January 21, 2025 • 45 mins
Anil Dharni, CEO and Founder of Sense, joins host James Mackey to discuss how AI and automation are transforming the landscape of talent acquisition, streamlining recruitment processes while enhancing candidate engagement.
He shares insights about the importance of candidate experience, ethical AI use, and the need for HR leaders to adapt to emerging technological innovations to remain competitive in hiring.
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):
Hello, welcome to the Breakthrough Hiring Show.
I'm your host, james Mackey,really excited about today's
episode.
We have Anil Darni, theco-founder and CEO of Sense, on
the show.
Anil, thanks for joining us.
Speaker 2 (00:11):
Yeah, thank you for having me Really appreciate it.
Speaker 1 (00:13):
Yeah, I'm very excited about this episode.
I'm really excited to learnabout what you're building over
at Sense.
To start us off, would you mindjust giving us a high-level
product overview on what you'rebuilding?
Speaker 2 (00:25):
Yeah, absolutely so.
We've been building Sense forabout eight years and we started
the company.
We are based out of SanFrancisco Bay Area, but we have
a global footprint.
We have teams in Europe, teamsin India, and we are focused on
talent, engagement, engagementand that.
What that means is it can meetthem.
It can mean a mouthful, butit's focused around how are
(00:47):
large companies with complexhiring needs basically
simplifying their hiring needsand figuring out how to use
automation and AI technology tobasically create a great
candidate experience and movecandidates faster through the
funnel?
You know, that's, that's thethe most important thing that is
relevant for talent acquisitionleaders today, especially in
(01:07):
today's market.
Speaker 1 (01:09):
Yeah, and I'm looking at the website right now and it
seems like it's a verycomprehensive platform, so I'm
excited to learn more about that.
But just to take a step back,it would be really cool to learn
a little bit more about you andthe founding story.
How did you come up with theidea to start Sense?
Speaker 2 (01:27):
Yeah, absolutely.
You know I was.
My career started inengineering, so I was building
engineering products back in theboom days of the late 90s.
You obviously probably don'tremember that, but soon.
You know I've been a serialentrepreneur started.
This is my third startup done acouple of companies.
I've done social networking.
I've done mobile gaming, thelast company that that's the one
that we ended up selling.
(01:47):
And after that, you know, Irealized it's the people you
hire that make or break thecompany.
So how much fun would it be tostart a company in the
recruiting space, helping otherorganizations, especially larger
organizations, helping themhire and find the right talent
for their businesses as they'relooking to grow and further
expand.
So that was the genesis.
(02:08):
It's the same core team thatstarted eight years back, the
same folks that were buildingthe mobile gaming company.
I brought them on here and it'sbeen quite the ride.
We have about 1,000 customersand we have basically two
pillars.
I would say, like what reallydifferentiates Sense versus
other sort of HR tech vendors inthe market is we truly have one
segment that's focused on youknow, you have that background
(02:30):
around agencies and staffingcompanies, so we have a large
footprint in that customer base.
And then we have the secondsegment of the market, which is
the direct hire the corporateside of the business right, so
we do both because the solutionis pretty flexible, and direct
hire the corporate side of thebusiness right.
So we do both because thesolution is pretty flexible and
pretty broad, and it's been anexciting ride, yeah that's
(02:51):
awesome.
Speaker 1 (02:51):
Are you still having fun?
Is the big question.
Speaker 2 (02:54):
I think that's why I'm still here and that's why,
like, the founding team isabsolutely intact and we're
having fun.
We are signing up like justamazing logos.
We are impacting a lot of folks.
I think it's a two-sidedproblem, right, as you are very
well familiar the whole thingaround organizations and what
are they trying to achieve anddo.
But it's also about the jobseekers and how can we improve
(03:18):
that and the interaction theyhave, not just with a particular
job they're applying to, butalso with the organization that
they want to work ultimatelywith.
So I think, every day I wake upand there are just so many
challenges.
But of course, with the adventof LMS and the new JNI
technology, it's like reallygiven, I think, a new set of
tailwinds to the market and theopportunity that is in front of
(03:40):
us.
Speaker 1 (03:41):
Yeah, absolutely.
I always like to ask thatquestion whenever somebody
starts a business in recruiting.
It's a really cool space that Ithink now is actually the most
exciting time to be in it.
I mean, at least in my careerand recruiting, which I guess at
this point for my first job inthe industry.
I guess it was about 12, 13years ago, so at least from what
I seen, I got started.
(04:02):
I think my first job instaffing was like 2012.
And this is definitely at leastfrom where I've been working in
the industry definitely the mostexciting time, I think, with
(04:23):
LLMs and some of the innovationthat's coming out of the space,
and how quickly it's happeningis pretty amazing.
That's right.
I couldn't read more.
Well, look, I mean I'm justtrying to figure out where to
start right, Because I'm on yourwebsite and there's a lot.
I mean, your product suite isreally comprehensive and there's
a lot of really cool stuff herethat I'd love to dial in on.
Maybe we could start with whatyou would consider to be your
core product.
Maybe the largest percentage ofyour customer segment is
utilizing the most.
I mean, what is that?
Speaker 2 (04:44):
Yeah, yeah, no, I can jump into that.
I think we can start with thebuyer's perspective, right,
that's what you also alluded toat the start of the conversation
.
But when you think about theselarge organizations and the
director of TA or the VP of TAwho's coming to us and trying to
have a conversation, the needof the hour today definitely
seems to be around recruitmentautomation.
(05:04):
So which is all around.
I need to drive more and moreefficient recruiting process.
I need to make sure that myrecruiters are happy and
productive and they'redelivering on the ROI and the
outcome that we are ultimatelyfocused on.
So that's the mindset throughwhich TA leaders are coming and
the product where that we builda core product is.
(05:27):
Basically, think about it as aworkflow orchestration engine.
So it takes your differentrecruiting processes and it
digitizes them and brings thatonline through basically a
WYSIWYG editor, right?
So you can just drag and dropyour processes, whether those
processes lie in the ADS or theylie on the CRM, and what kind
(05:48):
of through those processes, whatkind of communication and
engagement you want to run onthe candidate side, right?
So think of it as a coreorchestration engine.
So that's the bread and butterof what we sell and some of
these companies will come to usand say, hey, I want to
basically automate from all theway from hello to hire to
onboarding to the first day andonwards, especially in high
(06:10):
volume hiring use cases.
It's like my first 90 day ofemployee engagement is also
pretty important because I havea pretty high attrition rate
there, right?
So all the way, the entiretalent lifecycle automation is
why people come to us and withinthat there's a whole slew of
products that we can go into,you know, all the way from AI
chatbots to voice, ai forscreening sort of use cases,
(06:33):
two-way text messaging, masstext messaging for recruiters so
they get the superpower throughwhich they can do these kind of
mass broadcasts, and it's allhuman operated.
There's no automation in thatpiece.
