October 27, 2023 • 27 mins
Uncover the transformative power of HR Tech with industry expert George LaRocque, Founder of WorkTech. We dissect AI's impact, from streamlining data integration to uncovering hidden talents in Slack channels. Question what you know about AI, exploring the allure and risks of borrowed language models, comparing the emergence of AI to the rise of social media.
This conversation promises an enlightening exploration of the future of HR Tech, one you don't want to miss.
Special mini series recorded with Oleeo at HR Tech 2023 with hosts Ryan Leary, Brian Fink, and Shally Steckerl
Listen & Subscribe on your favorite platform
Apple | Spotify | Google | Amazon
Visit RecruitingDaily
Twitter @RecruitingDaily
Join the Secret Sourcing Group
Learn more about #HRTX Events
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:07):
And we are back.
This is Shali Stekaral comingat you live from the HR Tech
Expo floor Powered today by theOlio booth.
We are broadcasting here in theOlio booth, taking names and
prisoners and just having a goodold time.
Our next guest is a long, longtime friend, old friend of mine,
(00:32):
george LaRocque, and George isfrom Workforce Tech, work Tech,
work Tech, work Tech.
How appropriate, hr Tech, workTech.
I think Work Tech is betterthan HR Tech because, like it's
about work.
Speaker 2 (00:47):
Right.
Speaker 1 (00:48):
Yeah, I like it.
So, george, why are you here atthe show and what's been your
impression so far?
Speaker 2 (00:55):
Well, I'm here at the show for a couple of reasons.
One, I do a lot of work at theshow, so I had a pre-conference
bringing together investors andstartups and scale-ups that this
show puts on, and I'm the chairof that.
I've been hosting the startuppitch fest for the last few days
(01:16):
.
Oh you introduce people andeverything.
Yeah, I do all that.
I provide a little insight on,you know, with the data that we
track and the trends that wetrack, all of the investment
coming into the space and youknow what are the trends, what
are people buying, et cetera.
You know in between.
And then I also come here tomeet people, see friends and
learn, see what's going on.
(01:37):
It's a great place to get apulse on the tech side of what's
happening.
Speaker 1 (01:40):
It is Definitely.
So what is your take on thepulse?
That's what's happening now.
Not the pitch fest, but likelooking at the booths and
walking around.
Yeah.
Speaker 2 (01:51):
So it seems like the two conversations are skills and
AI.
Of course we knew it.
Speaker 1 (01:58):
There's a bingo word for you.
We got ATS, we got AI and wegot CRM.
Bingo, yeah.
So skills meaning like skillingup, learning and development,
or like finding skills orontologies, yeah.
Speaker 2 (02:12):
Well, because everybody's talking about it,
yes, all of that and more.
You know.
So, depending on where they'recoming from, you've got even the
labor market data.
Folks are starting to look atskills market data.
So looking at, you know, notjust the jobs and geographic
location and comp by job, butskill, skills and comp by skill
(02:34):
and that sort of thing.
And then you've got, as you,you know, start from the very,
very, very, very top of thefunnel and go all the way into,
you know, post hire and learningand development.
It's, you know, skills, somarketplaces and upskilling and
re-skilling, and it's, you know,everybody's got an angle on
(02:56):
skills, got you Everybody.
Speaker 1 (02:59):
So I have two questions for you, all right,
based on the skills conversation.
The first of those is from apersonal perspective.
Recently started not recently,it lasts several months I
started using LinkedIn's.
I don't know skills catalog orontology so you're probably a
(03:20):
lot more familiar with this thanI am, but when you are posting
jobs to LinkedIn recruiterthrough the API and, for example
, like greenhouse publishes them, you know directly.
It used to not be very catalog,so if you had a software
engineer, it was just posted asa software engineer.
It might have maybe kind ofnarrowed down by industry.
(03:42):
But now they have this bigcatalog of skills that are
interrelated.
Are you familiar with what I'mtalking about?
Speaker 2 (03:49):
I'm familiar because I've seen the content on it.
But I haven't gone deep on that.
Speaker 1 (03:57):
So I started working with that and tweaking kind of
what.
Because there's two problemsthat happen very frequently with
job postings, and they have alot to do with classification.
One of them is the job titleitself.
If a company has a let's justcall it a cool job title, it
might be cool to the company, itmight set them apart, it might
(04:19):
be unique, but people don't knowwhat it means.
And so you end up withcandidates that don't understand
what the job is and they applyanyway because they're figured
out it might as well take achance and they're not qualified
because they're just not reallygetting the job title.
And then the opposite of thathappens you got people who are
in the job that you want to hirefor, but their job title is
completely different than yourcool title and so they don't see
(04:41):
themselves in it.
So they might think well, I'm asoftware engineer and this job
is for a you know, back-end datadeveloper or whatever, and
that's not me, so they don'tapply.
So the good ones don't applybecause the job title doesn't
resonate and the ones that don'tmatch apply because they don't
quite understand it.
So it's a problem in that youend up with just a higher
(05:04):
percentage of not reallyqualified candidates.
This has helped combat that, inthat these skills are matched,
and so I think the confidencelevel in applicants is
increasing Over time.
It'll get better.
Where a software engineer looksat the job and goes, ah,
go-lang, back-end data scientistdeveloper, that doesn't sound
(05:26):
right.
But everything else on the pageis me, so this has to be me.
So that's kind of.
What you're talking about isthe skills.
Speaker 2 (05:35):
That's one of the big use cases, both for hiring and
for internal mobility orupskilling.
So I did a case study last yearwith Manulife and they went and
Manulife is a big insurancecompany, global, and this was in
their engineering group.
So IT and development theproject took them from 1,400
(06:02):
roles to 40, so the biggest pullin the tent to your point was
change management, whereas, hey,I know your title is software
engineer but you're an analystand let me show you where we're
going with the organization.
Or we know you've got yourdepartment structured with these
titles, but we're moving inthis direction and some people
it wasn't a bad thing, not allchange is bad.
(06:23):
It was like, oh, you'reactually an engineer, good for
you.
That makes sense, but it was abig project.
Consultants travel around theworld.
That was a change managementprocess Once they got that to 40
, then maintenance of thatbecame ongoing.
