September 5, 2024 • 38 mins
James Mackey, CEO of a leading RPO provider, SecureVision, and Elijah Elkins, CEO of Avoda, a highly rated global recruiting firm, co-host Kwal’s Co-Founder, David Tell, in our special series on AI for Hiring.
David shares how Kwal’s generative AI voice agent is evolving recruitment for staffing firms and RPOs by automating initial screenings and interviews. The conversation highlights the challenges of bias in AI-driven evaluations and the need for companies to address these concerns.
0:00 Generative AI voice agents for hiring
14:08 Regulatory concerns and AI bias detection
22:00 Future trends in AI hiring solutions
27:00 Broader implications and market trends
Thank you to our sponsor, SecureVision, for making this show possible!
Our host James Mackey
Follow us:
https://www.linkedin.com/company/82436841/
#1 Rated Embedded Recruitment Firm on G2!
https://www.g2.com/products/securevision/reviews
Thanks for listening!
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Hey everyone, Welcome to the show.
This is your host, James Mackey.
Today, I'm joined by my co-host, Elijah Elkins.
Elijah, what's up, man?
How are you?
Hey Daman, how are you Good?
And we're also joined by DavidTell, who's the founder and CEO
of Qual.
David, how are you doing today?
Speaker 3 (00:16):
Doing great.
Thanks so much for having me onthis.
Speaker 1 (00:18):
Yeah, absolutely.
And just looking at theLinkedIn profile of your company
AI Breakthrough that conductsinterviews at warp speed and
sounds just like a human theperfect recall and adaptive
questioning.
No detail gets overlooked.
So that sounds pretty cool andwe're looking forward to
learning more about that.
Do you want to start off justwith an introduction of yourself
and what brought you tofounding Qual?
Speaker 3 (00:39):
Sure, qual is a generative AI voice agent for
hiring.
Prior to Qual, I helped startanother company called TaskAid
and during that process it's acurated marketplace sold to
enterprises effectively, and wehad to hire a lot of people at
scale globally in about 70different countries, and it was
a pretty challenging environment.
Just with the time changes andI always wanted to provide a
(01:01):
really great experience tocandidates but also difficult to
scale that quickly, and sothat's the genesis for Qual
really is just like livingthrough that pain, and so it's
exciting to see that it's built,it's working, it's operating
and customers are finding valuewith it.
Speaker 1 (01:14):
Yeah, that's awesome.
That's really great to hear.
And now you've been in business.
Now what like better part of Ithink you said what a year or
two.
Speaker 3 (01:20):
Yeah, about a year and a half.
Speaker 1 (01:23):
Okay, cool.
Can you tell us more about yourcustomers?
What types of companies you'reworking with?
Speaker 3 (01:27):
Yes, typically working with staffing firms and
RPOs, and typically we'reworking with companies that are
dealing with roles from anywherefrom customer support to
nursing type of roles orphysician assistant type of
roles, so across a wide range ofindustries, everything again
from healthcare to construction.
Speaker 1 (01:46):
Okay.
So when you say like,partnering with, like RPOs and
staffing companies, like what isit Like?
You're there.
Essentially your product isreplacing the screening call for
, so the initial, when RPOsreach out, like how does that
work exactly?
Speaker 3 (02:00):
Yeah, so primarily it's going to be the first call
is what we would say.
In some cases that first callis the entire interview,
particularly like lightindustrial is a great example
where there's no secondaryinterview.
So you could say that's thewhole interview and kit and
caboodle.
In that type of industry, otherroles, and depending upon the
speed of the need of the hire,it might be the first call that
(02:22):
happens, but then there might bea secondary call.
Really depends upon thecustomer or the use case.
Speaker 1 (02:27):
Yeah, that makes sense.
So my question, more so on theRPO side, is when we plug into a
customer, we're representingtheir brand.
So when we screen candidates,it's we're talking as if we're
an intern part of their internalteam.
We have their email address andwe're fully embedded, so to
speak.
So how does that work?
Is it essentially, essentiallylike this kind of like white
(02:48):
label situation, which it's likethe rpo makes it part of the
package that they're essentiallyselling to a customer?
Speaker 3 (02:57):
yeah, it depends on the setup of the rpo.
Some have their own middlewarethat they work with.
Others will spin up, forexample, like a bullhorn
instance for each end client.
Some will use their ownmiddleware that they work with.
Others will spin up, forexample, like a Bullhorn
instance for each end client.
Some will use their own excuseme, the clients and ATA system.
So it really depends on the RPOsetup, but we can support any
three of those types of usecases.
And then in terms of thedialogue and like the tone, the
(03:18):
flavor, all those things thatare possible now with generative
AI, that's where we like honein with the customer right.
A lot of people think that Ithink this is a misnomer about
enterprise sales and generativeAI just generally, which is that
it's just like a product, justslide it in.
But I think that's actuallylike a massive misnomer.
I think a lot of times you'respending a lot of time with the
(03:38):
service provider or the customer, whoever that might be really
like.
What is their brand?
What is their dialogue?
What do they want these callsto sound like?
And every customer is slightlydifferent.
Speaker 1 (03:48):
Yeah, for sure.
Oh yeah, I'm looking forward todiving more into that.
Elijah, do you have any highlevel product questions you want
to cover?
Speaker 2 (03:58):
Yeah, yeah, I'm curious, David.
So we've interviewed so far Benat BrightHire and Mark at
Pillar, so both in thatinterview intelligence space
right, they're probably dealinga lot more with like tech
companies or knowledge workertype jobs.
I'm curious is Qual doing muchwith knowledge worker screening
(04:20):
calls or do you foresee Qualbeing able to support those more
knowledge worker type screeningcalls at some point?
Speaker 3 (04:28):
Yeah, we already do support knowledge worker types
of calls.
But just to give you a sense ofthe entire industry, I think
50% of the market is lightindustrial alone from just a
pure hiring standpoint withinthe US, of course, just by the
sheer numbers.
We're going to do more of that,especially specifically like
the contingent workforce space.
What do you think about?
Yeah, can we do knowledgeworkers and we?
(04:49):
And do we do them?
The answer is yes, unprivately,we do them, but in terms of
just pure volume and where themarket is, a lot of this is
going to be in the category ofthings that are not necessarily
what we, what you and I mightconsider knowledge work.
And then, as you move up thatchain right, you'll see that a
lot of jobs are in other areasthat are I would call
semi-skilled, maybe likecustomer support, things like
(05:12):
that, and I think that's likeanother, like step function.
So when you just think aboutthe overall market, I think
that's the biggest predicator ofwhere we will fit in.
Speaker 2 (05:21):
Nice.
And how does it so?
How does it work?
Speaker 3 (05:38):
Take us through, say I'm a candidate for a sales job
or you could say, lightindustrial staffing job.
I spent years working in lightindustrial staffing when I was
late teens, so take candidatedepends on the company.
But if they are, some companieshave an open pipeline and
others will have the applyfiltration system, typically
using something like an ATS,which might have their own
version of that, or somethinglike a Daxstra or a Texkernel
which was just acquired byBullhorn, and apply that sort of
(05:58):
filter match score, then sendthe candidates to us.
So we're typically not in thegame of filtration or
identification.
We're in the business ofexecution and thinking about
tools like BrightHire, forexample, or MetaView would be
like another one.
Those are great tools.
We're not competing with thosetools.
Those tools are really like youcan imagine us as probably like
(06:20):
an earlier step of that process.
But then, right when youactually want somebody
internally to actually have acall with somebody to see
whether or not they're going tofit within your team, which
would be a huge thing is where Ithink those tools fit in.
And interestingly they havetheir own competitors, right
Microsoft and some of the moregeneral note-taking tools.
We're not really in that spaceor that sort of preliminary
(06:42):
space.
Speaker 1 (06:43):
Do you see the customers that basically, upon
applying their candidates, areimmediately prompted to engage
in this screening call?
Or I was just wondering howfrequent that process is,
because, if one of the thingsthat I've been thinking through
is, just think about the sheernumber of applicants, a lot of
what I do is in the techindustry, so of course we're
(07:05):
seeing a ton of applicants, butwhether it may be in light
industrial, there's a ton ofapplicants too.
Is that accurate?
Just real quick?
In light industrial, do youtypically see high levels of
inbound applicants?
Speaker 3 (07:14):
It really depends on the company's internal strategy
around whether or not they wantto spend marketing dollars.
Typically, what we see so it'smore of an internal strategy,
but so it depends on thecustomer is what we see.
We also support source callstoo.
Just to be like clear, if it'snot somebody, that's, if it's
somebody already internally inthe database or somebody that
they sourced, we do have achronic extension that allows
people to basically addcandidates.
(07:36):
That way we're not just a pureapply system, if that sort of
gives you a little bit morecontext.
Speaker 1 (07:41):
Yeah, so that was actually like.
Like if somebody there's youwant to get to all the
applicants, for instance,whether you have a few or a ton,
and maybe it just makes senseto okay, anybody who applies can
immediately go through thisprocess without scheduling an
additional screening call.