To you know, moving thecandidate all throughout the
candidate journey and makingsure that.
You know, from a competitiveperspective, this organization
(06:54):
has a massive lead over theircompetitors who are?
not using our platform.
Speaker 1 (06:59):
Absolutely so.
Talk to me.
I'd love to learn more aboutthe workflow and integration
with Epic and tracking systems,because it does seem like you're
able to plug in automationthroughout the entire talent
acquisition funnel, and it seemslike you also mentioned that it
sounds like this could be evengoing into onboarding and post
hiring.
So this is also an HR productfor HR teams, is that right?
Speaker 2 (07:22):
Yeah, so it depends on where the customer wants to
push us.
Yeah, so the answer is yes, butvery much focused more on the
engagement and communicationaspect of it versus, like, the
transactional.
So we are not like we won'thave DocuSign, we won't do like
on document onboarding, right,that's all.
Whatever the system that thecustomer is using will take care
of it.
But the problem is you areaware of, like, whether it's
(07:45):
assessments, whether it'sbackground checks, whether it's
interviews, it's all about thefollow-up.
It's all about the problem,right, it's about nudging the
recruiter, it's about nudgingthe candidate, it's about
nudging the hiring manager,right, and that's where these
processes are broken.
It truly takes a village to geta candidate, to get a lead to
become a candidate, to get alead to become a candidate, to
(08:05):
get from a candidate to actuallymake the hire right.
And that's the gap that weinitially saw and that's why I
go to market initially you know,eight years back, seven years
back was all around staffingindustry.
So we looked at the staffingindustry and we said, okay, high
volume recruiting, but it'srecruiting in professional, in
admin, clerical, in lightindustrial, in blue collar, it's
(08:26):
all over the map, but it's highvolume.
And what's truly missing is anew age recruitment automation
system that can come in it canbe an overlay on top of the
existing applicant trackingsystems, and that's how we work.
And then that brings thefirepower of speed to the hiring
process, of better conversionrates to the hiring process,
(08:46):
right?
The second realization we had isone really needs a multimodal
strategy, right?
So the way you communicate tothe candidate or with the
candidate is not around justemail driven, which is what's
still, you know, even if yousquint your eyes and you look at
a lot of like CRMs today, theyare still email focused.
But in these markets things arechanging so fast, right?
(09:07):
So it's email, it's text-based,it's chatbot interaction, it's
voice interaction, so it reallydoesn't matter, it's through
WhatsApp.
You just choose your mode ofhow you want to engage with the
candidate and we can enable that, unleash that power, right?
So we've thought about it inthose multiple dimensions when
(09:27):
we built the product out.
Speaker 1 (09:28):
Got it and so I totally understand the use case
for staffing and recruiting.
I think that's a really smartplace to start right, because it
is particularly for high volume, right, high volume searches.
They're working with several.
You know a lot of customers,particularly staffing firms,
that are focused on high volumeindustries, right, like
industrial, as you mentioned.
They do have high volume needs.
(09:49):
They have a lot of roles.
They're always hiring.
It's literally their business tohire, right?
So if they're in business,they're hiring their services
companies, meaning they havelower margins.
So they're really looking athow to scale effectively, which
is incredibly hard because a lotof staffing companies can
increase top line growth butthey're making the same amount
or less money.
So they're often early adoptersto automation type of
(10:11):
technology like you're building.
So there's a lot of value there.
But I'm sure, as you've learned, it's like the use case isn't
just relevant to staffing andrecruiting.
It's for in-house corporateteams as well.
Are you noticing, like yourcorporate customers?
Is there kind of like?
You know, some features andproducts are more so leveraged
(10:32):
by corporate teams and then someare more so leveraged by
staffing.
I'd be curious to learn moreabout that.
Speaker 2 (10:39):
Yeah, great question and something that we learn
every day, differences betweenthe two.
But you know, I would say thestaffing companies and the
agencies are pushing theboundaries of what we can do,
and for obvious reasons, exactlywhat you nailed right, which is
this, is your bread and butterright.
You make a placement, you makerevenue, right, yeah, and the
(11:02):
staffing company that makes theplacement first gets that
revenue right.
Otherwise, all that effort thatyou spend trying to get that
candidate submitted and hiredwent to waste because another
staffing company came in andplaced the hire for you right.
So it's highly competitive.
Speed is of the essence.
Quality is still of the essence, because you still need to find
the needle in the haystack asfast as you can.
(11:23):
You alluded to light industrial, but let me tell you nearly 35%
of that business isprofessional.
So it surprises us.
But when you're hiring 1,000,2,000, 3,000 professionals every
single year, that's still highvolume hiring for you, right?
So it's interesting that westill play very well in the
knowledge workspace.
So back to your question aroundcorporate versus staffing.
(11:45):
Corporate is just a laggard.
They are going to adopt exactlythe same things.
Give them a few years, becausethey are going to come to the
same exact conclusion.
So I'll tell you.
A great example is we launchedVoice AI, so it's a voice
product.
You know, there are still somegray lines and boundaries as to
(12:07):
what is compliant, what's notcompliant.
The staffing will push right,whereas a corporation will say,
or direct hire will say, hey,wait a minute, I'm a large
corporation, this could put meat some risk, so I need more
time to evaluate this newtechnology.
It's so brand new that I needsome proof points.
I need my infosec, I need my ITguys, I need my legal, I need
(12:27):
my compliance stuff to come infirst and then give me a green
signal.
So that could take a year, thatcould take a couple of years,
but you know what it's coming.
There's no stopping it.
But I think that's the bigdifference that we see.
Number two I will say is and Ithink you are very familiar with
this is inbound use case versusoutbound.
So in corporate, what hashappened is, I feel a lot of HR
(12:48):
tech vendors has pushed thisnotion of like it's all about
spending as much money as youcan on job boards, because it's
like it's as if they'veconvinced them that you don't
have a database of candidates.
You might be one of the largestautomakers in the world, but
they have convinced successfullyyou don't have.
So all your focus and all yourinvestment in technology needs
(13:11):
to be on inbound Captureeveryone that's coming through
your career side, through jobboards, where you're spending
tens of millions of dollars andyou need to put a lot of
technology on there.
And then the question we come tothem because we have a staffing
background, is like what thehell are you doing with your
outbound right Database ofpotentially millions of
(13:31):
candidates and do you evenunderstand the quality of that
database?
Because let me tell you, whenwe work with staffing companies,
they don't care about theinbound use case.
They are all about why would Igive Indeed LinkedIn a single
dollar when I have 10 millioncandidates already sitting in my
database right?
(13:54):
So I think it's a little bit ofan education, a little bit of a
mountain for us to climb.
But once they realize that allthe techniques of typical
marketing outbound use cases,sales, outbound use cases that
exist that direct hires canstart leveraging on their own
internal databases, I think itunlocks a huge untapped
potential today in the market.
So that's what we are prettyexcited about.