Speaker 1 (06:40):
The flywheel fed itself.
Speaker 2 (06:42):
But now with technology.
So this is, I'm going to say,the promise of large language
models.
But this is where a lot ofpeople are heading.
So I think what would be coolin that example you gave is if,
dynamically, those skills wereto look at you.
I didn't the titles that youused.
But if it was like yeah, backin engineer versus software
(07:03):
developer versus software.
Speaker 1 (07:05):
So here's your title and here's this job, and you
could reflect that.
Speaker 2 (07:09):
Here's how your job now compares to this job
regardless of the job title isnot mapping that promises.
Speaker 1 (07:15):
That is what I would like to see happen organically,
in the background.
I'm manually mapping it becausethere are skills that are
required.
For example, if an organizationis not really language
dependent, like they'redeveloping in multiple languages
, the job posting will haveseveral.
You know Go Lang, java, python,et cetera.
(07:38):
What that tends to sometimestranslate into is that the
candidate must have all of theabove, which is really important
.
What it actually supposed tomean is we really don't care if
you're good at Go Lang, good atPython and good at Java, as long
as you're good at one of themand we would like to know which
one is your preference but we'lltake them.
(07:59):
But instead what people see isman, I love Go Lang, but I don't
know enough about it and Ihaven't used it at work, so I'm
not going to apply, and that'screating that.
The other question that I havefor you folks that are listening
George and I go back I don'tknow what to 10, 15 years at
least.
Oh, at least, at least it'slonger than that.
Speaker 2 (08:20):
Yeah, so I used to work for, we'll say 10, 15.
Yeah, something like that.
It's more than a decade.
I used to work for this companythat was a pioneer in the
distribution of jobs and Iliterally went after George.
Speaker 1 (08:33):
I recruited him.
You were a recommendation.
Did you know?
You were a recommendation fromsomeone, an angel investor that
I knew.
Speaker 2 (08:41):
I don't know if I ever told you that I don't think
you did yeah.
Speaker 1 (08:46):
So I called this friend, my friend of him, like
dude, I need somebody that'slike no nonsense, that's going
to come in and just hit theground running and can run sales
for a mature enterprise product.
And they're like, I can onlythink of one person that gave me
your name and I reached out andthe rest is history.
And so from there, when we bothleft that organization.
(09:09):
But you left and you startedyour own thing and you became
essentially like I want to say,like the Uber analyst, Like an
analyst is not a good title foryou.
You're the mega analyst of worktechnology.
Speaker 2 (09:22):
All right, thank you, sure Thanks.
Speaker 1 (09:24):
Because you're not just like out there writing an
article or two.
I mean, you're evaluatingproducts, You're advising
products, You're you know, andpeople depend on your opinion.
So two things related to that.
Number one literally puteverything I just said away for
a moment and pretend that youdon't know anything about it on
the industry and you walk around.
What is the most awesome booth?
No judgment about what they do.
(09:45):
What are?
Just the booth itself?
Oh wow, Because these boothsare darn cool and last year they
weren't this pretty.
Speaker 2 (09:53):
Yeah, that's a hard question for me because I've
been running that pitch fast, soI just told you, for like this
is you haven't walked around.
Speaker 1 (10:00):
I walked from there to unfair unfair question but
but they're like massive, likewe got.
We got baseball field sizeboots here.
Speaker 2 (10:08):
OK, I am, but here there's one that looks pretty
cool.
I am not endorsing this product.
Speaker 1 (10:15):
Not at all, it's just a boot.
But I mean like I am literallynot endorsing this product.
Speaker 2 (10:21):
It's that's going to get me in trouble.
It's Ripley.
Speaker 1 (10:25):
Oh yeah.
Speaker 2 (10:26):
And so they've got that the sphere I just walked by
and I'm going to stop.
It's like a sphere, and thenthey have one of those.
You see it at all the tradeshows now, but it's the sphere
around it.
That's kind of cool and it tiesto the I think thematically to
the sphere in Vegas.
But one of those camera thingsthat spins around you.
Speaker 1 (10:43):
We're in there, ok.
Speaker 2 (10:45):
But it just looks.
You know it's, it's, they havenice colors and it's and that
very looks very appealing.
But that's good yeah.
Speaker 1 (10:52):
So now put all your knowledge back on the table.
Speaker 2 (10:55):
OK, all right.
Speaker 1 (10:56):
And think about what you've seen or heard pitchfaster
, not that in the last week orso here.
Conversation wise, what is,what is a truly innovative?
It's a tough question because Ihaven't seen a whole lot of
innovation.
I've seen a lot of expansioninto other areas.
I've seen a lot of rebranding.
(11:17):
Honestly, I haven't seen a lotof new companies, except for
startup corridor Yep, and thoseare all you know new, but
unfortunately it's a small booththat you almost have to go to
their website to figure out whatthey do.
Speaker 2 (11:28):
Right, right.
Speaker 1 (11:29):
So all of what you've taken in conversations and
everything is there.
Is there something that isinnovative?
Dare I say, game changing?
Speaker 2 (11:40):
Yeah, in work tech it's, it's, it's been.
I'm very, I'm so jaded, charlie.
Speaker 1 (11:49):
I've been, I've been around for so long, and that's
what we're talking about isyou're the Uber animal, so it's
you know all this stuff.
Speaker 2 (11:56):
So it's, it's a tough question.
It's also not just about mebeing jaded, but also I'm going
to give you an answer, but Iit's also about you know, to
your point just to explain forit to stand out for you means
yeah.
Well, but just like to to.
So those pitches right that wewatched, like this year there
were like the finalists, likereally tight pitches and like
(12:17):
immediately you know what theydo, not that?
Not, I'm not saying that'swhere the innovation was, but
most of these booths I have noidea what they do when I stand
in front of it or even when theytalk to me, for four or five
minutes.
Speaker 1 (12:29):
I've noticed that, yeah.
Speaker 2 (12:30):
And so so it's, you need to peel back the the layers
a little bit.
There was one of the startupsthat and this is really
foundational, but I just wentover after that they were one of
the finalists Aragorn AI andwhat they, what they've built,
is remember back in the day wewould have these like hub and
(12:52):
spoke like connector models todo integrations across platforms
.