I'm just curious if you seeadoption from customers that, or
(08:02):
do they typically want thathuman touch and review?
Do they want to speak to likewhere, what is?
What are you seeing when itcomes to that automation around
that first step?
Speaker 3 (08:13):
Yeah, that's where I was going with that earlier part
was this idea of open funnel,which is you don't apply a
filter, Like you have candidatesthat come in.
We do a lot of that wherepeople are saying hey, either
the resumes are not of highquality because of the industry.
That it is so light industrialis sort of notorious.
Like a lot of the resumes arenot very good in light
industrial.
So in those cases, right, it'smore about speed and timing and
(08:36):
all those types of things to geta candidate in the door working
right, and so you don't want toforego an opportunity to filter
somebody out that might be likea hidden gem for that type of
role.
So that's where we see, to yourpoint, this concept of open
funnel.
The other area we see that inright now is for more of these,
like I say, mid-skill to maybeknowledge worker type of jobs.
You have a flood of applicants,right.
(08:56):
So a friend of mine who runs atech company literally had the
other day, within a couple hours, 1,200 applicants, right, he
doesn't want to spend time andresources doing that, he just
wants to say hey, like a lot ofthese resumes I can't tell like
they've clearly been throughChatGPT and I don't know if this
person actually has theexperience.
So that's another sort of usecase that we serve.
Is you just have a lot ofapplicants, a lot of the resumes
(09:19):
, look good.
How do you like actuallywhittle through that pile pretty
effectively?
So it's pretty hard forrecruiters to even do, and
that's another way that we seethis sort of manifesting to your
original question I.
Speaker 1 (09:30):
One thing I noticed on your website is you have this
I see the evaluation part waslike candidate summary and it's
like a match score.
So what's interesting is when wewere speaking with a pillar
market pillar and then ben Benover at BrightHire, they
actually stayed away from doingmatches or actual like
evaluation or stack ranking,which to me it's.
(09:52):
That is great.
I understand the risks they'rein and it shouldn't be, we
shouldn't be filtering peopleout.
But evaluating match does seemlike that's.
It seems like an intuitive, uh,important does seem like that's
.
It seems like an intuitive, uh,important, somewhat obvious
part of what should be part ofthese generative ai solutions
and well within currentcapabilities to some extent.
(10:12):
So I'm just wondering it's likeI was curious to get your
thoughts on that, if you likewhy did you decide to do that
and if you have any insight onwhy you think maybe bright
higher or pillar or some ofthese other solutions don't do
that.
I'd be curious to get yourthoughts there too, why you
think maybe BrightHire or Pillaror some of these other
solutions don't do that.
Speaker 3 (10:27):
I'd be curious to get your thoughts there too.
Yeah, I think maybe one insighton why they don't do it is
because they're maybe morecompeting with the note-taking
space and so scoring may not beas maybe intuitive in that space
.
Maybe you're really just tryingto identify what happened on a
call and so maybe that's justthe way, the natural way, the
industry has adopted.
I think In our space, wherewe're applying more of a
(10:48):
secondary filter often, orsometimes a primary filter
through a voice call, I thinkthe score probably maybe matters
more a little bit for customersbecause they're using again, in
those cases you might only,have, like, with a bright hire,
three or four people you mightbe interviewing, right.
So you know, the delta onhaving a human sort of do a
rubric or whatever they're doingis not it's not a large lift,
(11:10):
but we were talking about likethousands, right?
Potentially, I think that'swhere a score is helpful,
because it's effectivelyproviding some sort of filter.
That being said, this scoringis something that's optional, so
if a customer asks us to turnit off, we can do that, so we
effectively just then startproducing a summary, right?
So same thing as brain hire.
The challenge is you're going tohave to go through each one of
(11:32):
them individually determined.
So that takes time.
So it's still less timeintensive than trying to place
100 calls with your existingrecruiting staff or internal
staff, whatever type of businessyou are.
But if you think about it froma time-saving standpoint, that's
where I think the scoringhappens, and I think one of the
reasons is people are nervousabout this is obviously the
(11:53):
legal implications aroundscoring.
But there's great guidance outthere about this whole area and
I think, generally speaking, Ithink regulators are not dumb.
I think they understand thatit's not exactly like a perfect
world, right, like when you havelarge volumes.
This is like what you need todo to be effective, but it's
(12:15):
like how you do it, and I thinksome companies are just going to
be farther along on the riskcurve, and so I think that's
what you're going to see.
I think you're going to see alot of vendors either play very
risk adverse others say maybeit's not as important and then
others which are like us, whichare going to be more adaptive to
sort of customer needs and alsogeographies too, because these
rules are not uniform.
Speaker 1 (12:35):
I was wondering.
That's where it's too.
It's like these enterprisecompanies there, it seems from a
compliance standpoint and arisk standpoint it would seem to
make sense as to why that wouldbe a blocker to them, and maybe
there's just generally loweradoption to new aspects of
process.
So maybe we'll just start tosee more of it in the next six
to 12 months.
But I'm not familiar with theregulatory restrictions and
(12:59):
requirements for AI essentiallydoing matching, but I'm
wondering specifically what themain concerns are.
I know there has to betransparency around, like how
I'm assuming how the ai is doingthe matching, but I don't.
I think one thing that might bea blind spot for me.
I don't know if you either ofyou guys know this, but it's.
I still don't fully understandwhy ai is considered like
(13:22):
riskier than having people do it.
I agree on, I agree so, like Iwas talking with daniel chate,
founder of Greenhouse, and wetalked about this and he was
like, because it's like the riskof bias at such large scale.
Speaker 2 (13:37):
So he was like if the system's doing it wrong, it
could do it wrong with athousand people or something
right, yeah, but every one ofyour recruiters and HR people
have biases too, and at leastwith the AI you might be able to
know what some of thoseinherent biases are that are
going to be multiplied across ahuge number of candidates versus
.
You don't know all the biasesof your team and they're all
(13:58):
going to be different, so youalmost have you know what I mean
Like unfair biases, because youdon't know what they are.
With AI, I'm assuming you mightbe able to identify with some
of what those are.
I don't know what you think,david.
Speaker 3 (14:13):
I think you really touched on something that's
pretty interesting, which is howmany companies you think today,
using like a bright hire or oneof these tools, takes all those
calls and then runs themthrough an algorithm to
determine whether or not therewas bias that actually happened
on the call.
My guess is probably not a lot,but the beautiful thing about
(14:34):
where we are with large languagemodels is that a lot of
companies that are building inthe space they use the LLM as
the judge right.
These things are prettyeffective at judging what
happened on a call right anddetermining if there is bias.
So what's interesting isthere's so much fear of the bias
from the AI, but are mostcompanies even taking the
recordings of their actualrecruiters today, identifying
(14:56):
bias when it's not happening?
And I guess the answer isprobably there's another aspect
to that.
Speaker 1 (15:01):
I don't think companies are necessarily coming
out and saying but just havingoperating with HR tech companies
as well, even outside ofrecruiting technology, sometimes
providing analytics around riskand compliance if it's not
legally required.
I'm not an attorney, but Iwould say that there could be
(15:25):
some concerns there.
It's okay.
When more of this is brought tolight, then we're in more of a
position at a risk in terms ofneeding to immediately address
it and solve it.
So maybe I have no idea whatI'm talking about here, but I'm
just trying to think out loud.
I wonder if, like part of it'slike we're hitting our
compliance requirements Right.
Do we really want a nice reportoutlining more than what we
(15:48):
need to provide that could beaudited, or something of that
nature?
Does this make any sense?
I'm curious to get yourthoughts there.
Speaker 3 (15:56):
Yeah, maybe a little bit of background too.
I'm actually.
I'm an attorney, I'm licensedin California and Florida oh
nice and I've worked in acompliance department before.
So I could think if there'sanybody that might be able to
answer this question, I thinkmaybe I'm the right guy.
Oh, that's awesome.
I'm so happy I asked on thispodcast.
(16:16):
Look, I think companies are notalways good at determining, like
, what the risk is.
There are certain things whereit's just super high risk, right
, Like you're doing atransaction right and the guys
that you run.
That's super high risk type oftransaction.
And I think, for as much ascompanies have tried to develop
risk profiles and things likethat over the years and it often
involves insurance and riskmanagement, sometimes like
treasury it's effectively like ahedge within a business,
(16:38):
especially a large business Ithink that it makes sense why
they're risk adverse.
It's like a lot of it is.
They don't know, they hearthings they say, oh my God,
worst case scenario.
But when you actually look atthe entire risk profile of what
a business has going on, Iactually would put this the way
low end of the risk curve theregulators in the EU would say,
(16:58):
David, no, you're completelywrong, but I would tell you
they're completely wrong.
Speaker 1 (17:02):
It's just weird.
What's the big?
I don't know.
So I was just wondering.
For instance, brighthire andPillar, I was wondering if one
of the reasons like they're notdoing the evaluation pieces just
simply because they'retargeting mid-market, upper mid
to enterprise customers and theyjust don't want to, they don't
(17:23):
want to make it harderessentially in their sales
process, and I didn't discussthis directly with them, but I'm
wondering if that has somethingto do with it.