But that's also the differencebetween the two segments.
Speaker 1 (14:15):
Yeah, yeah, that's.
That's really cool.
So what are some of the morerecent features, products that
you've been rolling out tocustomers that there seems to be
the most excitement around?
Speaker 2 (14:27):
Yeah, I mean there's a whole slew around actually the
capabilities of theorchestration engine, right.
So, as you can imagine, like inthe past, the orchestration
engine was built just like theRPAs, or robotic process
automation companies that youall might be familiar with.
It's like you know, if anaction is taken, then do this,
(14:48):
if something happens, then dothis.
So it's all very rules-basedautomation.
So eight years back, when welaunched the product, that's
exactly what we were doing,right?
That's exactly what HubSpotdoes, exactly what we do, that's
what RPAs do.
So we are moving from arules-based world.
And then, you know, then in therecent years, when the LLMs
(15:08):
came out, it's become prettymuch a goal-based world.
So we are moving into the worldof, like the same engine
orchestration engine that washard-coded to some extent and
then could do all thesecommunications and automations,
which is still great and added alot of efficiency.
Now we got to move to a worldwhere it's all goal-based and
agent-ticket nature and that'struly exciting and it's like
(15:31):
it's going to unleash a slew ofbusiness use cases that in the
past were very hard to stitchtogether for companies and
organizations.
And suddenly I think the eyesare opening and the art of
what's possible is coming true,right?
So what does that mean?
That basically means I can tellyou like.
So, for example, if you'rescreening a candidate, right In
(15:53):
the past it would be ask themquestion A, then ask B, then C,
then D, and there's a linearprogression of what you could do
right.
And if in the chat, thecandidate would say, hey, I have
a different question to ask you, the chatbot would just unwrap.
It's like, okay, what justhappened here and I don't
understand it and I'm going tohand you to a human or it's just
(16:16):
a failed chatbot experience.
Right Now, with chatbot or voice, we have this amazing rule base
hey, listen, you can have aconversation if you want,
especially voice if you want,for 30 minutes with this
candidate, but I want you to askthese five knockout questions
one way or the other, becausethat's the goal.
(16:37):
The ultimate goal is, is thiscandidate like an example in
light industry could be?
You know, if they can't standon their feet for more than
eight hours, we shouldn't reallybe wasting our time talking to
that candidate, right?
So now the candidate can spend15 minutes asking different
questions which might not berelated to that one that
screening knockout question butin the end, the AI is smart
(16:58):
enough to bring them back tothat question to ensure that
that knockout question is infact asked right.
So I think the goal-based worldis turning out to be super
exciting for us and there are somany use cases for us to unlock
that right there.
Speaker 1 (17:12):
Yeah, that's a lot of really cool things that the LLM
is able to do and have more ofthose interactive conversations
and loop back, as you mentionedtoo, which is really, really
cool, and I'm just looking atagain, like your website, and
there's just so many differentareas that we can dial into.
I'm curious on just to pivot alittle bit and talk about
(17:38):
sourcing.
I'm wondering if we could dialinto that for a minute.
I mean, I'm talking right nowbecause you're mentioning
accessing candidate pools thatare existent in CR CRMs.
You know there's some companiesdoing some innovative things on
the sourcing side.
Particularly, there's twoepisodes that folks should check
out Steve Vartel with Gem.
They're doing a lot of coolstuff with sourcing right now
Different functionality,different products and features
(18:00):
coming out and also the Workableepisode.
We had the CEO of Workable onthe show, nikos, and he was able
to break down different AIsolutions that they're
incorporating into their product.
Speaker 2 (18:21):
But yeah, I'm curious about, from a sourcing motion,
how you're thinking aboutleveraging LLMs to provide
better sourcing experiences andautomation and speed and quality
and that type of thing.
Yeah, yeah, I mean, that's somuch to unpack over there, but I
can start with a simple thing.
But first, before we getstarted on that, one caveat is
we do not bring a database toour customers, right?
So we assume that they havecertain sources of candidates
where they can access them.
The whole database world isriddled with issues and, I think
(18:42):
, ethical challenges, franklyspeaking, from where people have
sourced those databases.
So we try and stay out of that.
So what we see is one is openingup the aperture for the
organizations in terms of, youknow, giving candidates the
opportunity to apply in multipleways, right, so it could be QR
(19:03):
codes, text to apply career site.
They could be sitting to yourpoint, they could be sitting in
the CRM, they could be sittingin other databases.
There's also access to jobboard databases and things like
that, right?
So I think one layer one has tobuild as an HR tech vendor is
like how do you bring all thosedatabases together in a unified
search experience so when a jobdescription comes in, you
(19:25):
understand what it's looking for, then you can go and run a
really highly qualified searchon those multiple sources of
data.
So you have your passivecandidates that might be in any
of these data sets, you haveyour active candidates that are
applying inbound, and then youhave to mesh together and search
in that context and ifsearching is just not enough,
(19:49):
once you search, then you needto quickly reach out to them as
fast as humanly possible.
And that's where the automationlayer comes in right.
So all the way, extracting thejob description from the job
description, extracting likewhat is the semantic search that
needs to be run on thesedatabases, is the semantic
search that needs to be run onthese databases, and then from
(20:09):
there coming up with a list andhopefully that list at least is
ranked in some way or the otherright.
So you might have 100, 200candidates that match the
profile but you're looking fornow.
Once you have that list, thenyou need to reach out to them
within seconds.
So that's where the whole flowruns.
So you use an app to create ajob description.
From the job description youextract the search.
(20:30):
From the search you search onall these databases.
That whole flow is just anagentic flow, right.
And then you start thecommunication flow.
So now you start reaching outto those candidates.
Maybe you need to reach out to100.
So you start going one by oneand you start having either chat
conversations or voiceconversations.
Now what we've understood fromchat is chat is great and it
(20:53):
serves this purpose.
Not everyone wants to be on thephone and have a conversation
in public or even at home, butthe chat responses are pretty
short in nature, so the amountof information you can actually
get out of the candidate islimited.
So we've had a very successfulchatbot.
There are other chatbotcompanies.
(21:14):
It's great, but text is limitedand voice definitely leads to a
richer conversation and you canbuild a much richer profile
about the candidate through avoice conversation.
So anyway, but going back tosort of these knockout screening
questions, then those happenand then you have kind of create
a shortlist that yourrecruiters can work off of.
(21:35):
So hopefully, you know I'veexplained at least some part of
the sourcing side that is, yeah,that can basically become super
fast and agentic in nature.
Speaker 1 (21:45):
But yeah, that was what I'm curious about, because
that's I'm seeing a lot of usecases around sourcing as well,
and then going into thispre-screening use case, which I
think is really interesting,right, and so I'm curious,
though, from when you're gettinginto more so the pre-screening
where the AI agent is reachingout to folks.
What about from a consentperspective?
Is there anything that folkshave to consent in?