So they, what they've built, isthey've leveraged AI.
So they had to build, you know,integration I was talking about
this yesterday.
Yeah that they've leveraged AI.
Yeah, to do a lift and shiftthe ETL, yeah, yeah, yeah,
that's a really good I wastalking about that being a
(13:12):
really good application for AI,and that's what they're doing.
Speaker 1 (13:15):
That's yeah, they're doing a lot of higher them to be
that I'm going to use a reallyold word, middleware.
Yeah, yeah absolutely thetranslator between Product A and
.
Speaker 2 (13:24):
Product B yeah, they're like the next wave, like
beyond.
You know we're getting pastlike the.
I'll call it manual, but youknow the hub and spoke sort of
connector model, or like Zapier,or yeah, yeah and then what
that does for the ecosystem.
Here is you know you could beon Workday or you could be on
(13:45):
Seridian and you could haveGreenhouse or whatever you have
and the because it's more fluidright, the automation, the
opportunities to do interestingthings with your data and your
workflow is way better than ahard-coded, you know.
Speaker 1 (14:05):
Or you have to write a custom API or hire yeah, you
got to go back, yeah, we want todo this now.
Software engineering, do theintegration specifically for
your version of whatever.
Because everybody you know,ultimately, because not very
many maybe in your experiencebut in mine not very many
organizations will like kind ofuse the out-of-the-box.
Yeah, they'll, you know.
Speaker 2 (14:24):
Yeah, I'm going to.
I'll give you another startupanswer To Argonne.
Yeah yeah, eric Gorn, eric Gorn,Eric Gorn, ai.
Yep, then I can't remember thename of this company because it
was one of the 33 pitches, butthis is a creepy innovation.
They did a facial recognition.
Like you're sitting at yourdesk and this thing is on all
(14:44):
the time and doing facialrecognition and measuring
employee sentiment.
You know they're like claimingto measure, like you know,
mental wellness, and I was thehost so I didn't get to ask all
the questions that I had butthat's I think that's a really
(15:05):
scary place to go.
Speaker 1 (15:06):
It's innovative, I'm assuming that they're not
recording but analyzing, right,because that would be kind of
the so you know, first of all,recording massive amounts of
video and story is like kind ofuseless but also the privacy
infringement in that, you know,even if you're at work, the
security camera in your officemight record on a loop, but it's
(15:27):
not intended to recordpermanently.
So let's assume that they'reessentially analyzing, you know,
overall wellness based on yourfacial expressions or whatever
right?
So they're analyzing the data.
I can see the applications onthat, but I can also see how
that would be.
You know, you know mybackground.
(15:48):
My undergraduate is inintercultural and non-verbal
communications and to me one ofthe biggest mistakes that we
make, generally speaking ashumans, is that we misinterpret
what other people are thinkingor feeling based on what our
particular proclivities might beRight.
So, for example, I'll just giveyou a really easy example you
(16:11):
and I being from a Westernculture, to us a smile that is a
nice big smile, is sort oftoothy.
Right, you open up wine, youshow your pearly whites and
that's a very friendly, warmsmile, where there are cultures
where if you show your teeth isan act of open aggression.
So my concern with thisintelligence is that if it
(16:35):
doesn't detect toothy smiles,it's gonna say, oh, this
person's unhappy, when in factit might be that they're quite
happy.
They're just culturallypredisposed to not, you know,
have a toothy smile, or it'sjust one thing.
And then you've gotneurodivergent people and you've
got people who have facial tics.
I mean, I'm thinking about allthe possible misinterpretations
(16:56):
that could happen there?
Speaker 2 (16:57):
Yeah, I think there.
Yeah, I wasn't surprised to seeit, given you know.
Speaker 1 (17:03):
AI, all the AI hype and everything.
Speaker 2 (17:06):
Here's another one.
You might like.
This one collab work thatsounds familiar.
They do non-traditional postingof jobs and recruitment, so
what they call the hidden talent.
So they're not going to jobboards, they're not going to
social media, they're gettingjobs out into Slack channels and
(17:30):
they're getting into otherdestinations where the talent
that you're looking for is notnecessarily there.
Speaker 1 (17:39):
It's not about jobs, Hidden talent, not intentionally
hidden, but yeah and that's thephrase they use.
Speaker 2 (17:47):
You know, they've got AI, you know, and everybody's
got AI in the mix.
I'm not sure how they're usingit, but it was pretty, I think,
with all the pressure on finding, finding talent through new
channels.
Speaker 1 (18:01):
Just tab sources.
Speaker 2 (18:02):
I think you know I think they'll definitely get
some traction.
Speaker 1 (18:05):
For sure I can think of applications for that in very
niche areas where you don'treally have people looking on
the traditional job boardsbecause they know that there's
not a lot of jobs for them.
Speaker 2 (18:17):
Yeah, I mean I go back when I was recruiting.
I don't know if you rememberthe Usenet news group days, yup,
usenet, I mean I got a lot ofgood candidates out of Usenet.
Yeah, they'd sit there on theMac and, just you know, put the
job every day, set aside an hourto get the jobs.
That was the old boards.
Speaker 1 (18:37):
Yeah, oldest one of those would probably be like the
well, from back in the, youknow you dial up days, but you
know.
So you've mentioned AI a coupletimes and I know it's one of
our bingo words, yeah.
Yeah, I would love for you toexplain what, in your perception
, is really called AI, because Iknow, and you probably know
(19:02):
that I know that it's not AI,right, it's not really true AI.
So they call it generative AIbecause they want to qualify it
and you know, basically justlarge language models and it's
data science and it's machinelearning.
But what, in your estimation,as an analyst, looking at this
industry deeply, what you know?
They're calling it AI.
They used to call it semanticsearch.
They used to call it, you know,whatever, but what is it really
(19:25):
?
Speaker 2 (19:26):
Yeah, and I think you bring up a really good point.
If you know it could be machinelearning, it could be.
There are so many flavors ofwhat's what they call it yeah,
the AI is like the catch allterm.
Yeah, I think at an event likethis, I think that's probably a
good way to look at it.
Speaker 1 (19:44):
It's a catch all term .
Speaker 2 (19:46):
It's more of a catch all term.