Speaker 3 (17:31):
so yeah, I think if you think about like the
enterprise sales processtypically, let's just say that
used to be nine months.
I think at enterprise hr techthat's probably way north of 12
months now 10 years they don'twant to risk a sale, so they're
going to do whatever it takes tomake sure the buyer feels
satisfied.
And the buyer has to go througha lot of hoops right.
Typically in those orgs there'sa procurement person, as you
(17:52):
guys know, and then there's, youknow, it security team that
needs to sign off and thentypically legal is involved in
the process, and those could belike treacherous waters to
navigate to get a sale acrossthe line.
So I think you're ontosomething Like.
I think that's probably anadaptation to a lot of those
like processes, and I wouldn'tbe surprised if BrandHire goes
(18:15):
through multiple committees toget a sale done right.
Like it might be, they mighthave to talk to 16 to 20 people
right over the course of thatperiod of time, find the
champions and then navigatethrough the entire process.
I think it's a function of theenvironment that the sales
process, and then therefore theproduct, is adapting.
Speaker 2 (18:32):
I'm trying to think do you see anything with how the
tool is developing further withthe advancements in AI?
Did you guys, once GPT kind ofcame out, really see a lot more
advancement and really launchoff from there, or were you
working with other tools beforethat?
I'm just curious how thatimpacted the trajectory of the
(18:53):
business.
Speaker 3 (18:54):
So the first company that I helped start is called
TaskAid, but it's still around.
It's a Series B company.
The search engine there isactually developed on
transformer architecture, sopre-GPT.
So it's called BERT.
It's very well known.
It was Google published it andmade it open source, and so we
basically built our entiresearch function around that, so
we're very familiar withtransformers.
The company really startedaround time of GPT-3, prior to
(19:18):
chat GPT explosion, which is 3.5.
So, yeah, this was an evolution, right, like in terms of is the
technology mature enough?
Can it do all the things thatwe want it to do?
Because there still are a lotof vendors in the market today
that are pre-generative AI, thatare based on the old piece
workflow builder right, if this,that it still has some AI
(19:40):
elements in terms of it's likeit's transcribing, which is
certainly part of AI.
It's doing some naturallanguage processing, some intent
detection, but it's not what wewould call today a generative
AI driven system or an agenticsystem, which is like a further
step up past this generativesystem.
Speaker 2 (19:59):
And have you seen with your customer base as they
implement Qual and start gettingvalue from it and let's say
they are using the score, as youmentioned earlier, are they
reducing the number ofassessments like later in the
funnel, like if they're using,like a metal or some sort of
(20:19):
assessment platform, personalityor what have you?
Was that providing enough valueon that first call that they're
reducing those assessments downthe funnel?
Speaker 3 (20:35):
Definitely so.
In the case where there is anassessment, for example, like a
classic one, also might be likeEnglish language assessment,
right, others which can be moretechnical assessments, clearly
it's acting as a sort offiltration step.
So it's pretty interestingbecause there's a lot of
companies that are in that space, like iMocha and Glider, and
they're going to be.
You're not going to send athousand candidates to those.
They're just typically tooexpensive or too time consuming.
You're not going to want to putcandidates through that.
So, yeah, qual acts as a greatsort of again filter step before
(20:59):
you would send somebody to anassessment.
But not every role requires anassessment.
Speaker 1 (21:04):
Cool, yeah, I was curious.
So, david, I'm curious to getyour thoughts on how you're
thinking about the future forQual over the next six months
here, if you happen to know whatyou're going to be building
over a year out.
But just say what feedback areyou hearing from customers?
What additional functionalitydo they?
What use cases do they needhelp solving?
(21:25):
What's next?
Speaker 3 (21:27):
Yeah, I think the biggest thing that we're going
to move towards is we'reobviously a voice solution.
We do SMS and email andvoicemail, but I think the real
future is like what I would callintelligent agent, omni channel
, and so we're driving towardsthat.
It's not just about a callhappening and being
bidirectional or text beingbidirectional.
It's like how do you tie allthese things together in a
(21:49):
really intelligent way?
There's a lot of oldersolutions that sort of claim
that they do this, but againthey're stringing things
together and like duct tapingthem.
But I think the real future islike how do you merge all these
channels so it becomes like asingle channel and the agent is
like the back plane behind thatright, this idea that you have
access to the logic, the, thememory and the tools and
(22:12):
functions that agents can haveright to execute different
things, whether it's on a callor whether it's a text.
So I think that's the futurethat we're going to drive to and
I think that resonates with alot of customers that we talk to
.
Speaker 1 (22:25):
Yeah, it makes a lot of sense and I know you said
most of your business is on the.
It sounds like a lot of yourcustomers on the RPO staffing
side.
Do you see that growing alsointo like more in-house
companies or what's your?
Speaker 3 (22:39):
Yeah, I think so.
I think so.
If you look at RPO, rpo istypically hired by an enterprise
and staffing is typically hiredby enterprise.
But I think it has to makesense.
Do they have this sort ofvolumes?
We're probably not a solutionfor a two-person type of shop
and also from our standpoint, wereally like working with RPO
and staffing.
We think it's pretty dynamicspace.
We get to see a lot ofdifferent things working with
(23:01):
them because they typically havemore than one customer.
So it's a great channel for usand I think we'll spend a lot of
time in this space.
But as we sort of grow and allthose types of things, obviously
engaging directly with theenterprise is something that we
will take on if there's mutualinterest.
Speaker 1 (23:16):
Yeah, makes a lot of sense and suppose, do you like
just talking about the future ofAI and hiring, maybe even a
little less related to the usecase that you're helping your
customers solve right now, tothe use case that you're helping
your customers solve right now?
Curious to get your thoughts onother product plays, other use
(23:37):
cases that are technicallyfeasible or will be technically
feasible to solve for the nextyear.
Curious to just get yourgeneral thoughts on our space,
so to speak, and what you thinkis coming over the next 12, 18
months.
Speaker 3 (23:49):
Biggest thing is going to be outside of what
we're doing is going to be inthe browser-based space, browser
agent space specifically.
You can think about that asexecuting a query right, and
then the agent is going to golog into your LinkedIn right, go
, click through profiles,determine whether or not this
person's a good fit and thengenerate a report for you or
like a list.
I think we're going to startseeing that there's been a lot
(24:11):
of companies out there that arebuilt chrome extensions and
stuff like that, or they workwith people data labs, as an
example, and these like large,like list building companies
similar to, like, zoom info, butI think the real thing is that
those are typically like stale,oftentimes right those databases
are.
They have to get constantlyupdated.
A lot of that time that'shappening from humans.
But I think this idea of theseagents that can actually act on
(24:34):
your behalf, like on yourbrowser, like execute the task
of identification of candidatesit's going to be really big.
The curiosity I have theregenerally is like how is indeed,
or these types of companies ina or linkedin, yeah?
Speaker 2 (24:47):
are they gonna block it?
How are they gonna?
Speaker 3 (24:49):
block it.
Um, yeah, they've clearly donethat with, like, the scrapers.
This isn't typically a scraper,right, it's a lot different
type of technology and so Ithink there's like some
questions around that and, yeah,probably we'll probably see
some lawsuits at some point forwhether or not people really
have permission to do that ifthey.
But it's interesting becauseit's like an extension of you,
right, these agents, right likethe whole concept of the agent
(25:10):
is you're the principal, that'sthe agent, just like hiring a
contractor, right, and it's sortof the same idea, except it's a
digital entity that's doingthat.
And I think that's where thisis going to get really
interesting.
It's like, how many of thesetasks can we offload today that
were really painful?
And I think it's really greatfor recruiters that actually
want to be connecting withcandidates, right Like selling
(25:31):
candidates on roles.
So I don't look at it as likethere's probably a lot of fear
about AI for, like, recruitersand sourcers and a lot of folks,
but I think if you're in thebusiness of selling roles and
culture like that's not going tochange, like an AI is not going
to do that.
But I think that recruitersthat really also understand the
nuances of the business, right,like the end business, really
are engaged and understand likethat piece of it.
(25:53):
There's a moat there, right,and AI is not going to come in
and replace that, but these sortof ancillary tasks that people
are conditioned to doing becauseit's just the way it's been
done, I think that's where AI isgoing to be really disruptive,
and I would expect that we wouldsee either two things happen.
One is you're going to have asmaller amount of recruiters or
(26:14):
you'll have the same amount ofrecruiters, but the service
level quality will just rise,and I don't think that's a bad
thing.
Speaker 1 (26:22):
Yeah, I mean, I think that's spot on.
I think the role of recruitersis going to change drastically.
One of the reasons we're doingthis series and one of several
is, I think talent acquisitionleaders, probably more now than
ever, are getting veryinterested in the technologies
that are available in the market, which is funny because it's
happening at a time where theyreally don't have the budget.