(22:05):
And then, like, how, how do youensure that you're able to do
that for your customers so thatyour agent can actually do the
outreach?
Speaker 2 (22:17):
Yeah, so, so, absolutely so.
These databases, hopefully.
Yeah, the only way we work withthem is you know if the client
has gotten the permission toreach out to that.
So, and we follow GDPR and allthose compliance and depending
even on sometimes which stateyou're dealing with, which
country you're dealing with, soit's all localized and we comply
with that.
And voice is also a newterritory.
Like just because you gotconsent for text doesn't really
(22:39):
mean that you got consent forvoice.
So there are nuances there.
But I think even there, likethe way it's pretty nice, like
the orchestration layer allowsyou to text somebody and say,
hey, I'm going to call you in afew minutes, would that be okay
to have a conversation, and thenthe person gives you consent
and then you have a conversation, right?
So at all points we areobviously following the opt-in
(23:01):
and opt-out criteria, which iseither maintained by the ATS or
maintained by us.
And it's bidirectional sync atany given time.
Of course we don't go blind, wedon't go and get into a
database where they have noconsent and we start spamming.
That's the worst experiencepossible.
Yeah, yeah, yeah.
Speaker 1 (23:20):
But you also mentioned for outbound.
So I guess for is there a usecase for when folks are like
recruiters are doing outreach tocandidates, like, let's say,
doing like human touch, likereach, linkedin and mails, to
then run into prescript Like howdoes how does the workflow work
if that's like it's newoutbound source candidates?
Speaker 2 (23:41):
Yeah, so you know.
So there is the.
As you know, all the CRMs andADSs have come.
You know, those are 20 year oldtechnologies, 25 year old
technologies.
They've come into the worldwith a view of, like it's a
recruiter or a sourcer use case,right, that's the insertion
point.
I want recruiters to come inand do a take certain action,
sources to come in and takecertain action.
We came in from an automationfirst world so we believe in the
(24:08):
future.
Like recruiters don't need tohang out for hours in the CRM or
the ATS.
Either way, they don't likethose, right?
So everything that I can dowhether I'm negotiating, whether
I am selling them the positionof what the value proposition of
this job is, and the company itcan all be done offline without
the ATS or the CRM involved.
Right, and now the AI is justhelping record all those calls
(24:29):
and put them out right.
So, yeah, so our approach isdifferent, you know.
So we do allow for if therecruiter or the sourcer comes
in and has to build a talentpool first manually, they can do
it with the platform.
Ideally, we just like to buildeverything that is just fully
(24:49):
automated, right?
The future is pretty simple.
You have your annual operatingplan and your annual hiring plan
uploaded in the system.
Adss have that capability.
From that you extract all theroles people are hiring for.
You know exactly the roles thatare coming up.
You potentially know the rolesin which month they need to hire
.
You have the data to tell youlike how, what kind of lead
times you need.
You automatically create thosetalent pools.
A recruiter doesn't need to doanything and you start building
(25:17):
those talent pools and startmessaging and asking those
people for having conversations.
You're having voiceconversations, chatbot
conversations.
Hey, I know you were a silvermedalist six months back.
Where are you today?
What's your job title?
Would you be looking for anopportunity in the next six
months?
And you know which candidatesare warm, which candidates are
hot, which candidates areactually cold.
They don't want you to talk tothem again, you know.
So that's where all of this isheaded, and I think there's a
(25:41):
mix of AI and human power willalways remain, but I think the
human, the amount that will behuman powered, will keep
diminishing over time.
At least, that's ourperspective.
Speaker 1 (25:54):
Okay, yeah, for sure.
What about?
So can we dial in on AIevaluation and how you're
thinking about that?
So you have pre-screening rightFor just like the top knockout
questions, right, which is afairly straightforward use case,
and then you start to get alittle bit down funnel and I'm
wondering if that's a use caseyou're thinking about.
(26:17):
I think we're seeing probablymore demand top of funnel or you
probably have more datasurrounding this than I do, but
my assumption is top of funnelbecause it's higher volume.
It's just a lot more work,right.
And then when you get likeparticularly for professional
service knowledge work, by thetime you get to like a second or
a third round, potentially peropening you have fewer
(26:37):
candidates.
So maybe there's potentially,arguably maybe, a little less
demand in the market currentlyfor down funnel evaluations, and
maybe there's also a trustaspect and maybe there's also
concerns around potentialupcoming regulation or whatever
else.
I don't know.
I don't know.
I mean, I guess I'm justthinking like how are you
thinking about like deeperevaluation use cases and is that
(27:01):
something, an area you're goingto play?
Speaker 2 (27:02):
yeah, no, I think I think you kind of nailed it like
.
The first part is top of thefunnel, is a very much an
efficiency use case, right, and?
And also a use case where theTA leaders are realizing my
recruiters are getting boggeddown by the sheer scale of
candidates that are coming in.
There's just no way, and Ithink if you're a TA leader that
has not made an investment oris not thinking of making an
(27:23):
investment at the top of thefunnel to do something about it,
I think you're failing yourorganization, right.
So you need to like, reallybring some of these technologies
, because they give youefficiency, they give you scale,
they reduce the noise, so, butthere's no value, not
necessarily like code evaluationhappening right.
And, by the way, this is validnot just for professional roles,
even for high volume roles.
(27:43):
In high volume roles, let metell you, like, if you're a
large BPO center, you arespending tens of millions of
dollars on assessment toolscenter.
You are spending tens ofmillions of dollars on
assessment tools.
They don't want the noise tohit the assessment tool, they
don't want fraudulent applicantsto hit the assessment tool.
They want to keep screening,keep knocking out people as much
(28:04):
as they can to only have aselect list of candidates.
At the end of the day, inprofessional it might be I just
need 10, like, give me reallywell screened folks that could
match this role and then I willgo super deep with them.
That's why I also want myhiring managers to spend a lot
more time right.
So, yeah, so to your point.
(28:24):
There is this screeninghappening, there is the noise
reduction happening at the top,and companies are investing
massively.
But they're also trying toavoid costs that come later,
either cost to the hiringmanager, cost for assessments,
cost for background checks, andif they are low margin
businesses that need that rapidhiring, they really want to
(28:46):
reduce any of this OPEX thatthey can.
So it's a real ROI pitch forthem.
On the evaluation scale at thebottom, towards sort of the as
you get closer to the hiring,yes, I think interview tools
that are at least able totranscribe what happened in the
interview and very quickly sortof compare those to the
scorecards that people mighthave to use to evaluate.
(29:09):
You know the final.
Maybe it's the silver medalist,the final person they help
picking, but at the end of theday it's still core human in
nature, according to us, and itwill probably remain there, at
least in knowledge, work andprofessional, for a while,
that's not going away.
Yeah, I mean, that's a goodthing, and I think that's a good
thing.
(29:29):
You want it that way.
Speaker 1 (29:31):
Yeah, yeah, I think so.