I think what generative AI didwas open up everybody's minds to
the possibilities, and a lot ofwhat we have out there is
aspirational.
It's, you know, superaspirational right now.
I don't think, you know, Idon't think there's anything on
(20:08):
this floor that's going to passthe Turing test yeah.
So you know, there's Rightright.
Speaker 1 (20:14):
So you know the traditional definition of a true
artificial intelligence that'sable to extract subtext from
context and can make sense of avariety of different things.
You know that's not what we'relooking at here.
That's what I was kind ofwondering is if you have a sense
of what, in general, they'rereferring to when they say AI.
(20:35):
Is it, generally speaking, theyou know large language models,
or is it something else?
Because the video processingone they're doing, that's not
large language models.
Right, and the you know theother ones that I've seen on
here that are more around likechat bots.
That's also not really a largelanguage model.
There's a gentleman that was atthe podcast earlier talking
(20:58):
about recruiting sourcing toolthat they're using.
He's a data scientist and whatthey've created is not a large
language model, the type thatyou would see in chat GPT, for
example, which is unhomogenousdata.
His is specifically homogenous.
So you used to be a boardmember of the HRXML consortium.
Speaker 2 (21:17):
Right right.
Speaker 1 (21:20):
Literally, what is the HRXML equivalent these days?
You know that doesn't existanymore.
No, no.
But what is the?
Are these language models?
Essentially a new way of sortof categorizing data into some
sort of Homogenous pool so youcan later use it Like.
These are conversations, theseare resumes, these are, you know
(21:40):
, is that what we're talkingabout?
Because machines can do thatreally well.
Speaker 2 (21:47):
I think some of them are talking about that.
So like an eight-fold theybefore they ever came out with
anything generative they put alot of effort into large
language models.
Their founders are, you know,scientists from Google.
Speaker 1 (22:02):
Data science.
Yeah, so that's an ontology, ataxonomy.
Yeah, exactly that's whatsemantic search used to be.
Speaker 2 (22:08):
Right, and so you look there and I would say
there's an example.
They're doing that.
Now, your average recruitingtool that is talking about AI is
probably got a field or an areain the app where it'll write an
email for you.
It's using generative AI andit's using not their own large
(22:32):
language model.
Speaker 1 (22:33):
No, they're borrowing from somebody else's which is
scary because it becomeseverybody is using the same
language.
Speaker 2 (22:40):
It's very bland.
Speaker 1 (22:41):
Right, Very middle of the road, which is not very
convincing yeah now others are.
Speaker 2 (22:50):
You know, back to the skills conversation or if
you're thinking about somethingthat's a little more sensitive,
like when you're talking aboutjob descriptions, those are
going external and those areexposed anyhow so you don't have
to be as careful with those,and it's good to capture more of
a global sort of learn from theworld at large, but if you're
(23:15):
looking at anything close to orinvolving PII or sensitive data,
you need to have enough data tohave your own language model,
and it's got to be private,otherwise you're exposing a lot
of risk.
So there are a lot of and don'tget me wrong a tool that makes
(23:37):
your email sound better, thatengages candidates better,
that's effective.
Speaker 1 (23:43):
And that's really helpful.
Most of us are not greatwriters.
Speaker 2 (23:47):
Yeah, so that's great , but that's becoming table
stakes, like really fast.
Speaker 1 (23:52):
That's my point.
Speaker 2 (23:53):
That is table stakes.
Now, like every tool that I useI'll say almost every tool that
I use on my desk has somethinglike that.
Speaker 1 (24:02):
Yeah, there's something you know it's an
assistive thing.
Yeah, it's clippy, it's clippy.
Speaker 2 (24:05):
Yeah, it's just yeah, remember clippy.
Yeah, I do remember clippy,it's clippy.
Speaker 1 (24:09):
Yeah, okay, so that makes me feel a little bit
better about kind of the stateof the industry, because then we
have a lot of different placesto go from here rather than just
everybody calling it, becauseit used to be all everybody
called it semantic search and wewrote that white paper and then
people stopped talking about itand now it's this other thing
and you know, probably gonna bethe buzz for a while.
People stopped talking about it.
(24:30):
It kind of gives me the youknow, social media, Way back
when we started talking aboutsocial media and we were talking
about, ooh, social media, nowit's just media man.
I mean it's not social, it'sjust, it is part of the mix.
So AI is going to be somewherealong the line part of the mix.
Speaker 2 (24:47):
Yeah, I think that's an excellent analogy and I
actually use that example a lot.
And so we used to talk aboutsocial recruiting and it's just
recruiting now and it took about10 years to really we had to
become mainstream and we hadsuccess and we had all along the
(25:07):
way, but it took seven to 10years.
I'll say before you really,mobile recruiting, right, yeah,
right, mobile, yeah.
Speaker 1 (25:14):
So this will be AI in five years.
It'll just be recruiting, yeah,I think yeah.
Speaker 2 (25:20):
So this cycle, I think, is more like five years
on the money and another thingthat should make you feel good.
So I just in the recruitingspace, specifically recruitment
marketing space, I've been on abig journey, talking to
everybody, looking at becausethere are a lot of threats
around with what's coming here,and I ventured outside.
(25:45):
I went and said, well, what arethey doing in B2C marketing and
in B2B marketing?
So I tapped into thought leaderstuff there.
They sound just like us.
They have the conversation wejust had.
We could have been talking aboutmarketing platforms and CRMs in
B2C in B2B and so with that,when I say it should make you
(26:05):
feel good, we're all in it.
Yeah, well, we're on par.
So we always used to say we'refive to 10 years behind B2B and
B2C, maybe not this time.
Speaker 1 (26:15):
We're not, we're not, yeah, maybe not this time.
Speaker 2 (26:19):
In some areas we may be when you're talking about the
more true AI that you're pokingaround, but I'm talking about
the application of this stuff onuser desktop.
We're neck and neck witheverybody.
It's exciting, it's reallyexciting.
Speaker 1 (26:33):
It is.
Yeah.
Well, you know, what reallymakes me feel really good is to
have you here and see you.
It's been a few, quite a bit,since we last got a chance to
talk and, yeah, I'm really happyyou came by.
