to go on, here's
(13:51):
training that.
You want to take, or take thisclass or read this book and,
and, and we have the ability totell people, by the way, there's
this other job that you mightwant to think about that it's
not because of job title, butyou have these interesting
skills because everybody hasgreat skills.
Have you ever thought aboutthis?
Almost like a career counseloras opposed to just a job
(14:14):
matching thing, right?
Which,
William Tincup (14:16):
which, which everyone needs, everyone needs
somebody to be an advocate forthem.
And if the technology can be,uh, lead to this advocacy and
just, even if they don't apply,I mean, even if they're like,
no, man, I really like my gig.
I like my boss, I like the team,you know, first of all, that's
cool.
But it's someone's, I saysomeone, the technology is
advocating for them in a waythat's, that's soft, but also I
(14:40):
see really powerful.
Yeah.
Jack Berkowitz (14:42):
I mean, think about it, right?
Did you ever think you'd bedoing what you're doing for a
living 15 years ago?
I, I certainly didn't even, myjob title never even existed.
And so, yeah, it's, it's, it's,things are always changing.
And so for us to be able to helppeople on that journey is the
most important thing.
William Tincup (15:01):
So we've made it about 20 minutes and we haven't
said generative AI.
So I applaud both of us.
Uh, Now, we're about to go to HRTech, and I believe every other
booth will talk about generativeAI?
Oh, I'm gonna,
Jack Berkowitz (15:15):
I'm gonna talk about it.
There's, there's no doubt, I'llsay.
William Tincup (15:19):
Well, you know, we have to talk about it,
because it is actually reallycool, and I can see people Using
it in, in a lot of really,really cool ways.
So, so as we talk about it, whatare you thinking about
generative AI right now, justkind of on the onset?
Jack Berkowitz (15:35):
So, so we gave some really nice visions last
week and I'll do it again at HRTech, but we've got it in the
hands of people for three,three, three cases right now.
So we've got it in the hands ofpeople.
Um, for doing that law, right?
Interpreting all these policiesand things.
And it's, we've shipped it toabout 2, 000 clients at this
(15:55):
moment and learning from that.
So they're able to ask for, youknow, about legal issues or, or,
or HR policy stuff.
Um, we also have built a, um,natural language interface to
our reporting tool.
Remember I said, nobody wants towrite reports, you don't need
to, you can just ask it what youwant.
And we have this little trickwhere, that I'll show next week,
(16:16):
where you can actually copy anemail and the system will just
build the report off that.
Right.
It's really cool.
And then, and then we've got itin the hands of our associates
that support people.
Cause one of the things aboutADP is, you know, we have
thousands of people whose wholejob is to help HR people do
their job.
And, you know, we're not just asoftware vendor or SAS provider,
but we also have all thesepeople, this expertise.
(16:39):
And so to help them interpreteverything.
On behalf of clients, we've,we've got, you know, thousands
of people using generative AI tocome up with summaries of
problems and give the rightanswers.
Yeah.
And it
William Tincup (16:53):
builds, what I love about that one, uh, is it
builds a library.
It builds a library of, of, of,okay, this is, you know,
customer A has this problem.
Okay.
Well, listen, there's threedifferent ways that we've seen
this being tackled.
You know, let's, let's talkabout your company and what's
going to be the mostappropriate, et cetera, like
that.
I could see that really helping.
What I love about what you aredoing is you're not, you know,
(17:15):
pushing all your chips in andsaying generative AI, blah,
blah, blah.
You're like, let's go find aspecific use case for these
things.
And you know, let's build themslowly, build them
appropriately, and then deploythem.
And then make sure that peopleunderstand how to use them, et
cetera.
Jack Berkowitz (17:31):
That's right.
That's right.
You know, we know that there'llbe a new technology down the
road, but we're, we're here withclients for a long time.
Right.
And we want to be there withclients for a long time.
That's part of our journey withthem.
And so we want to make sure thatwhatever we field and whatever
we provide, Either to apractitioner or to, uh, to an
(17:54):
employee is the right thing forthem at the right time.
And so we, we, we do it slowly,uh, we do it with, with a
cadence, but it's impactful inthe long run.
William Tincup (18:05):
Right.
I ran sales a hundred years agoand, uh, one of the things that,
uh, talked to my sales teamabout was FUD, fear, uncertainty
and doubt.
So how do you, how do you, howdo you get around FUD or how do
you displace it with a prospect?
And so I'm thinking about HR.
And I wonder, is it, is, do weneed to have explicit
(18:25):
conversations about AI?
Does it just need to be bakedinto what we do?
Jack Berkowitz (18:33):
Uh, you know, I wish I, I wish I knew the
answer.
I, I like, I like it.
Damn it,
William Tincup (18:38):
Jack, that's why I had you on the podcast.
Jack Berkowitz (18:39):
Yeah, I know.
Uh, I like the, I like thelatter, right?
I like, it's just baked intowhat you do.
However, I think you do need tohave a bit of a discussion.
Right.
Because we.
Um, we have an ethical principlehere, like we published our
ethics about AI back in 2019.
And one of our ethicalprinciples is that people need
to be aware that they're workingwith an accelerator or augmented
(19:04):
intelligence, right?
They need to know it's a bot,that it's not a person.
So we do have to have thatconversation at some point.
How we do it is different.
I don't mean putting people in aclassroom for eight hours, maybe
the agent just pops up and says,Hey, I'm an agent.
I'm here to help.
And if you, if you need to bailout of me, click this button,
you'll talk to a human.
(19:24):
We can do that.
I
William Tincup (19:25):
like that.
I like that a lot.
Then it gives the person thepower to then say, I'd want to
talk to a person.