I mean, I think that's theconclusion that I've come to as
well just from hosting thisseries, talking with a lot of
brilliant CEOs such as yourself,right, learning from folks who
have a lot of access to data andare constantly having these
very high level strategicconversations with customers,
and so I think that that makes alot of sense for me.
(29:54):
Another question that I have isyou know, you have companies
like BrightHire and Pillar,companies that are doing like I
suppose at this point they'recalling it interview
intelligence, essentiallyjoining, co-piloting, zoom
interviews, which I'm sure, iftheir CEOs heard me explain it
that way, maybe I should let mejust, it's a lot more than that,
right, but basically they'rejoining recruiters on the
(30:16):
interviews.
They have all the rolerequirements on the back end and
they're essentially they haveall the custom questions and
packaging the data.
To summarize, you know how well, how many of the questions that
folks answer and hey, did youforget to ask some?
So they push that along to thenext interview.
They're, of course, generatingjob descriptions and doing some
of those generation tasks toofor customer questions, but
(30:38):
they're really ensuring thatcandidates are going through a
consistent, thorough,comprehensive interview process.
Is that an area that your teamis building in that direction,
or are you staying away fromthat?
Speaker 2 (30:52):
Yeah, we are staying away from that, but we like
aspects of that.
We help enable that in the way.
We have a large number ofcustomers that have bought our
interview automation schedulingproduct, which is a very big
need of the hour, right.
So you're talking about theactual interview process and the
interview itself.
I'm talking about the processaround the interview, which is
(31:15):
key because hiring managers arefrustrated, recruiters are
frustrated, candidates arefrustrated Everybody's
scheduling, rescheduling hiringmanager are not prepared and
they're coming into theseinterviews right.
So that's a big problem, hugepain point.
We have some of the largesthealthcare hospital systems that
we sell that product to.
We have, I think, the fourthlargest, the fifth largest
automaker that is using us.
(31:35):
So tons of examples.
The example is as simple asthat.
I have 20,000 hiring managers.
Please fix interview automationscheduling for them right.
So it's such a natural extensionof an automation platform.
Right Now.
Your point where you're comingto is we've seen this work like
(31:56):
consistency is required Beyondconsistency.
That's why chatbots and voiceAI is great.
It's asking the exact samequestion over and over again to
tens of thousands of candidatesfor that particular role.
That's great.
It's removing actually biasfrom that screening right that
used to exist.
So same thing is absolutelytrue, probably for the interview
(32:16):
platforms, like at leastthey're helping the hiring
managers ask the same questionsand be consistent, I think.
Number two instead of talkingabout the candidates, it's
important to talk about thehiring managers.
We know so many hiring managerseven probably me we have gaps
when we interview candidates.
We are probably not asking thequestions the right way.
Speaker 1 (32:34):
There's a lot of training that we can provide to
these hiring managers,especially the first-time hiring
managers, and if there's an AItool, ai co-pilot that's helping
me train my own internalmanager teams to ask better
questions, to put the best footforward.
That's a great win, I think.
Yeah, I mean, I think it's likealmost to some extent there,
and that's a whole other side ofthe product too that I'm glad
(32:56):
you're bringing up, becauseBrightHire and Pillar are also
focused on essentially likeevaluating what the hiring
managers are doing.
I think even BrightHire has afair amount of this
functionality where they'reactually, I think, data like,
even even like stuff like didyou start the interview on time?
Did you ask questions the rightway?
Did you spend five minutestalking about the football game
you watched on Sunday and notreally dialing into the
(33:16):
screening questions?
Right, just helping them stayorganized and producing a score?
Almost like, maybe like a gongright On the sales side they're
doing some of that kind ofanalytics functionality on the
call itself, which I think ispretty cool too.
But yeah, it seems like they'vegone like the interview
intelligence, like co-pilot,packaging data, showing the
candidates answers, responses,how relevant it is to the role,
(33:38):
requirements, and then alsoproviding analytics and data on
how the hiring team is doinginterviewing.
You know cause I?
You know it is havingconsistent interviews.
Having hiring managers performat a high level when it comes to
hiring and interviewing ischallenging, we interviewers.
Having hiring managers performat a high level when it comes to
hiring and interviewing ischallenging.
We had Daniel Chait on the showa lot, the co-founder and CEO
of Greenhouse.
There's something funny healways likes to say is hiring
managers struggle with twothings hiring and managing Right
(34:01):
.
Speaker 2 (34:01):
So you know, yeah, no , no, absolutely.
Speaker 1 (34:04):
And the same thing.
Speaker 2 (34:05):
Like you know, when we have the text message, we
have a whole text messagingsuite.
Right, as I told you, therecruiters have.
We try and measure the samething.
We try and measure how manymessages they send, what kind of
speed do they have?
How fast do they respond tocandidates?
Which recruiters have betterresponse rates?
And we are also trying to justgive these sort of signals back
(34:26):
to the TA teams.
And now the teams are like okay,it's not the recruiter that is
pointing the finger.
Now it's like I actually have atranscribed chat of yours in an
interview of Mr Hiring Managerand here are the things that
went wrong.
Right, so I would like you toimprove, because that's leading
to a poor candidate experiencefor us.
Right, so that loop, like ourautomation, can measure the
(34:49):
candidate experience at anypoint point before the interview
, before the screen, after thescreen, after the interview,
after the disposition, the finaldecision was made and we can
give that intelligence back.
Right, so it's super importantand I think LLMs and AI is
really going to help change thegame, as long as they're not
making the final decision, weare all good, but yeah, you know
(35:11):
, I think it's a truly excitingtime.
Speaker 1 (35:15):
Yeah, it's a lot of really cool, really cool stuff
happening for sure.
What about, from a regulationperspective, for companies that
are a little bit concerned aboutincorporating AI into their
workflow?
Right, I mean, like right now,for instance, like one thing
I've brought up several times onthe show is, like workday right
now there's a class actionlawsuit for their AI system,
(35:37):
potentially like discriminatingfolks over several different
reasons, right, but you know,people are, I think some folks
might be a little bit concernedabout applying AI technology and
I'm wondering, like you know, Ithink evaluation is a key
sticking point on AI making afinal decision, or, and I'm
curious to get your thoughts onthat I mean, I think, honestly,
maybe there's nuance to thedefinition of what we consider
(35:59):
evaluating Right and, and youknow, is it considered?
For instance, like, is itconsidered?
Maybe it's it's more soevaluating if, if ai is
generating role requirements,generating a jd, and then
basically deciding what the rolerequirements are and then
evaluating against those, maybe,maybe that's what evaluating is
(36:22):
.
Maybe, if a hiring team putstogether the role requirements
and then essentially, the is itand the AI is evaluating or is
interviewing for that, Is thatreally evaluating or is that
just matching Right, like, andso I'm just wondering, like, how
do you think like the nuancesis like where?