Speaker 2 (26:46):
Thanks, it was great to see you last night for a few
minutes and then, yeah, to chat.
This was a bonus.
Speaker 1 (26:52):
So here we are wrapping up at the Oleo booth,
the HR Tech Expo with WorkTech.
George Larac, Thank you verymuch.
Speaker 2 (27:00):
Thank you, thanks for having me and thanks everybody
for listening Take care bye-bye.
And we are back.
This is Shali Stekaral comingat you live from the HR Tech
Expo floor Powered today by theOlio booth.
We are broadcasting here in theOlio booth, taking names and
prisoners and just having a goodold time.
Our next guest is a long, longtime friend, old friend of mine,
(00:32):
george LaRocque, and George isfrom Workforce Tech, work Tech,
work Tech, work Tech.
How appropriate, hr Tech, workTech.
I think Work Tech is betterthan HR Tech because, like it's
about work.
Speaker 2 (00:47):
Right.
Speaker 1 (00:48):
Yeah, I like it.
So, george, why are you here atthe show and what's been your
impression so far?
Speaker 2 (00:55):
Well, I'm here at the show for a couple of reasons.
One, I do a lot of work at theshow, so I had a pre-conference
bringing together investors andstartups and scale-ups that this
show puts on, and I'm the chairof that.
I've been hosting the startuppitch fest for the last few days
(01:16):
.
Oh you introduce people andeverything.
Yeah, I do all that.
I provide a little insight on,you know, with the data that we
track and the trends that wetrack, all of the investment
coming into the space and youknow what are the trends, what
are people buying, et cetera.
You know in between.
And then I also come here tomeet people, see friends and
learn, see what's going on.
(01:37):
It's a great place to get apulse on the tech side of what's
happening.
Speaker 1 (01:40):
It is Definitely.
So what is your take on thepulse?
That's what's happening now.
Not the pitch fest, but likelooking at the booths and
walking around.
Yeah.
Speaker 2 (01:51):
So it seems like the two conversations are skills and
AI.
Of course we knew it.
Speaker 1 (01:58):
There's a bingo word for you.
We got ATS, we got AI and wegot CRM.
Bingo, yeah.
So skills meaning like skillingup, learning and development,
or like finding skills orontologies, yeah.
Speaker 2 (02:12):
Well, because everybody's talking about it,
yes, all of that and more.
You know.
So, depending on where they'recoming from, you've got even the
labor market data.
Folks are starting to look atskills market data.
So looking at, you know, notjust the jobs and geographic
location and comp by job, butskill, skills and comp by skill
(02:34):
and that sort of thing.
And then you've got, as you,you know, start from the very,
very, very, very top of thefunnel and go all the way into,
you know, post hire and learningand development.
It's, you know, skills, somarketplaces and upskilling and
re-skilling, and it's, you know,everybody's got an angle on
(02:56):
skills, got you Everybody.
Speaker 1 (02:59):
So I have two questions for you, all right,
based on the skills conversation.
The first of those is from apersonal perspective.
Recently started not recently,it lasts several months I
started using LinkedIn's.
I don't know skills catalog orontology so you're probably a
(03:20):
lot more familiar with this thanI am, but when you are posting
jobs to LinkedIn recruiterthrough the API and, for example
, like greenhouse publishes them, you know directly.
It used to not be very catalog,so if you had a software
engineer, it was just posted asa software engineer.
It might have maybe kind ofnarrowed down by industry.
(03:42):
But now they have this bigcatalog of skills that are
interrelated.
Are you familiar with what I'mtalking about?
Speaker 2 (03:49):
I'm familiar because I've seen the content on it.
But I haven't gone deep on that.
Speaker 1 (03:57):
So I started working with that and tweaking kind of
what.
Because there's two problemsthat happen very frequently with
job postings, and they have alot to do with classification.
One of them is the job titleitself.
If a company has a let's justcall it a cool job title, it
might be cool to the company, itmight set them apart, it might
(04:19):
be unique, but people don't knowwhat it means.
And so you end up withcandidates that don't understand
what the job is and they applyanyway because they're figured
out it might as well take achance and they're not qualified
because they're just not reallygetting the job title.
And then the opposite of thathappens you got people who are
in the job that you want to hirefor, but their job title is
completely different than yourcool title and so they don't see
(04:41):
themselves in it.
So they might think well, I'm asoftware engineer and this job
is for a you know, back-end datadeveloper or whatever, and
that's not me, so they don'tapply.
So the good ones don't applybecause the job title doesn't
resonate and the ones that don'tmatch apply because they don't
quite understand it.
So it's a problem in that youend up with just a higher
(05:04):
percentage of not reallyqualified candidates.
This has helped combat that, inthat these skills are matched,
and so I think the confidencelevel in applicants is
increasing Over time.
It'll get better.
Where a software engineer looksat the job and goes, ah,
go-lang, back-end data scientistdeveloper, that doesn't sound
(05:26):
right.
But everything else on the pageis me, so this has to be me.
So that's kind of.
What you're talking about isthe skills.
Speaker 2 (05:35):
That's one of the big use cases, both for hiring and
for internal mobility orupskilling.
So I did a case study last yearwith Manulife and they went and
Manulife is a big insurancecompany, global, and this was in
their engineering group.
So IT and development theproject took them from 1,400
(06:02):
roles to 40, so the biggest pullin the tent to your point was
change management, whereas, hey,I know your title is software
engineer but you're an analystand let me show you where we're
going with the organization.
Or we know you've got yourdepartment structured with these
titles, but we're moving inthis direction and some people
it wasn't a bad thing, not allchange is bad.
(06:23):
It was like, oh, you'reactually an engineer, good for
you.
That makes sense, but it was abig project.
Consultants travel around theworld.
That was a change managementprocess Once they got that to 40
, then maintenance of thatbecame ongoing.
Speaker 1 (06:40):
The flywheel fed itself.
Speaker 2 (06:42):
But now with technology.
So this is, I'm going to say,the promise of large language
models.
But this is where a lot ofpeople are heading.
So I think what would be coolin that example you gave is if,
dynamically, those skills wereto look at you.
I didn't the titles that youused.
But if it was like yeah, backin engineer versus software
(07:03):
developer versus software.