But I think that the besttalent acquisition executives
(26:46):
are going to ultimately be onesthat understand how to leverage
these different products and,essentially, the right balance
of technology and people and howto most effectively, based on
the nuances of their industryand their business, have those
things that integrate, thosepieces integrate.
That's going to be the coreskill set, and this is probably
(27:08):
the case of most departmentswithin a company, right, but
it's going to be the core skillset.
That a great, and this isprobably the case of most
departments within a company,right, but it's going to change.
It's not going to be good enoughto just understand process.
You're going to have tounderstand the nuance of, based
on your business, yourcandidates, your hiring goals,
what should technology solve andwhat should people solve, and
what parts of the process andand whatnot, and how to
(27:31):
incorporate those things in away where you still get the
visibility you need from areporting perspective and the
integrations you need and tomake things run smoothly and, of
course, as we discussed, likethe compliance aspects as well.
So it's going to get.
I think it's just the role isgoing to, of course, get like
more, more technical than it'sprobably ever been before.
(27:53):
It's not going to just be like,okay, slap on an ATS, it's okay
.
We really got to be thoughtfulof our process and where we're
leveraging tech and how.
Speaker 3 (28:00):
Yeah, I think people are scared of this.
I think again because it's anew level of automation.
They haven't seen it before.
They used to be the person thatwould spend 30 minutes building
a job description.
Now it takes about 10 secondsand you can imagine every ATS is
building this if it doesn'talready have it.
We already have it.
It didn't take very long for usto develop.
So I think those are thelow-hanging fruit, but there's
(28:21):
quite a bit of fruit everywherein this tree.
It will shake out, certainly,but we've had automation for a
long time.
Email automation is somethingthat we've had automation for a
long time.
Right, like email automation issomething that we've had for
years.
Right, recruiters hit 10different checkboxes and the 10
candidates get the rejectionemail at the same time, or it's
automated completely.
We've had this automation.
I think it's just now at thepoint where we can generate
(28:42):
personalization.
We can generate things at theedge of the node rather than the
sort of like centralizedfunction or centralized matter
that really makes generative AIand agents again, which is like
a next level past that, likesuper unique.
It'll be a very large learningcurve, I think, an adoption
curve.
I think the nature of any sortof transitional change is that
there's going to be, like earlyadopters, people in the middle,
(29:04):
the sort of late.
But this is not one of thosetrends where I would want to be
late, for I would if I was aleader.
I would be evaluating thingslike constantly, like multiple
products a week, because thespace is changing really fast.
It's still going to be acompetitive war for great talent
.
Always, or even in areas whereyou might think it's like a blue
collar role, some of those arethe most competitive, right?
(29:26):
You wouldn't?
You would be surprised, right?
There's only so many peoplethat can do this role and 30
miles outside of Wichita, kansas, that's a really specific role.
So I think we're just going tosee this adoption happen.
But the companies that reallymove fast are going to be,
they're going to get amazingyield very quickly and they'll
realize the value.
(29:46):
And I think the laggards arereally going to struggle, like
really going to struggle, andthat's why it's always great
working with people that can seethe future For sure.
Speaker 1 (29:55):
Look, this has been really, really insightful.
I appreciate you coming on theshow.
Is there anything else that youwant to share with the audience
about your product or anythingelse?
We've got a couple of minuteshere.
Speaker 3 (30:09):
Anything else you want to share.
Yeah, I think the biggest thingto share is just whether you're
evaluating like our technologyor a different vendor.
It doesn't matter the space,maybe it's not even recording,
maybe it's sales or whatever thetypes of tools you're looking
at.
Look at when these companiesstarted always right, I think
you know.
If you're looking at companiesthat started three to five years
ago, right, typically whatyou're going to see is they're
adding things on.
But what's really nice aboutthe newer companies that are
(30:31):
coming out is that they're allAI native companies, right.
They're built from the groundup being AI native and the
experiences will be reallydifferent.
So, even though the end productmight look similar, the
marketing might look similar,the actual experience when you
actually dive into theseproducts are very different, and
the only way to do that iseither getting on a demo call,
right?
(30:51):
So I would say to leaders thatare listening to this podcast be
open-minded, right.
There's a lot of great newsoftware that's coming to the
market.
That's completely changing theentire paradigms, from what
we're seeing and being aware,meeting the founders, ceos of
these companies, I think it's agreat thing because you'll get,
even if you don't decide topurchase, you get a sense of
what is coming and you can beknowledgeable about that and
(31:13):
share that with other people.
And I think, again, going backto what I said earlier, you
don't want to have your headburied in the sand, right?
You've heard that expressionbefore.
This is a biggest revolution, Ithink, in my opinion, since the
birth of the internet and priorto that, probably
semiconductors, which is so youdon't want to miss out on this
trend and you don't want to beunder your head buried in the
sand, even if you're not goingto purchase.
Go out there, be curious, meetpeople.
(31:34):
It will only benefit you andyour organization as you go
through these motions, and it'stime consuming, but it's
important.
Speaker 1 (31:42):
Yeah, I totally agree , Elijah.
Do you have any other finalthoughts?
Speaker 2 (31:50):
No, I think David's right.
I think you need to keep apulse on the market to be a good
TA leader, and I think we'regoing to see that more people
may be in like recruiting ops,like we saw with, like sales ops
and rev ops.
Thinking about the whole techstack, I think the tools those
AI native tools that David'smentioning are going to make the
TA leaders like tech stack alot more yeah, a lot more like
powerful and intelligent andwe'll be able to do things that
(32:14):
we weren't able to do previously.
But we're definitely still alot of catching up to do
compared to the tools that, likesales leaders have at their
disposal to grow revenue, togrow the company and find great
people to join the team.
There's a lot of great toolsout there, but, yeah, probably
not near as many as there are inlike the sales world.
Speaker 3 (32:32):
Yeah, literally sales .
If you think TA is busy in HRtech, sales is just like
incredible, like how many toolscome into market every day.
Speaker 1 (32:41):
I feel like a lot of hopefully not in the songs,
maybe that sounds mean, but alot of them went out of business
.
It was just too many.
Lot of them went out ofbusiness, it was just too many.
There was too many Me Tooproducts in sales.
Like how many differentproducts that do the same thing?
Do you need, good Lord, somecategories?
You'd have like 30, 40 players,like early stage companies,
essentially doing the same thing.
Speaker 2 (33:03):
The pricing's interesting too.
Sorry, david, I was just goingto throw out there.
Sales tools often are a lotmore affordable than some of the
recruiting and HR tech toolsthat I've seen.
Probably a whole differentpodcast episode.
But the pricing differencemaybe that's like you're saying,
james, the number ofcompetitors.
But an email sequencing tool,for example, you could get for,
(33:37):
let's say, $20 to $50 a monthfor a license, some of the
recruiting versions to do emailsequencing or anywhere from what
150 to 300 per license and theyoften don't have the same
feature set either, like theyhave fewer features.
Speaker 3 (33:42):
So, anyways, what were you gonna say, david?
Yeah, I've definitely comeacross orgs where it's
especially smaller orgs maybeyou guys have where, like
hubspot is like their recruitingtool, it's like they've taken a
sales tool or crm and likeHubSpot is like their recruiting
tool, it's like they've taken asales tool or a CRM and like
have tried to like build aroundthat.
It's sort of interesting.
But I think there's just a lotof orgs that, in my mind,
probably don't want to deal withthe complexity of again duct
taping things.
So I think that's probably whyyou still see higher rates and I
(34:05):
don't think that TA is not acompetitive market or recruiting
tech is not a competitivemarket.
It definitely is.
But yeah, I think sales is justit's revenue driving function
and I hate to say that as an HRperson, right, every TA leader
would probably like yell at meright at the top of their lungs
no, we're a revenue generatingfunction and it's true Like HR
is responsible for findingpeople and engaging people and
(34:25):
retaining people that are inrevenue driving positions, and
it is all about the people.
There's no doubt about it.
But I think, if you think aboutfrom a sales perspective, if
you're selling a tool, mostpeople want to sell it to a
revenue generating function.
Right, it's pretty, pretty muchthat's why that is, and so
you're always going to see themost amount of tools that are
(34:46):
going to probably drive revenueor give you insights into
revenue, which is why you see somany of these tools that are to
your point, like RevOps toolsand analytics tools.
It's just like the numbers arejust like staggering in terms of
how many there are out there.
Speaker 1 (34:59):
Yeah, it's just, I don't know how you expect to
scale an incredibly successfulcompany if you had 20
competitors doing the same thing, but I don't know.
Anyways, yeah, the marketcorrected, hopefully is doing
the same thing, but I don't know.
Anyways, yeah, the marketcorrected.
Hopefully it's not as saturatednow.
But anyways, we digress.
We're coming up on time here.
David, thank you so much forjoining us today.
We really appreciate you comingon with us on the show.
Speaker 3 (35:18):
Yeah, it was a lot of fun.
I really appreciate you guyshaving me Great dialogue and I
look forward to stayingconnected, and our company
website is wwwqualai for thosethat are listening.