Where do you think regulationis going to kind of fall in this
in the evaluation AIdecision-making piece?
Speaker 2 (36:43):
Well, yeah, yeah, I'm no EUOC expert or regulation
expert, so I won't, I won't makeany statements around that, but
yeah, I mean, listen, it'sfairly complicated.
It does to your point.
It does start from the JD.
Everything falls flat from theJD, the noise starts from the JD
(37:04):
.
You just copy something fromthe job description, from Google
or some past job description,and then you just run with it.
Right, that's how the screeningquestions are generated.
That's how the screeningquestions get on a chatbot,
that's how they get on a voiceAI.
You know, if you made themistake in the job description,
everything flows from there.
Right, that is the freakingreality.
Right, and nobody wants toaccept it, right, but that is
(37:24):
the reality.
So you know, the kind of thingsthat we are experimenting is
like imagine a hiring managerable to talk to a voice agent
and tell them exactly whothey're looking for, the core
skills I'm looking for.
Hey, it is still a forkliftoperator, or it is still a
cashier at a franchise.
But like, this is what makesthis particular role special, or
(37:45):
this is what makes thislocation special, you know, and
let me tell you more about it.
And can the voice agent capturethat and really humanize the
job description where, yes,there are certain knockout
things like you have to be here,but here's where I can give you
more context around that role,even though you know I'm Bath
and Body Works and you'reLululemon and you're Nike store,
(38:05):
but it's still a store job.
But here's what makes us trulydifferentiated, right?
So we are going to try and pushthe boundaries on that front,
but I did want to make the pointlike, job descriptions are the
worst and they start the wholeproblem chain, right, but there
is bias, potential bias.
So what are we measuring itagainst?
I think the regulations need tounderstand we are measuring
(38:26):
against humans, and I think it'stoo much bent on.
Hey, ai is biased, but humansare not.
The reality is, when we takethe human data and we feed it
into AI, ai is horribly biased.
That's when we realize, oh shit, we shouldn't be talking about
which school they went to, whatwas their name, what was their
(38:47):
gender, what was their location.
We have to keep all the dataout before we train the models.
Right and well, that data lies.
When, as soon as that resumehits a human, there's no way the
human's not looking at it,right?
So I think one is trulyfiguring out what has to be
screened out of these resumes.
(39:08):
The good news is for a lot ofjobs and I would say I don't
know whether you know the statslike literally 65% of the
world's workforce is not onLinkedIn.
How many of those peopleactually have a resume?
So what we are matching towardsand what we are training our
models towards is what I wouldcall progressive matching, and
(39:33):
what that means is there mightnot be a resume.
There's a good chance thatthere won't be a resume, right?
So I want to understand who youare.
I want to have a conversationbased on the skills that the job
is saying it absolutely musthave, but then I'm trying to
understand your preferences.
You know, can you work overtime?
Can you do the shift or not?
Can you work on the weekends?
Do you even have a transport toget to work?
(39:55):
I'm trying to understand who isAnil as a person, and even if
you tell me I've done customersupport work before, okay, but
were you in a situation wherethere was a blow up?
Can you talk to me more aboutthat?
So, trying to drill a littlebit deeper than just what a
resume might have, and even ifyou have a resume, a resume is
just another data point to use,but you've got to use the job
(40:19):
seekers' preferences, what theyexpect.
You've got to take into accountwhat the manager is saying and
then try and intersect it.
So the future, we believe, istry and remove the biases from
the training models as much asyou can.
Get yourself audited all thetime in real time.
Hopefully Make the AI moreexplainable.
If the AI matches a certainperson to a role, can it
(40:44):
actually explain why it did that?
And then, if you make itcompletely blind, will it match?
Will it match the same personagain?
You know that's where it getsvery tricky.
So I think explainability is ahuge sort of actually a killer
feature if you can get it goingin your product, and then you
need to make sure that thatexplanation satisfies the
auditors at some point when youaudit it right or can pass a
(41:08):
bias audit right.
So that's those are all theinvestments that at Sense we've
made to make sure that we aredoing it the right way and we
are following at least the localregulations.
I will say one last thing whichpeople miss.
We are so focused constantly oncandidates, inbound candidates
coming in, screening them.
It's not just screening anddispositioning them and saying
(41:31):
you're rejected, it's actuallyscreening them.
You know, it's not justscreening and dispositioning
them and saying you're rejected,it's actually screening them,
saying that you're not a goodfit.
But here are three other jobswhere you are a better fit for
and let me help you apply tothose jobs right, especially in
the high volume scenario,because that's truly possible.
But even if you're a technologycompany, maybe you apply for an
AI job and you're not a greatfit for the AI job.
(41:52):
You don't have that level ofexpertise that was required, but
maybe I can deflect you to aregular software engineer job
and get you hired there.
So the whole point of the systemthat we are trying to build is
and actually our data is proving70% of the candidates that
apply to a job are not a goodfit for that job.
(42:14):
They just think they are, theybelieve, they want to believe
that they are.
It's a aspirational apply, it'snot a true apply.
But if we can deflect them andput them in roles and jobs and
show them the jobs that theymight be a better fit at, we can
actually improve the conversionrates, we can improve the fill
rates and we can make both sideshappy.
(42:35):
Right, we can explain to themwhy.
Why are we taking the actionthat we are taking?
So that's our vision and that'swhy we think the candidate
experience also needs to change.
The good news is a lot of TAleaders understand that and they
know that that's what'sactually happening in the market
.
Speaker 1 (42:51):
Lot of TA leaders understand that and they know
that that's what's actuallyhappening in the market.
Yeah, wow, this has been anincredibly insightful
conversation.
I've certainly learned a wholelot.
I know our audience is going toreally enjoy this and I would
actually one day like tocontinue the conversation if you
want to come back on the showman.
You just shared so much greatinsight with us.
This has really been fantastic.
(43:11):
I think it's really incrediblewhat you're building over.
At Sense, it seems like areally impressive platform, a
lot of different functionalityto help out through the entire
workflow.
So I'm really excited tocontinue to follow what your
team is building and it soundsit just sounds all like really
impressive and incrediblyhelpful to your customers.
So, anyways, thank you forcoming on and sharing a little
bit about what you're doing.
Speaker 2 (43:32):
Likewise Thanks, james, and let's stay connected.
Speaker 1 (43:35):
Yeah, absolutely, and for everybody else tuning in.
Thank you so much for joiningus and we'll talk to you soon.
Take care.
Hello, welcome to the Breakthrough Hiring Show.
I'm your host, james Mackey,really excited about today's
episode.
We have Anil Darni, theco-founder and CEO of Sense, on
the show.
Anil, thanks for joining us.
Speaker 2 (00:11):
Yeah, thank you for having me Really appreciate it.
Speaker 1 (00:13):
Yeah, I'm very excited about this episode.
I'm really excited to learnabout what you're building over
at Sense.
To start us off, would you mindjust giving us a high-level
product overview on what you'rebuilding?