Speaker 1 (07:05):
So here's your title and here's this job, and you
could reflect that.
Speaker 2 (07:09):
Here's how your job now compares to this job
regardless of the job title isnot mapping that promises.
Speaker 1 (07:15):
That is what I would like to see happen organically,
in the background.
I'm manually mapping it becausethere are skills that are
required.
For example, if an organizationis not really language
dependent, like they'redeveloping in multiple languages
, the job posting will haveseveral.
You know Go Lang, java, python,et cetera.
(07:38):
What that tends to sometimestranslate into is that the
candidate must have all of theabove, which is really important
.
What it actually supposed tomean is we really don't care if
you're good at Go Lang, good atPython and good at Java, as long
as you're good at one of themand we would like to know which
one is your preference but we'lltake them.
(07:59):
But instead what people see isman, I love Go Lang, but I don't
know enough about it and Ihaven't used it at work, so I'm
not going to apply, and that'screating that.
The other question that I havefor you folks that are listening
George and I go back I don'tknow what to 10, 15 years at
least.
Oh, at least, at least it'slonger than that.
Speaker 2 (08:20):
Yeah, so I used to work for, we'll say 10, 15.
Yeah, something like that.
It's more than a decade.
I used to work for this companythat was a pioneer in the
distribution of jobs and Iliterally went after George.
Speaker 1 (08:33):
I recruited him.
You were a recommendation.
Did you know?
You were a recommendation fromsomeone, an angel investor that
I knew.
Speaker 2 (08:41):
I don't know if I ever told you that I don't think
you did yeah.
Speaker 1 (08:46):
So I called this friend, my friend of him, like
dude, I need somebody that'slike no nonsense, that's going
to come in and just hit theground running and can run sales
for a mature enterprise product.
And they're like, I can onlythink of one person that gave me
your name and I reached out andthe rest is history.
And so from there, when we bothleft that organization.
(09:09):
But you left and you startedyour own thing and you became
essentially like I want to say,like the Uber analyst, Like an
analyst is not a good title foryou.
You're the mega analyst of worktechnology.
Speaker 2 (09:22):
All right, thank you, sure Thanks.
Speaker 1 (09:24):
Because you're not just like out there writing an
article or two.
I mean, you're evaluatingproducts, You're advising
products, You're you know, andpeople depend on your opinion.
So two things related to that.
Number one literally puteverything I just said away for
a moment and pretend that youdon't know anything about it on
the industry and you walk around.
What is the most awesome booth?
No judgment about what they do.
(09:45):
What are?
Just the booth itself?
Oh wow, Because these boothsare darn cool and last year they
weren't this pretty.
Speaker 2 (09:53):
Yeah, that's a hard question for me because I've
been running that pitch fast, soI just told you, for like this
is you haven't walked around.
Speaker 1 (10:00):
I walked from there to unfair unfair question but
but they're like massive, likewe got.
We got baseball field sizeboots here.
Speaker 2 (10:08):
OK, I am, but here there's one that looks pretty
cool.
I am not endorsing this product.
Speaker 1 (10:15):
Not at all, it's just a boot.
But I mean like I am literallynot endorsing this product.
Speaker 2 (10:21):
It's that's going to get me in trouble.
It's Ripley.
Speaker 1 (10:25):
Oh yeah.
Speaker 2 (10:26):
And so they've got that the sphere I just walked by
and I'm going to stop.
It's like a sphere, and thenthey have one of those.
You see it at all the tradeshows now, but it's the sphere
around it.
That's kind of cool and it tiesto the I think thematically to
the sphere in Vegas.
But one of those camera thingsthat spins around you.
Speaker 1 (10:43):
We're in there, ok.
Speaker 2 (10:45):
But it just looks.
You know it's, it's, they havenice colors and it's and that
very looks very appealing.
But that's good yeah.
Speaker 1 (10:52):
So now put all your knowledge back on the table.
Speaker 2 (10:55):
OK, all right.
Speaker 1 (10:56):
And think about what you've seen or heard pitchfaster
, not that in the last week orso here.
Conversation wise, what is,what is a truly innovative?
It's a tough question because Ihaven't seen a whole lot of
innovation.
I've seen a lot of expansioninto other areas.
I've seen a lot of rebranding.
(11:17):
Honestly, I haven't seen a lotof new companies, except for
startup corridor Yep, and thoseare all you know new, but
unfortunately it's a small booththat you almost have to go to
their website to figure out whatthey do.
Speaker 2 (11:28):
Right, right.
Speaker 1 (11:29):
So all of what you've taken in conversations and
everything is there.
Is there something that isinnovative?
Dare I say, game changing?
Speaker 2 (11:40):
Yeah, in work tech it's, it's, it's been.
I'm very, I'm so jaded, charlie.
Speaker 1 (11:49):
I've been, I've been around for so long, and that's
what we're talking about isyou're the Uber animal, so it's
you know all this stuff.
Speaker 2 (11:56):
So it's, it's a tough question.
It's also not just about mebeing jaded, but also I'm going
to give you an answer, but Iit's also about you know, to
your point just to explain forit to stand out for you means
yeah.
Well, but just like to to.
So those pitches right that wewatched, like this year there
were like the finalists, likereally tight pitches and like
(12:17):
immediately you know what theydo, not that?
Not, I'm not saying that'swhere the innovation was, but
most of these booths I have noidea what they do when I stand
in front of it or even when theytalk to me, for four or five
minutes.
Speaker 1 (12:29):
I've noticed that, yeah.
Speaker 2 (12:30):
And so so it's, you need to peel back the the layers
a little bit.
There was one of the startupsthat and this is really
foundational, but I just wentover after that they were one of
the finalists Aragorn AI andwhat they, what they've built,
is remember back in the day wewould have these like hub and
(12:52):
spoke like connector models todo integrations across platforms
.
So they, what they've built, isthey've leveraged AI.
So they had to build, you know,integration I was talking about
this yesterday.
Yeah that they've leveraged AI.
Yeah, to do a lift and shiftthe ETL, yeah, yeah, yeah,
that's a really good I wastalking about that being a
(13:12):
really good application for AI,and that's what they're doing.