And yeah, I'd love to just talkto people in the space and see
how we can help you.
Speaker 1 (35:33):
Yeah, that sounds great and we're going to put
everybody tuning in some notesand descriptions so you'll have
the website, david's LinkedInprofile, all that good stuff.
And, by the way, yeah, david,if there's anything you want us
to drop in the show notes, justlet us know.
Speaker 3 (35:46):
For sure Awesome Cool .
Speaker 1 (35:50):
Thank.
Hey everyone, Welcome to the show.
This is your host, James Mackey.
Today, I'm joined by my co-host, Elijah Elkins.
Elijah, what's up, man?
How are you?
Hey Daman, how are you Good?
And we're also joined by DavidTell, who's the founder and CEO
of Qual.
David, how are you doing today?
Speaker 3 (00:16):
Doing great.
Thanks so much for having me onthis.
Speaker 1 (00:18):
Yeah, absolutely.
And just looking at theLinkedIn profile of your company
AI Breakthrough that conductsinterviews at warp speed and
sounds just like a human theperfect recall and adaptive
questioning.
No detail gets overlooked.
So that sounds pretty cool andwe're looking forward to
learning more about that.
Do you want to start off justwith an introduction of yourself
and what brought you tofounding Qual?
Speaker 3 (00:39):
Sure, qual is a generative AI voice agent for
hiring.
Prior to Qual, I helped startanother company called TaskAid
and during that process it's acurated marketplace sold to
enterprises effectively, and wehad to hire a lot of people at
scale globally in about 70different countries, and it was
a pretty challenging environment.
Just with the time changes andI always wanted to provide a
(01:01):
really great experience tocandidates but also difficult to
scale that quickly, and sothat's the genesis for Qual
really is just like livingthrough that pain, and so it's
exciting to see that it's built,it's working, it's operating
and customers are finding valuewith it.
Speaker 1 (01:14):
Yeah, that's awesome.
That's really great to hear.
And now you've been in business.
Now what like better part of Ithink you said what a year or
two.
Speaker 3 (01:20):
Yeah, about a year and a half.
Speaker 1 (01:23):
Okay, cool.
Can you tell us more about yourcustomers?
What types of companies you'reworking with?
Speaker 3 (01:27):
Yes, typically working with staffing firms and
RPOs, and typically we'reworking with companies that are
dealing with roles from anywherefrom customer support to
nursing type of roles orphysician assistant type of
roles, so across a wide range ofindustries, everything again
from healthcare to construction.
Speaker 1 (01:46):
Okay.
So when you say like,partnering with, like RPOs and
staffing companies, like what isit Like?
You're there.
Essentially your product isreplacing the screening call for
, so the initial, when RPOsreach out, like how does that
work exactly?
Speaker 3 (02:00):
Yeah, so primarily it's going to be the first call
is what we would say.
In some cases that first callis the entire interview,
particularly like lightindustrial is a great example
where there's no secondaryinterview.
So you could say that's thewhole interview and kit and
caboodle.
In that type of industry, otherroles, and depending upon the
speed of the need of the hire,it might be the first call that
(02:22):
happens, but then there might bea secondary call.
Really depends upon thecustomer or the use case.
Speaker 1 (02:27):
Yeah, that makes sense.
So my question, more so on theRPO side, is when we plug into a
customer, we're representingtheir brand.
So when we screen candidates,it's we're talking as if we're
an intern part of their internalteam.
We have their email address andwe're fully embedded, so to
speak.
So how does that work?
Is it essentially, essentiallylike this kind of like white
(02:48):
label situation, which it's likethe rpo makes it part of the
package that they're essentiallyselling to a customer?
Speaker 3 (02:57):
yeah, it depends on the setup of the rpo.
Some have their own middlewarethat they work with.
Others will spin up, forexample, like a bullhorn
instance for each end client.
Some will use their ownmiddleware that they work with.
Others will spin up, forexample, like a Bullhorn
instance for each end client.
Some will use their own excuseme, the clients and ATA system.
So it really depends on the RPOsetup, but we can support any
three of those types of usecases.
And then in terms of thedialogue and like the tone, the
(03:18):
flavor, all those things thatare possible now with generative
AI, that's where we like honein with the customer right.
A lot of people think that Ithink this is a misnomer about
enterprise sales and generativeAI just generally, which is that
it's just like a product, justslide it in.
But I think that's actuallylike a massive misnomer.
I think a lot of times you'respending a lot of time with the
(03:38):
service provider or the customer, whoever that might be really
like.
What is their brand?
What is their dialogue?
What do they want these callsto sound like?
And every customer is slightlydifferent.
Speaker 1 (03:48):
Yeah, for sure.
Oh yeah, I'm looking forward todiving more into that.
Elijah, do you have any highlevel product questions you want
to cover?
Speaker 2 (03:58):
Yeah, yeah, I'm curious, David.
So we've interviewed so far Benat BrightHire and Mark at
Pillar, so both in thatinterview intelligence space
right, they're probably dealinga lot more with like tech
companies or knowledge workertype jobs.
I'm curious is Qual doing muchwith knowledge worker screening
(04:20):
calls or do you foresee Qualbeing able to support those more
knowledge worker type screeningcalls at some point?
Speaker 3 (04:28):
Yeah, we already do support knowledge worker types
of calls.
But just to give you a sense ofthe entire industry, I think
50% of the market is lightindustrial alone from just a
pure hiring standpoint withinthe US, of course, just by the
sheer numbers.
We're going to do more of that,especially specifically like
the contingent workforce space.
What do you think about?
Yeah, can we do knowledgeworkers and we?
(04:49):
And do we do them?
The answer is yes, unprivately,we do them, but in terms of
just pure volume and where themarket is, a lot of this is
going to be in the category ofthings that are not necessarily
what we, what you and I mightconsider knowledge work.
And then, as you move up thatchain right, you'll see that a
lot of jobs are in other areasthat are I would call
semi-skilled, maybe likecustomer support, things like
(05:12):
that, and I think that's likeanother, like step function.
So when you just think aboutthe overall market, I think
that's the biggest predicator ofwhere we will fit in.
Speaker 2 (05:21):
Nice.
And how does it so?
How does it work?
Speaker 3 (05:38):
Take us through, say I'm a candidate for a sales job
or you could say, lightindustrial staffing job.
I spent years working in lightindustrial staffing when I was
late teens, so take candidatedepends on the company.
But if they are, some companieshave an open pipeline and
others will have the applyfiltration system, typically
using something like an ATS,which might have their own
version of that, or somethinglike a Daxstra or a Texkernel
which was just acquired byBullhorn, and apply that sort of
(05:58):
filter match score, then sendthe candidates to us.
So we're typically not in thegame of filtration or
identification.
We're in the business ofexecution and thinking about
tools like BrightHire, forexample, or MetaView would be
like another one.
Those are great tools.
We're not competing with thosetools.
Those tools are really like youcan imagine us as probably like
(06:20):
an earlier step of that process.
But then, right when youactually want somebody
internally to actually have acall with somebody to see
whether or not they're going tofit within your team, which
would be a huge thing is where Ithink those tools fit in.
And interestingly they havetheir own competitors, right
Microsoft and some of the moregeneral note-taking tools.
We're not really in that spaceor that sort of preliminary
(06:42):
space.
Speaker 1 (06:43):
Do you see the customers that basically, upon
applying their candidates, areimmediately prompted to engage
in this screening call?
Or I was just wondering howfrequent that process is,
because, if one of the thingsthat I've been thinking through
is, just think about the sheernumber of applicants, a lot of
what I do is in the techindustry, so of course we're
(07:05):
seeing a ton of applicants, butwhether it may be in light
industrial, there's a ton ofapplicants too.
Is that accurate?
Just real quick?
In light industrial, do youtypically see high levels of
inbound applicants?
Speaker 3 (07:14):
It really depends on the company's internal strategy
around whether or not they wantto spend marketing dollars.
Typically, what we see so it'smore of an internal strategy,
but so it depends on thecustomer is what we see.
We also support source callstoo.
Just to be like clear, if it'snot somebody, that's, if it's
somebody already internally inthe database or somebody that
they sourced, we do have achronic extension that allows
people to basically addcandidates.
(07:36):
That way we're not just a pureapply system, if that sort of
gives you a little bit morecontext.
Speaker 1 (07:41):
Yeah, so that was actually like.
Like if somebody there's youwant to get to all the
applicants, for instance,whether you have a few or a ton,
and maybe it just makes senseto okay, anybody who applies can
immediately go through thisprocess without scheduling an
additional screening call.
I'm just curious if you seeadoption from customers that, or
(08:02):
do they typically want thathuman touch and review?
Do they want to speak to likewhere, what is?
What are you seeing when itcomes to that automation around
that first step?
Speaker 3 (08:13):
Yeah, that's where I was going with that earlier part
was this idea of open funnel,which is you don't apply a
filter, Like you have candidatesthat come in.
We do a lot of that wherepeople are saying hey, either
the resumes are not of highquality because of the industry.