I mean, I do this all the timewhen I call, you know, a credit
card or something like that, orI just want to talk to a human.
And, uh, and then sometimes Ijust want to talk to a bot.
Like, hey, I have this problem.
Oh, okay.
Um, so I like the, I like theputting the power in the, in the
(19:47):
hands of the user, if you will,to then decide what, how that
plays out.
Um, and I think you, you nailedit at the beginning, privacy,
security, compliance, ethics,having the discussion about all
of those things, because we'redealing with data, I think is a
good thing.
Uh, I don't think it's probablythe first thing a salesperson
should talk to a prospect about,but before maybe they signed the
(20:10):
contract, it's like, okay,listen, at one point, you're
probably going to be asked aboutthese things.
Let's roll through this, youknow, like, and again, if they
bring it up first, that's finebecause y'all have already kind
of, y'all have already kind ofoutlined those things.
So it's easy.
It's an easy discussion to have.
But if, if, uh, I think.
(20:31):
You know, from my perspective,the less we talk about AI, the
better in the sense of it shouldjust be a part of the product,
like what you do, theirexperience.
Jack Berkowitz (20:44):
Yeah.
I mean, look, look at the greatproducts that we all use, right?
Whether it's Spotify or Uber orsomething else, right?
Those are just data drivensystems.
They're data aware AI systems.
And you know, you, you noticethat thing about you were
saying, uh, sometimes you justwant to talk to the bot, you
know, like if you get a badcharge on Uber or Lyft, you just
(21:05):
go into their bot and you tellit and it takes care of it.
There's nobody on the otherside.
But it gives you that refund andit's fantastic.
Yeah.
And it's like, well, no, I can'tdo it.
No, please look at this again.
Oh, I did.
Thank you.
Right.
Um, and, and those are the typesof experiences that, that we in
the HR industry, you know, wereally should be driving
(21:27):
towards.
Yeah.
William Tincup (21:29):
We should really be driving towards those types
of experiences.
Pun intended.
Uh, so, so.
Why wouldn't an HR leader, whywould, I mean, I'm trying to
figure out the no, like whywould they say no to this?
I
Jack Berkowitz (21:42):
think, I think, I think there's a couple of
things, right?
Um, HR people, justifiably, havebuilt the fact that they're
there to either champion theemployee or champion the
company.
And it is around trust, right?
It is around trust.
And Unless they understand theimplications of everything,
they're rightfully so doubting,right?
(22:05):
Rightfully so doubting.
And, uh, and, and I, and I thinkthat's part of it, right?
So we not, we need to not justgive them the answer, but a
little bit of explanation, or,or at least give them the
ability to ask for theexplanation.
And I think it's that it's, Ithink it's that the HR person
(22:26):
has classically been the one attimes deliver bad news.
Right.
Right.
And so therefore giving them allthe tools that they need to be
able to do that and to, to dothat confidently, uh, it's a
tricky, but we should do that.
William Tincup (22:43):
Yeah, and I think that, as you said, it's a
slow burn with trust.
You just start building trustover time, like Spotify and its
recommendation engine.
You just trust it.
Amazon and its recommendation,you just trust it.
It's like, yeah, okay, uh, I, Itrust
Jack Berkowitz (22:57):
it.
That's right, but, but you alsoknow some of them that go bad.
Yeah.
Uh, one of the video services,it's like, okay, I watched one
video about somebody hikingthrough the Alps.
I have other interests, right?
So, so,
William Tincup (23:10):
in fact, that was, that was not an interest at
all.
I was forced to watch that by mywife.
Exactly.
Jack Berkowitz (23:17):
Exactly.
And so, um, so, so that becomesa really big thing.
is to make sure whoever you'reworking with, whether it's a
vendor or you build it in houseor whatever, that, that the,
that the team supporting isactually looking at this stuff
over time and that it's beingmonitored and that.
(23:40):
You know, it's being checked,not just for bias, but it's
being checked for correctness oreven usefulness over time.
And that's another aspect,accuracy, right?
Otherwise, we're all going to bewatching videos of
William Tincup (23:53):
people walking through the OPS office.
That's a very specific example,by the way, very, very personal,
very personal.
So what the last question I haveis around risk, uh, because HR's
job in large parts is managingrisk, uh, and, and, and they.
You know, they do this job well,they take it seriously, et
cetera.
Do, do you think that we have tohave a risk discussion with them
(24:16):
around AI and kind of mitigatetheir, any concerns that they
might have?
Jack Berkowitz (24:22):
I think so.
I think so.
And, and by the way, that's notjust because of the vendor or
the technology provider.
I mean, the government.
Right?
Yes.
Government and agencies andgroups, you know, whether it's
California or the FTC or theEEOC is, is, is opening these
discussions about risk.
And there is risk, right?
There, there are, there arerightfully, you know, um, bad
(24:45):
implementations that, that weall need to be aware of.
Right.
Yeah.
We're dealing with people'slives.
I think that's the thing.
It's maybe not medical systems.
But it's still people's lives.
And whether it's, it's becausethis is a career thing or simply
equal pay for, for, for, for,for jobs, right?
It's people's lives.
And so understanding the riskand having that discussion,
(25:09):
being upfront about what you aredoing to control for it, to
account for it, to tell peopleyou don't know when you don't
know, um, that's important forus to have that conversation.
William Tincup (25:21):
Jobs Mike walks off stage.
Thank you so much.
I know you're crazy busy.
I can't even, I wouldn't want tolook at your calendar, but, uh,
but thanks again, Jack, for thetime.
Absolutely appreciate it.
Same
Jack Berkowitz (25:33):
here.
Thank you for your time.
William Tincup (25:35):
It's been great.
And thanks, and thanks everyonefor listening.
Until next time.
Popular Podcasts
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.