Speaker 1 (13:15):
That's yeah, they're doing a lot of higher them to be
that I'm going to use a reallyold word, middleware.
Yeah, yeah absolutely thetranslator between Product A and
.
Speaker 2 (13:24):
Product B yeah, they're like the next wave, like
beyond.
You know we're getting pastlike the.
I'll call it manual, but youknow the hub and spoke sort of
connector model, or like Zapier,or yeah, yeah and then what
that does for the ecosystem.
Here is you know you could beon Workday or you could be on
(13:45):
Seridian and you could haveGreenhouse or whatever you have
and the because it's more fluidright, the automation, the
opportunities to do interestingthings with your data and your
workflow is way better than ahard-coded, you know.
Speaker 1 (14:05):
Or you have to write a custom API or hire yeah, you
got to go back, yeah, we want todo this now.
Software engineering, do theintegration specifically for
your version of whatever.
Because everybody you know,ultimately, because not very
many maybe in your experiencebut in mine not very many
organizations will like kind ofuse the out-of-the-box.
Yeah, they'll, you know.
Speaker 2 (14:24):
Yeah, I'm going to.
I'll give you another startupanswer To Argonne.
Yeah yeah, eric Gorn, eric Gorn,Eric Gorn, ai.
Yep, then I can't remember thename of this company because it
was one of the 33 pitches, butthis is a creepy innovation.
They did a facial recognition.
Like you're sitting at yourdesk and this thing is on all
(14:44):
the time and doing facialrecognition and measuring
employee sentiment.
You know they're like claimingto measure, like you know,
mental wellness, and I was thehost so I didn't get to ask all
the questions that I had butthat's I think that's a really
(15:05):
scary place to go.
Speaker 1 (15:06):
It's innovative, I'm assuming that they're not
recording but analyzing, right,because that would be kind of
the so you know, first of all,recording massive amounts of
video and story is like kind ofuseless but also the privacy
infringement in that, you know,even if you're at work, the
security camera in your officemight record on a loop, but it's
(15:27):
not intended to recordpermanently.
So let's assume that they'reessentially analyzing, you know,
overall wellness based on yourfacial expressions or whatever
right?
So they're analyzing the data.
I can see the applications onthat, but I can also see how
that would be.
You know, you know mybackground.
(15:48):
My undergraduate is inintercultural and non-verbal
communications and to me one ofthe biggest mistakes that we
make, generally speaking ashumans, is that we misinterpret
what other people are thinkingor feeling based on what our
particular proclivities might beRight.
So, for example, I'll just giveyou a really easy example you
(16:11):
and I being from a Westernculture, to us a smile that is a
nice big smile, is sort oftoothy.
Right, you open up wine, youshow your pearly whites and
that's a very friendly, warmsmile, where there are cultures
where if you show your teeth isan act of open aggression.
So my concern with thisintelligence is that if it
(16:35):
doesn't detect toothy smiles,it's gonna say, oh, this
person's unhappy, when in factit might be that they're quite
happy.
They're just culturallypredisposed to not, you know,
have a toothy smile, or it'sjust one thing.
And then you've gotneurodivergent people and you've
got people who have facial tics.
I mean, I'm thinking about allthe possible misinterpretations
(16:56):
that could happen there?
Speaker 2 (16:57):
Yeah, I think there.
Yeah, I wasn't surprised to seeit, given you know.
Speaker 1 (17:03):
AI, all the AI hype and everything.
Speaker 2 (17:06):
Here's another one.
You might like.
This one collab work thatsounds familiar.
They do non-traditional postingof jobs and recruitment, so
what they call the hidden talent.
So they're not going to jobboards, they're not going to
social media, they're gettingjobs out into Slack channels and
(17:30):
they're getting into otherdestinations where the talent
that you're looking for is notnecessarily there.
Speaker 1 (17:39):
It's not about jobs, Hidden talent, not intentionally
hidden, but yeah and that's thephrase they use.
Speaker 2 (17:47):
You know, they've got AI, you know, and everybody's
got AI in the mix.
I'm not sure how they're usingit, but it was pretty, I think,
with all the pressure on finding, finding talent through new
channels.
Speaker 1 (18:01):
Just tab sources.
Speaker 2 (18:02):
I think you know I think they'll definitely get
some traction.
Speaker 1 (18:05):
For sure I can think of applications for that in very
niche areas where you don'treally have people looking on
the traditional job boardsbecause they know that there's
not a lot of jobs for them.
Speaker 2 (18:17):
Yeah, I mean I go back when I was recruiting.
I don't know if you rememberthe Usenet news group days, yup,
usenet, I mean I got a lot ofgood candidates out of Usenet.
Yeah, they'd sit there on theMac and, just you know, put the
job every day, set aside an hourto get the jobs.
That was the old boards.
Speaker 1 (18:37):
Yeah, oldest one of those would probably be like the
well, from back in the, youknow you dial up days, but you
know.
So you've mentioned AI a coupletimes and I know it's one of
our bingo words, yeah.
Yeah, I would love for you toexplain what, in your perception
, is really called AI, because Iknow, and you probably know
(19:02):
that I know that it's not AI,right, it's not really true AI.
So they call it generative AIbecause they want to qualify it
and you know, basically justlarge language models and it's
data science and it's machinelearning.
But what, in your estimation,as an analyst, looking at this
industry deeply, what you know?
They're calling it AI.
They used to call it semanticsearch.
They used to call it, you know,whatever, but what is it really
(19:25):
?
Speaker 2 (19:26):
Yeah, and I think you bring up a really good point.
If you know it could be machinelearning, it could be.
There are so many flavors ofwhat's what they call it yeah,
the AI is like the catch allterm.
Yeah, I think at an event likethis, I think that's probably a
good way to look at it.
Speaker 1 (19:44):
It's a catch all term .
Speaker 2 (19:46):
It's more of a catch all term.
I think what generative AI didwas open up everybody's minds to
the possibilities, and a lot ofwhat we have out there is
aspirational.
It's, you know, superaspirational right now.
I don't think, you know, Idon't think there's anything on
(20:08):
this floor that's going to passthe Turing test yeah.
So you know, there's Rightright.
Speaker 1 (20:14):
So you know the traditional definition of a true
artificial intelligence that'sable to extract subtext from
context and can make sense of avariety of different things.