Speaker 2 (00:25):
Yeah, absolutely so.
We've been building Sense forabout eight years and we started
the company.
We are based out of SanFrancisco Bay Area, but we have
a global footprint.
We have teams in Europe, teamsin India, and we are focused on
talent, engagement, engagementand that.
What that means is it can meetthem.
It can mean a mouthful, butit's focused around how are
(00:47):
large companies with complexhiring needs basically
simplifying their hiring needsand figuring out how to use
automation and AI technology tobasically create a great
candidate experience and movecandidates faster through the
funnel?
You know, that's, that's thethe most important thing that is
relevant for talent acquisitionleaders today, especially in
(01:07):
today's market.
Speaker 1 (01:09):
Yeah, and I'm looking at the website right now and it
seems like it's a verycomprehensive platform, so I'm
excited to learn more about that.
But just to take a step back,it would be really cool to learn
a little bit more about you andthe founding story.
How did you come up with theidea to start Sense?
Speaker 2 (01:27):
Yeah, absolutely.
You know I was.
My career started inengineering, so I was building
engineering products back in theboom days of the late 90s.
You obviously probably don'tremember that, but soon.
You know I've been a serialentrepreneur started.
This is my third startup done acouple of companies.
I've done social networking.
I've done mobile gaming, thelast company that that's the one
that we ended up selling.
(01:47):
And after that, you know, Irealized it's the people you
hire that make or break thecompany.
So how much fun would it be tostart a company in the
recruiting space, helping otherorganizations, especially larger
organizations, helping themhire and find the right talent
for their businesses as they'relooking to grow and further
expand.
So that was the genesis.
(02:08):
It's the same core team thatstarted eight years back, the
same folks that were buildingthe mobile gaming company.
I brought them on here and it'sbeen quite the ride.
We have about 1,000 customersand we have basically two
pillars.
I would say, like what reallydifferentiates Sense versus
other sort of HR tech vendors inthe market is we truly have one
segment that's focused on youknow, you have that background
(02:30):
around agencies and staffingcompanies, so we have a large
footprint in that customer base.
And then we have the secondsegment of the market, which is
the direct hire the corporateside of the business right, so
we do both because the solutionis pretty flexible, and direct
hire the corporate side of thebusiness right.
So we do both because thesolution is pretty flexible and
pretty broad, and it's been anexciting ride, yeah that's
(02:51):
awesome.
Speaker 1 (02:51):
Are you still having fun?
Is the big question.
Speaker 2 (02:54):
I think that's why I'm still here and that's why,
like, the founding team isabsolutely intact and we're
having fun.
We are signing up like justamazing logos.
We are impacting a lot of folks.
I think it's a two-sidedproblem, right, as you are very
well familiar the whole thingaround organizations and what
are they trying to achieve anddo.
But it's also about the jobseekers and how can we improve
(03:18):
that and the interaction theyhave, not just with a particular
job they're applying to, butalso with the organization that
they want to work ultimatelywith.
So I think, every day I wake upand there are just so many
challenges.
But of course, with the adventof LMS and the new JNI
technology, it's like reallygiven, I think, a new set of
tailwinds to the market and theopportunity that is in front of
(03:40):
us.
Speaker 1 (03:41):
Yeah, absolutely.
I always like to ask thatquestion whenever somebody
starts a business in recruiting.
It's a really cool space that Ithink now is actually the most
exciting time to be in it.
I mean, at least in my careerand recruiting, which I guess at
this point for my first job inthe industry.
I guess it was about 12, 13years ago, so at least from what
I seen, I got started.
(04:02):
I think my first job instaffing was like 2012.
And this is definitely at leastfrom where I've been working in
the industry definitely the mostexciting time, I think, with
(04:23):
LLMs and some of the innovationthat's coming out of the space,
and how quickly it's happeningis pretty amazing.
That's right.
I couldn't read more.
Well, look, I mean I'm justtrying to figure out where to
start right, Because I'm on yourwebsite and there's a lot.
I mean, your product suite isreally comprehensive and there's
a lot of really cool stuff herethat I'd love to dial in on.
Maybe we could start with whatyou would consider to be your
core product.
Maybe the largest percentage ofyour customer segment is
utilizing the most.
I mean, what is that?
Speaker 2 (04:44):
Yeah, yeah, no, I can jump into that.
I think we can start with thebuyer's perspective, right,
that's what you also alluded toat the start of the conversation
.
But when you think about theselarge organizations and the
director of TA or the VP of TAwho's coming to us and trying to
have a conversation, the needof the hour today definitely
seems to be around recruitmentautomation.
(05:04):
So which is all around.
I need to drive more and moreefficient recruiting process.
I need to make sure that myrecruiters are happy and
productive and they'redelivering on the ROI and the
outcome that we are ultimatelyfocused on.
So that's the mindset throughwhich TA leaders are coming and
the product where that we builda core product is.
(05:27):
Basically, think about it as aworkflow orchestration engine.
So it takes your differentrecruiting processes and it
digitizes them and brings thatonline through basically a
WYSIWYG editor, right?
So you can just drag and dropyour processes, whether those
processes lie in the ADS or theylie on the CRM, and what kind
(05:48):
of through those processes, whatkind of communication and
engagement you want to run onthe candidate side, right?
So think of it as a coreorchestration engine.
So that's the bread and butterof what we sell and some of
these companies will come to usand say, hey, I want to
basically automate from all theway from hello to hire to
onboarding to the first day andonwards, especially in high
(06:10):
volume hiring use cases.
It's like my first 90 day ofemployee engagement is also
pretty important because I havea pretty high attrition rate
there, right?
So all the way, the entiretalent lifecycle automation is
why people come to us and withinthat there's a whole slew of
products that we can go into,you know, all the way from AI
chatbots to voice, ai forscreening sort of use cases,
(06:33):
two-way text messaging, masstext messaging for recruiters so
they get the superpower throughwhich they can do these kind of
mass broadcasts, and it's allhuman operated.
There's no automation in thatpiece.
To you know, moving thecandidate all throughout the
candidate journey and makingsure that.
You know, from a competitiveperspective, this organization
(06:54):
has a massive lead over theircompetitors who are?
not using our platform.
Speaker 1 (06:59):
Absolutely so.
Talk to me.
I'd love to learn more aboutthe workflow and integration
with Epic and tracking systems,because it does seem like you're
able to plug in automationthroughout the entire talent
acquisition funnel, and it seemslike you also mentioned that it
sounds like this could be evengoing into onboarding and post
hiring.
So this is also an HR productfor HR teams, is that right?
Speaker 2 (07:22):
Yeah, so it depends on where the customer wants to
push us.
Yeah, so the answer is yes, butvery much focused more on the
engagement and communicationaspect of it versus, like, the
transactional.
So we are not like we won'thave DocuSign, we won't do like
on document onboarding, right,that's all.