That it is so light industrialis sort of notorious.
Like a lot of the resumes arenot very good in light
industrial.
So in those cases, right, it'smore about speed and timing and
(08:36):
all those types of things to geta candidate in the door working
right, and so you don't want toforego an opportunity to filter
somebody out that might be likea hidden gem for that type of
role.
So that's where we see, to yourpoint, this concept of open
funnel.
The other area we see that inright now is for more of these,
like I say, mid-skill to maybeknowledge worker type of jobs.
You have a flood of applicants,right.
(08:56):
So a friend of mine who runs atech company literally had the
other day, within a couple hours, 1,200 applicants, right, he
doesn't want to spend time andresources doing that, he just
wants to say hey, like a lot ofthese resumes I can't tell like
they've clearly been throughChatGPT and I don't know if this
person actually has theexperience.
So that's another sort of usecase that we serve.
Is you just have a lot ofapplicants, a lot of the resumes
(09:19):
, look good.
How do you like actuallywhittle through that pile pretty
effectively?
So it's pretty hard forrecruiters to even do, and
that's another way that we seethis sort of manifesting to your
original question I.
Speaker 1 (09:30):
One thing I noticed on your website is you have this
I see the evaluation part waslike candidate summary and it's
like a match score.
So what's interesting is when wewere speaking with a pillar
market pillar and then ben Benover at BrightHire, they
actually stayed away from doingmatches or actual like
evaluation or stack ranking,which to me it's.
(09:52):
That is great.
I understand the risks they'rein and it shouldn't be, we
shouldn't be filtering peopleout.
But evaluating match does seemlike that's.
It seems like an intuitive, uh,important does seem like that's
.
It seems like an intuitive, uh,important, somewhat obvious
part of what should be part ofthese generative ai solutions
and well within currentcapabilities to some extent.
(10:12):
So I'm just wondering it's likeI was curious to get your
thoughts on that, if you likewhy did you decide to do that
and if you have any insight onwhy you think maybe bright
higher or pillar or some ofthese other solutions don't do
that.
I'd be curious to get yourthoughts there too, why you
think maybe BrightHire or Pillaror some of these other
solutions don't do that.
Speaker 3 (10:27):
I'd be curious to get your thoughts there too.
Yeah, I think maybe one insighton why they don't do it is
because they're maybe morecompeting with the note-taking
space and so scoring may not beas maybe intuitive in that space
.
Maybe you're really just tryingto identify what happened on a
call and so maybe that's justthe way, the natural way, the
industry has adopted.
I think In our space, wherewe're applying more of a
(10:48):
secondary filter often, orsometimes a primary filter
through a voice call, I thinkthe score probably maybe matters
more a little bit for customersbecause they're using again, in
those cases you might only,have, like, with a bright hire,
three or four people you mightbe interviewing, right.
So you know, the delta onhaving a human sort of do a
rubric or whatever they're doingis not it's not a large lift,
(11:10):
but we were talking about likethousands, right?
Potentially, I think that'swhere a score is helpful,
because it's effectivelyproviding some sort of filter.
That being said, this scoringis something that's optional, so
if a customer asks us to turnit off, we can do that, so we
effectively just then startproducing a summary, right?
So same thing as brain hire.
The challenge is you're going tohave to go through each one of
(11:32):
them individually determined.
So that takes time.
So it's still less timeintensive than trying to place
100 calls with your existingrecruiting staff or internal
staff, whatever type of businessyou are.
But if you think about it froma time-saving standpoint, that's
where I think the scoringhappens, and I think one of the
reasons is people are nervousabout this is obviously the
(11:53):
legal implications aroundscoring.
But there's great guidance outthere about this whole area and
I think, generally speaking, Ithink regulators are not dumb.
I think they understand thatit's not exactly like a perfect
world, right, like when you havelarge volumes.
This is like what you need todo to be effective, but it's
(12:15):
like how you do it, and I thinksome companies are just going to
be farther along on the riskcurve, and so I think that's
what you're going to see.
I think you're going to see alot of vendors either play very
risk adverse others say maybeit's not as important and then
others which are like us, whichare going to be more adaptive to
sort of customer needs and alsogeographies too, because these
rules are not uniform.
Speaker 1 (12:35):
I was wondering.
That's where it's too.
It's like these enterprisecompanies there, it seems from a
compliance standpoint and arisk standpoint it would seem to
make sense as to why that wouldbe a blocker to them, and maybe
there's just generally loweradoption to new aspects of
process.
So maybe we'll just start tosee more of it in the next six
to 12 months.
But I'm not familiar with theregulatory restrictions and
(12:59):
requirements for AI essentiallydoing matching, but I'm
wondering specifically what themain concerns are.
I know there has to betransparency around, like how
I'm assuming how the ai is doingthe matching, but I don't.
I think one thing that might bea blind spot for me.
I don't know if you either ofyou guys know this, but it's.
I still don't fully understandwhy ai is considered like
(13:22):
riskier than having people do it.
I agree on, I agree so, like Iwas talking with daniel chate,
founder of Greenhouse, and wetalked about this and he was
like, because it's like the riskof bias at such large scale.
Speaker 2 (13:37):
So he was like if the system's doing it wrong, it
could do it wrong with athousand people or something
right, yeah, but every one ofyour recruiters and HR people
have biases too, and at leastwith the AI you might be able to
know what some of thoseinherent biases are that are
going to be multiplied across ahuge number of candidates versus
.
You don't know all the biasesof your team and they're all
(13:58):
going to be different, so youalmost have you know what I mean
Like unfair biases, because youdon't know what they are.
With AI, I'm assuming you mightbe able to identify with some
of what those are.
I don't know what you think,david.
Speaker 3 (14:13):
I think you really touched on something that's
pretty interesting, which is howmany companies you think today,
using like a bright hire or oneof these tools, takes all those
calls and then runs themthrough an algorithm to
determine whether or not therewas bias that actually happened
on the call.
My guess is probably not a lot,but the beautiful thing about
(14:34):
where we are with large languagemodels is that a lot of
companies that are building inthe space they use the LLM as
the judge right.
These things are prettyeffective at judging what
happened on a call right anddetermining if there is bias.
So what's interesting isthere's so much fear of the bias
from the AI, but are mostcompanies even taking the
recordings of their actualrecruiters today, identifying
(14:56):
bias when it's not happening?
And I guess the answer isprobably there's another aspect
to that.
Speaker 1 (15:01):
I don't think companies are necessarily coming
out and saying but just havingoperating with HR tech companies
as well, even outside ofrecruiting technology, sometimes
providing analytics around riskand compliance if it's not
legally required.
I'm not an attorney, but Iwould say that there could be
(15:25):
some concerns there.
It's okay.
When more of this is brought tolight, then we're in more of a
position at a risk in terms ofneeding to immediately address
it and solve it.
So maybe I have no idea whatI'm talking about here, but I'm
just trying to think out loud.
I wonder if, like part of it'slike we're hitting our
compliance requirements Right.
Do we really want a nice reportoutlining more than what we
(15:48):
need to provide that could beaudited, or something of that
nature?
Does this make any sense?
I'm curious to get yourthoughts there.
Speaker 3 (15:56):
Yeah, maybe a little bit of background too.
I'm actually.
I'm an attorney, I'm licensedin California and Florida oh
nice and I've worked in acompliance department before.
So I could think if there'sanybody that might be able to
answer this question, I thinkmaybe I'm the right guy.
Oh, that's awesome.
I'm so happy I asked on thispodcast.
(16:16):
Look, I think companies are notalways good at determining, like
, what the risk is.
There are certain things whereit's just super high risk, right
, Like you're doing atransaction right and the guys
that you run.
That's super high risk type oftransaction.
And I think, for as much ascompanies have tried to develop
risk profiles and things likethat over the years and it often
involves insurance and riskmanagement, sometimes like
treasury it's effectively like ahedge within a business,
(16:38):
especially a large business Ithink that it makes sense why
they're risk adverse.
It's like a lot of it is.
They don't know, they hearthings they say, oh my God,
worst case scenario.
But when you actually look atthe entire risk profile of what
a business has going on, Iactually would put this the way
low end of the risk curve theregulators in the EU would say,
(16:58):
David, no, you're completelywrong, but I would tell you
they're completely wrong.
Speaker 1 (17:02):
It's just weird.
What's the big?
I don't know.
So I was just wondering.
For instance, brighthire andPillar, I was wondering if one
of the reasons like they're notdoing the evaluation pieces just
simply because they'retargeting mid-market, upper mid
to enterprise customers and theyjust don't want to, they don't
(17:23):
want to make it harderessentially in their sales
process, and I didn't discussthis directly with them, but I'm
wondering if that has somethingto do with it.
Speaker 3 (17:31):
so yeah, I think if you think about like the
enterprise sales processtypically, let's just say that
used to be nine months.
I think at enterprise hr techthat's probably way north of 12
months now 10 years they don'twant to risk a sale, so they're
going to do whatever it takes tomake sure the buyer feels
satisfied.