You know that's not what we'relooking at here.
That's what I was kind ofwondering is if you have a sense
of what, in general, they'rereferring to when they say AI.
(20:35):
Is it, generally speaking, theyou know large language models,
or is it something else?
Because the video processingone they're doing, that's not
large language models.
Right, and the you know theother ones that I've seen on
here that are more around likechat bots.
That's also not really a largelanguage model.
There's a gentleman that was atthe podcast earlier talking
(20:58):
about recruiting sourcing toolthat they're using.
He's a data scientist and whatthey've created is not a large
language model, the type thatyou would see in chat GPT, for
example, which is unhomogenousdata.
His is specifically homogenous.
So you used to be a boardmember of the HRXML consortium.
Speaker 2 (21:17):
Right right.
Speaker 1 (21:20):
Literally, what is the HRXML equivalent these days?
You know that doesn't existanymore.
No, no.
But what is the?
Are these language models?
Essentially a new way of sortof categorizing data into some
sort of Homogenous pool so youcan later use it Like.
These are conversations, theseare resumes, these are, you know
(21:40):
, is that what we're talkingabout?
Because machines can do thatreally well.
Speaker 2 (21:47):
I think some of them are talking about that.
So like an eight-fold theybefore they ever came out with
anything generative they put alot of effort into large
language models.
Their founders are, you know,scientists from Google.
Speaker 1 (22:02):
Data science.
Yeah, so that's an ontology, ataxonomy.
Yeah, exactly that's whatsemantic search used to be.
Speaker 2 (22:08):
Right, and so you look there and I would say
there's an example.
They're doing that.
Now, your average recruitingtool that is talking about AI is
probably got a field or an areain the app where it'll write an
email for you.
It's using generative AI andit's using not their own large
(22:32):
language model.
Speaker 1 (22:33):
No, they're borrowing from somebody else's which is
scary because it becomeseverybody is using the same
language.
Speaker 2 (22:40):
It's very bland.
Speaker 1 (22:41):
Right, Very middle of the road, which is not very
convincing yeah now others are.
Speaker 2 (22:50):
You know, back to the skills conversation or if
you're thinking about somethingthat's a little more sensitive,
like when you're talking aboutjob descriptions, those are
going external and those areexposed anyhow so you don't have
to be as careful with those,and it's good to capture more of
a global sort of learn from theworld at large, but if you're
(23:15):
looking at anything close to orinvolving PII or sensitive data,
you need to have enough data tohave your own language model,
and it's got to be private,otherwise you're exposing a lot
of risk.
So there are a lot of and don'tget me wrong a tool that makes
(23:37):
your email sound better, thatengages candidates better,
that's effective.
Speaker 1 (23:43):
And that's really helpful.
Most of us are not greatwriters.
Speaker 2 (23:47):
Yeah, so that's great , but that's becoming table
stakes, like really fast.
Speaker 1 (23:52):
That's my point.
Speaker 2 (23:53):
That is table stakes.
Now, like every tool that I useI'll say almost every tool that
I use on my desk has somethinglike that.
Speaker 1 (24:02):
Yeah, there's something you know it's an
assistive thing.
Yeah, it's clippy, it's clippy.
Speaker 2 (24:05):
Yeah, it's just yeah, remember clippy.
Yeah, I do remember clippy,it's clippy.
Speaker 1 (24:09):
Yeah, okay, so that makes me feel a little bit
better about kind of the stateof the industry, because then we
have a lot of different placesto go from here rather than just
everybody calling it, becauseit used to be all everybody
called it semantic search and wewrote that white paper and then
people stopped talking about itand now it's this other thing
and you know, probably gonna bethe buzz for a while.
People stopped talking about it.
(24:30):
It kind of gives me the youknow, social media, Way back
when we started talking aboutsocial media and we were talking
about, ooh, social media, nowit's just media man.
I mean it's not social, it'sjust, it is part of the mix.
So AI is going to be somewherealong the line part of the mix.
Speaker 2 (24:47):
Yeah, I think that's an excellent analogy and I
actually use that example a lot.
And so we used to talk aboutsocial recruiting and it's just
recruiting now and it took about10 years to really we had to
become mainstream and we hadsuccess and we had all along the
(25:07):
way, but it took seven to 10years.
I'll say before you really,mobile recruiting, right, yeah,
right, mobile, yeah.
Speaker 1 (25:14):
So this will be AI in five years.
It'll just be recruiting, yeah,I think yeah.
Speaker 2 (25:20):
So this cycle, I think, is more like five years
on the money and another thingthat should make you feel good.
So I just in the recruitingspace, specifically recruitment
marketing space, I've been on abig journey, talking to
everybody, looking at becausethere are a lot of threats
around with what's coming here,and I ventured outside.
(25:45):
I went and said, well, what arethey doing in B2C marketing and
in B2B marketing?
So I tapped into thought leaderstuff there.
They sound just like us.
They have the conversation wejust had.
We could have been talking aboutmarketing platforms and CRMs in
B2C in B2B and so with that,when I say it should make you
(26:05):
feel good, we're all in it.
Yeah, well, we're on par.
So we always used to say we'refive to 10 years behind B2B and
B2C, maybe not this time.
Speaker 1 (26:15):
We're not, we're not, yeah, maybe not this time.
Speaker 2 (26:19):
In some areas we may be when you're talking about the
more true AI that you're pokingaround, but I'm talking about
the application of this stuff onuser desktop.
We're neck and neck witheverybody.
It's exciting, it's reallyexciting.
Speaker 1 (26:33):
It is.
Yeah.
Well, you know, what reallymakes me feel really good is to
have you here and see you.
It's been a few, quite a bit,since we last got a chance to
talk and, yeah, I'm really happyyou came by.
Speaker 2 (26:46):
Thanks, it was great to see you last night for a few
minutes and then, yeah, to chat.
This was a bonus.
Speaker 1 (26:52):
So here we are wrapping up at the Oleo booth,
the HR Tech Expo with WorkTech.
George Larac, Thank you verymuch.
Speaker 2 (27:00):
Thank you, thanks for having me and thanks everybody
for listening Take care bye-bye.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!