And the buyer has to go througha lot of hoops right.
Typically in those orgs there'sa procurement person, as you
(17:52):
guys know, and then there's, youknow, it security team that
needs to sign off and thentypically legal is involved in
the process, and those could belike treacherous waters to
navigate to get a sale acrossthe line.
So I think you're ontosomething Like.
I think that's probably anadaptation to a lot of those
like processes, and I wouldn'tbe surprised if BrandHire goes
(18:15):
through multiple committees toget a sale done right.
Like it might be, they mighthave to talk to 16 to 20 people
right over the course of thatperiod of time, find the
champions and then navigatethrough the entire process.
I think it's a function of theenvironment that the sales
process, and then therefore theproduct, is adapting.
Speaker 2 (18:32):
I'm trying to think do you see anything with how the
tool is developing further withthe advancements in AI?
Did you guys, once GPT kind ofcame out, really see a lot more
advancement and really launchoff from there, or were you
working with other tools beforethat?
I'm just curious how thatimpacted the trajectory of the
(18:53):
business.
Speaker 3 (18:54):
So the first company that I helped start is called
TaskAid, but it's still around.
It's a Series B company.
The search engine there isactually developed on
transformer architecture, sopre-GPT.
So it's called BERT.
It's very well known.
It was Google published it andmade it open source, and so we
basically built our entiresearch function around that, so
we're very familiar withtransformers.
The company really startedaround time of GPT-3, prior to
(19:18):
chat GPT explosion, which is 3.5.
So, yeah, this was an evolution, right, like in terms of is the
technology mature enough?
Can it do all the things thatwe want it to do?
Because there still are a lotof vendors in the market today
that are pre-generative AI, thatare based on the old piece
workflow builder right, if this,that it still has some AI
(19:40):
elements in terms of it's likeit's transcribing, which is
certainly part of AI.
It's doing some naturallanguage processing, some intent
detection, but it's not what wewould call today a generative
AI driven system or an agenticsystem, which is like a further
step up past this generativesystem.
Speaker 2 (19:59):
And have you seen with your customer base as they
implement Qual and start gettingvalue from it and let's say
they are using the score, as youmentioned earlier, are they
reducing the number ofassessments like later in the
funnel, like if they're using,like a metal or some sort of
(20:19):
assessment platform, personalityor what have you?
Was that providing enough valueon that first call that they're
reducing those assessments downthe funnel?
Speaker 3 (20:35):
Definitely so.
In the case where there is anassessment, for example, like a
classic one, also might be likeEnglish language assessment,
right, others which can be moretechnical assessments, clearly
it's acting as a sort offiltration step.
So it's pretty interestingbecause there's a lot of
companies that are in that space, like iMocha and Glider, and
they're going to be.
You're not going to send athousand candidates to those.
They're just typically tooexpensive or too time consuming.
You're not going to want to putcandidates through that.
So, yeah, qual acts as a greatsort of again filter step before
(20:59):
you would send somebody to anassessment.
But not every role requires anassessment.
Speaker 1 (21:04):
Cool, yeah, I was curious.
So, david, I'm curious to getyour thoughts on how you're
thinking about the future forQual over the next six months
here, if you happen to know whatyou're going to be building
over a year out.
But just say what feedback areyou hearing from customers?
What additional functionalitydo they?
What use cases do they needhelp solving?
(21:25):
What's next?
Speaker 3 (21:27):
Yeah, I think the biggest thing that we're going
to move towards is we'reobviously a voice solution.
We do SMS and email andvoicemail, but I think the real
future is like what I would callintelligent agent, omni channel
, and so we're driving towardsthat.
It's not just about a callhappening and being
bidirectional or text beingbidirectional.
It's like how do you tie allthese things together in a
(21:49):
really intelligent way?
There's a lot of oldersolutions that sort of claim
that they do this, but againthey're stringing things
together and like duct tapingthem.
But I think the real future islike how do you merge all these
channels so it becomes like asingle channel and the agent is
like the back plane behind thatright, this idea that you have
access to the logic, the, thememory and the tools and
(22:12):
functions that agents can haveright to execute different
things, whether it's on a callor whether it's a text.
So I think that's the futurethat we're going to drive to and
I think that resonates with alot of customers that we talk to
.
Speaker 1 (22:25):
Yeah, it makes a lot of sense and I know you said
most of your business is on the.
It sounds like a lot of yourcustomers on the RPO staffing
side.
Do you see that growing alsointo like more in-house
companies or what's your?
Speaker 3 (22:39):
Yeah, I think so.
I think so.
If you look at RPO, rpo istypically hired by an enterprise
and staffing is typically hiredby enterprise.
But I think it has to makesense.
Do they have this sort ofvolumes?
We're probably not a solutionfor a two-person type of shop
and also from our standpoint, wereally like working with RPO
and staffing.
We think it's pretty dynamicspace.
We get to see a lot ofdifferent things working with
(23:01):
them because they typically havemore than one customer.
So it's a great channel for usand I think we'll spend a lot of
time in this space.
But as we sort of grow and allthose types of things, obviously
engaging directly with theenterprise is something that we
will take on if there's mutualinterest.
Speaker 1 (23:16):
Yeah, makes a lot of sense and suppose, do you like
just talking about the future ofAI and hiring, maybe even a
little less related to the usecase that you're helping your
customers solve right now, tothe use case that you're helping
your customers solve right now?
Curious to get your thoughts onother product plays, other use
(23:37):
cases that are technicallyfeasible or will be technically
feasible to solve for the nextyear.
Curious to just get yourgeneral thoughts on our space,
so to speak, and what you thinkis coming over the next 12, 18
months.
Speaker 3 (23:49):
Biggest thing is going to be outside of what
we're doing is going to be inthe browser-based space, browser
agent space specifically.
You can think about that asexecuting a query right, and
then the agent is going to golog into your LinkedIn right, go
, click through profiles,determine whether or not this
person's a good fit and thengenerate a report for you or
like a list.
I think we're going to startseeing that there's been a lot
(24:11):
of companies out there that arebuilt chrome extensions and
stuff like that, or they workwith people data labs, as an
example, and these like large,like list building companies
similar to, like, zoom info, butI think the real thing is that
those are typically like stale,oftentimes right those databases
are.
They have to get constantlyupdated.
A lot of that time that'shappening from humans.
But I think this idea of theseagents that can actually act on
(24:34):
your behalf, like on yourbrowser, like execute the task
of identification of candidatesit's going to be really big.
The curiosity I have theregenerally is like how is indeed,
or these types of companies ina or linkedin, yeah?
Speaker 2 (24:47):
are they gonna block it?
How are they gonna?
Speaker 3 (24:49):
block it.
Um, yeah, they've clearly donethat with, like, the scrapers.
This isn't typically a scraper,right, it's a lot different
type of technology and so Ithink there's like some
questions around that and, yeah,probably we'll probably see
some lawsuits at some point forwhether or not people really
have permission to do that ifthey.
But it's interesting becauseit's like an extension of you,
right, these agents, right likethe whole concept of the agent
(25:10):
is you're the principal, that'sthe agent, just like hiring a
contractor, right, and it's sortof the same idea, except it's a
digital entity that's doingthat.
And I think that's where thisis going to get really
interesting.
It's like, how many of thesetasks can we offload today that
were really painful?
And I think it's really greatfor recruiters that actually
want to be connecting withcandidates, right Like selling
(25:31):
candidates on roles.
So I don't look at it as likethere's probably a lot of fear
about AI for, like, recruitersand sourcers and a lot of folks,
but I think if you're in thebusiness of selling roles and
culture like that's not going tochange, like an AI is not going
to do that.
But I think that recruitersthat really also understand the
nuances of the business, right,like the end business, really
are engaged and understand likethat piece of it.
(25:53):
There's a moat there, right,and AI is not going to come in
and replace that, but these sortof ancillary tasks that people
are conditioned to doing becauseit's just the way it's been
done, I think that's where AI isgoing to be really disruptive,
and I would expect that we wouldsee either two things happen.
One is you're going to have asmaller amount of recruiters or
(26:14):
you'll have the same amount ofrecruiters, but the service
level quality will just rise,and I don't think that's a bad
thing.
Speaker 1 (26:22):
Yeah, I mean, I think that's spot on.
I think the role of recruitersis going to change drastically.
One of the reasons we're doingthis series and one of several
is, I think talent acquisitionleaders, probably more now than
ever, are getting veryinterested in the technologies
that are available in the market, which is funny because it's
happening at a time where theyreally don't have the budget.
But I think that the besttalent acquisition executives
(26:46):
are going to ultimately be onesthat understand how to leverage
these different products and,essentially, the right balance
of technology and people and howto most effectively, based on
the nuances of their industryand their business, have those
things that integrate, thosepieces integrate.
That's going to be the coreskill set, and this is probably
(27:08):
the case of most departmentswithin a company, right, but
it's going to be the core skillset.
That a great, and this isprobably the case of most
departments within a company,right, but it's going to change.
It's not going to be good enoughto just understand process.
You're going to have tounderstand the nuance of, based
on your business, yourcandidates, your hiring goals,
what should technology solve andwhat should people solve, and
what parts of the process andand whatnot, and how to
(27:31):
incorporate those things in away where you still get the
visibility you need from areporting perspective and the
integrations you need and tomake things run smoothly and, of
course, as we discussed, likethe compliance aspects as well.
So it's going to get.
I think it's just the role isgoing to, of course, get like
more, more technical than it'sprobably ever been before.
(27:53):
It's not going to just be like,okay, slap on an ATS, it's okay
.
We really got to be thoughtfulof our process and where we're
leveraging tech and how.
Speaker 3 (28:00):
Yeah, I think people are scared of this.
I think again because it's anew level of automation.
They haven't seen it before.
They used to be the person thatwould spend 30 minutes building
a job description.
Now it takes about 10 secondsand you can imagine every ATS is
building this if it doesn'talready have it.
We already have it.
It didn't take very long for usto develop.
So I think those are thelow-hanging fruit, but there's
(28:21):
quite a bit of fruit everywherein this tree.
It will shake out, certainly,but we've had automation for a
long time.
Email automation is somethingthat we've had automation for a
long time.
Right, like email automation issomething that we've had for
years.
Right, recruiters hit 10different checkboxes and the 10
candidates get the rejectionemail at the same time, or it's
automated completely.
We've had this automation.
I think it's just now at thepoint where we can generate
(28:42):
personalization.
We can generate things at theedge of the node rather than the
sort of like centralizedfunction or centralized matter
that really makes generative AIand agents again, which is like
a next level past that, likesuper unique.
It'll be a very large learningcurve, I think, an adoption
curve.
I think the nature of any sortof transitional change is that
there's going to be, like earlyadopters, people in the middle,
(29:04):
the sort of late.
But this is not one of thosetrends where I would want to be
late, for I would if I was aleader.
I would be evaluating thingslike constantly, like multiple
products a week, because thespace is changing really fast.
It's still going to be acompetitive war for great talent
.
Always, or even in areas whereyou might think it's like a blue
collar role, some of those arethe most competitive, right?
(29:26):
You wouldn't?
You would be surprised, right?
There's only so many peoplethat can do this role and 30
miles outside of Wichita, kansas, that's a really specific role.
So I think we're just going tosee this adoption happen.
But the companies that reallymove fast are going to be,
they're going to get amazingyield very quickly and they'll
realize the value.
(29:46):
And I think the laggards arereally going to struggle, like
really going to struggle, andthat's why it's always great
working with people that can seethe future For sure.
Speaker 1 (29:55):
Look, this has been really, really insightful.
I appreciate you coming on theshow.
Is there anything else that youwant to share with the audience
about your product or anythingelse?
We've got a couple of minuteshere.
Speaker 3 (30:09):
Anything else you want to share.
Yeah, I think the biggest thingto share is just whether you're
evaluating like our technologyor a different vendor.
It doesn't matter the space,maybe it's not even recording,
maybe it's sales or whatever thetypes of tools you're looking
at.
Look at when these companiesstarted always right, I think
you know.
If you're looking at companiesthat started three to five years
ago, right, typically whatyou're going to see is they're
adding things on.
But what's really nice aboutthe newer companies that are
(30:31):
coming out is that they're allAI native companies, right.
They're built from the groundup being AI native and the
experiences will be reallydifferent.
So, even though the end productmight look similar, the
marketing might look similar,the actual experience when you
actually dive into theseproducts are very different, and
the only way to do that iseither getting on a demo call,
right?
(30:51):
So I would say to leaders thatare listening to this podcast be
open-minded, right.
There's a lot of great newsoftware that's coming to the
market.
That's completely changing theentire paradigms, from what
we're seeing and being aware,meeting the founders, ceos of
these companies, I think it's agreat thing because you'll get,
even if you don't decide topurchase, you get a sense of
what is coming and you can beknowledgeable about that and
(31:13):
share that with other people.
And I think, again, going backto what I said earlier, you
don't want to have your headburied in the sand, right?
You've heard that expressionbefore.
This is a biggest revolution, Ithink, in my opinion, since the
birth of the internet and priorto that, probably
semiconductors, which is so youdon't want to miss out on this
trend and you don't want to beunder your head buried in the
sand, even if you're not goingto purchase.
Go out there, be curious, meetpeople.
(31:34):
It will only benefit you andyour organization as you go
through these motions, and it'stime consuming, but it's
important.
Speaker 1 (31:42):
Yeah, I totally agree , Elijah.
Do you have any other finalthoughts?
Speaker 2 (31:50):
No, I think David's right.
I think you need to keep apulse on the market to be a good
TA leader, and I think we'regoing to see that more people
may be in like recruiting ops,like we saw with, like sales ops
and rev ops.
Thinking about the whole techstack, I think the tools those
AI native tools that David'smentioning are going to make the
TA leaders like tech stack alot more yeah, a lot more like
powerful and intelligent andwe'll be able to do things that
(32:14):
we weren't able to do previously.
But we're definitely still alot of catching up to do
compared to the tools that, likesales leaders have at their
disposal to grow revenue, togrow the company and find great
people to join the team.
There's a lot of great toolsout there, but, yeah, probably
not near as many as there are inlike the sales world.
Speaker 3 (32:32):
Yeah, literally sales .
If you think TA is busy in HRtech, sales is just like
incredible, like how many toolscome into market every day.
Speaker 1 (32:41):
I feel like a lot of hopefully not in the songs,
maybe that sounds mean, but alot of them went out of business
.
It was just too many.
Lot of them went out ofbusiness, it was just too many.
There was too many Me Tooproducts in sales.
Like how many differentproducts that do the same thing?
Do you need, good Lord, somecategories?
You'd have like 30, 40 players,like early stage companies,
essentially doing the same thing.
Speaker 2 (33:03):
The pricing's interesting too.
Sorry, david, I was just goingto throw out there.
Sales tools often are a lotmore affordable than some of the
recruiting and HR tech toolsthat I've seen.
Probably a whole differentpodcast episode.
But the pricing differencemaybe that's like you're saying,
james, the number ofcompetitors.
But an email sequencing tool,for example, you could get for,
(33:37):
let's say, $20 to $50 a monthfor a license, some of the
recruiting versions to do emailsequencing or anywhere from what
150 to 300 per license and theyoften don't have the same
feature set either, like theyhave fewer features.
Speaker 3 (33:42):
So, anyways, what were you gonna say, david?
Yeah, I've definitely comeacross orgs where it's
especially smaller orgs maybeyou guys have where, like
hubspot is like their recruitingtool, it's like they've taken a
sales tool or crm and likeHubSpot is like their recruiting
tool, it's like they've taken asales tool or a CRM and like
have tried to like build aroundthat.
It's sort of interesting.
But I think there's just a lotof orgs that, in my mind,
probably don't want to deal withthe complexity of again duct
taping things.
So I think that's probably whyyou still see higher rates and I
(34:05):
don't think that TA is not acompetitive market or recruiting
tech is not a competitivemarket.
It definitely is.
But yeah, I think sales is justit's revenue driving function
and I hate to say that as an HRperson, right, every TA leader
would probably like yell at meright at the top of their lungs
no, we're a revenue generatingfunction and it's true Like HR
is responsible for findingpeople and engaging people and
(34:25):
retaining people that are inrevenue driving positions, and
it is all about the people.
There's no doubt about it.
But I think, if you think aboutfrom a sales perspective, if
you're selling a tool, mostpeople want to sell it to a
revenue generating function.
Right, it's pretty, pretty muchthat's why that is, and so
you're always going to see themost amount of tools that are
(34:46):
going to probably drive revenueor give you insights into
revenue, which is why you see somany of these tools that are to
your point, like RevOps toolsand analytics tools.
It's just like the numbers arejust like staggering in terms of
how many there are out there.
Speaker 1 (34:59):
Yeah, it's just, I don't know how you expect to
scale an incredibly successfulcompany if you had 20
competitors doing the same thing, but I don't know.
Anyways, yeah, the marketcorrected, hopefully is doing
the same thing, but I don't know.
Anyways, yeah, the marketcorrected.
Hopefully it's not as saturatednow.
But anyways, we digress.
We're coming up on time here.
David, thank you so much forjoining us today.
We really appreciate you comingon with us on the show.
Speaker 3 (35:18):
Yeah, it was a lot of fun.
I really appreciate you guyshaving me Great dialogue and I
look forward to stayingconnected, and our company
website is wwwqualai for thosethat are listening.
And yeah, I'd love to just talkto people in the space and see
how we can help you.
Speaker 1 (35:33):
Yeah, that sounds great and we're going to put
everybody tuning in some notesand descriptions so you'll have
the website, david's LinkedInprofile, all that good stuff.
And, by the way, yeah, david,if there's anything you want us
to drop in the show notes, justlet us know.
Speaker 3 (35:46):
For sure Awesome Cool .
Speaker 1 (35:50):
Thank.
Popular Podcasts
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
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.