Whatever the system that thecustomer is using will take care
of it.
But the problem is you areaware of, like, whether it's
(07:45):
assessments, whether it'sbackground checks, whether it's
interviews, it's all about thefollow-up.
It's all about the problem,right, it's about nudging the
recruiter, it's about nudgingthe candidate, it's about
nudging the hiring manager,right, and that's where these
processes are broken.
It truly takes a village to geta candidate, to get a lead to
become a candidate, to get alead to become a candidate, to
(08:05):
get from a candidate to actuallymake the hire right.
And that's the gap that weinitially saw and that's why I
go to market initially you know,eight years back, seven years
back was all around staffingindustry.
So we looked at the staffingindustry and we said, okay, high
volume recruiting, but it'srecruiting in professional, in
admin, clerical, in lightindustrial, in blue collar, it's
(08:26):
all over the map, but it's highvolume.
And what's truly missing is anew age recruitment automation
system that can come in it canbe an overlay on top of the
existing applicant trackingsystems, and that's how we work.
And then that brings thefirepower of speed to the hiring
process, of better conversionrates to the hiring process,
(08:46):
right?
The second realization we had isone really needs a multimodal
strategy, right?
So the way you communicate tothe candidate or with the
candidate is not around justemail driven, which is what's
still, you know, even if yousquint your eyes and you look at
a lot of like CRMs today, theyare still email focused.
But in these markets things arechanging so fast, right?
(09:07):
So it's email, it's text-based,it's chatbot interaction, it's
voice interaction, so it reallydoesn't matter, it's through
WhatsApp.
You just choose your mode ofhow you want to engage with the
candidate and we can enable that, unleash that power, right?
So we've thought about it inthose multiple dimensions when
(09:27):
we built the product out.
Speaker 1 (09:28):
Got it and so I totally understand the use case
for staffing and recruiting.
I think that's a really smartplace to start right, because it
is particularly for high volume, right, high volume searches.
They're working with several.
You know a lot of customers,particularly staffing firms,
that are focused on high volumeindustries, right, like
industrial, as you mentioned.
They do have high volume needs.
(09:49):
They have a lot of roles.
They're always hiring.
It's literally their business tohire, right?
So if they're in business,they're hiring their services
companies, meaning they havelower margins.
So they're really looking athow to scale effectively, which
is incredibly hard because a lotof staffing companies can
increase top line growth butthey're making the same amount
or less money.
So they're often early adoptersto automation type of
(10:11):
technology like you're building.
So there's a lot of value there.
But I'm sure, as you've learned, it's like the use case isn't
just relevant to staffing andrecruiting.
It's for in-house corporateteams as well.
Are you noticing, like yourcorporate customers?
Is there kind of like?
You know, some features andproducts are more so leveraged
(10:32):
by corporate teams and then someare more so leveraged by
staffing.
I'd be curious to learn moreabout that.
Speaker 2 (10:39):
Yeah, great question and something that we learn
every day, differences betweenthe two.
But you know, I would say thestaffing companies and the
agencies are pushing theboundaries of what we can do,
and for obvious reasons, exactlywhat you nailed right, which is
this, is your bread and butterright.
You make a placement, you makerevenue, right, yeah, and the
(11:02):
staffing company that makes theplacement first gets that
revenue right.
Otherwise, all that effort thatyou spend trying to get that
candidate submitted and hiredwent to waste because another
staffing company came in andplaced the hire for you right.
So it's highly competitive.
Speed is of the essence.
Quality is still of the essence, because you still need to find
the needle in the haystack asfast as you can.
(11:23):
You alluded to light industrial, but let me tell you nearly 35%
of that business isprofessional.
So it surprises us.
But when you're hiring 1,000,2,000, 3,000 professionals every
single year, that's still highvolume hiring for you, right?
So it's interesting that westill play very well in the
knowledge workspace.
So back to your question aroundcorporate versus staffing.
(11:45):
Corporate is just a laggard.
They are going to adopt exactlythe same things.
Give them a few years, becausethey are going to come to the
same exact conclusion.
So I'll tell you.
A great example is we launchedVoice AI, so it's a voice
product.
You know, there are still somegray lines and boundaries as to
(12:07):
what is compliant, what's notcompliant.
The staffing will push right,whereas a corporation will say,
or direct hire will say, hey,wait a minute, I'm a large
corporation, this could put meat some risk, so I need more
time to evaluate this newtechnology.
It's so brand new that I needsome proof points.
I need my infosec, I need my ITguys, I need my legal, I need
(12:27):
my compliance stuff to come infirst and then give me a green
signal.
So that could take a year, thatcould take a couple of years,
but you know what it's coming.
There's no stopping it.
But I think that's the bigdifference that we see.
Number two I will say is and Ithink you are very familiar with
this is inbound use case versusoutbound.
So in corporate, what hashappened is, I feel a lot of HR
(12:48):
tech vendors has pushed thisnotion of like it's all about
spending as much money as youcan on job boards, because it's
like it's as if they'veconvinced them that you don't
have a database of candidates.
You might be one of the largestautomakers in the world, but
they have convinced successfullyyou don't have.
So all your focus and all yourinvestment in technology needs
(13:11):
to be on inbound Captureeveryone that's coming through
your career side, through jobboards, where you're spending
tens of millions of dollars andyou need to put a lot of
technology on there.
And then the question we come tothem because we have a staffing
background, is like what thehell are you doing with your
outbound right Database ofpotentially millions of
(13:31):
candidates and do you evenunderstand the quality of that
database?
Because let me tell you, whenwe work with staffing companies,
they don't care about theinbound use case.
They are all about why would Igive Indeed LinkedIn a single
dollar when I have 10 millioncandidates already sitting in my
database right?
(13:54):
So I think it's a little bit ofan education, a little bit of a
mountain for us to climb.
But once they realize that allthe techniques of typical
marketing outbound use cases,sales, outbound use cases that
exist that direct hires canstart leveraging on their own
internal databases, I think itunlocks a huge untapped
potential today in the market.
So that's what we are prettyexcited about.
But that's also the differencebetween the two segments.
Speaker 1 (14:15):
Yeah, yeah, that's.
That's really cool.
So what are some of the morerecent features, products that
you've been rolling out tocustomers that there seems to be
the most excitement around?
Speaker 2 (14:27):
Yeah, I mean there's a whole slew around actually the
capabilities of theorchestration engine, right.
So, as you can imagine, like inthe past, the orchestration
engine was built just like theRPAs, or robotic process
automation companies that youall might be familiar with.
It's like you know, if anaction is taken, then do this,
(14:48):
if something happens, then dothis.
So it's all very rules-basedautomation.
So eight years back, when welaunched the product, that's
exactly what we were doing,right?
That's exactly what HubSpotdoes, exactly what we do, that's
what RPAs do.
So we are moving from arules-based world.
And then, you know, then in therecent years, when the LLMs
(15:08):
came out, it's become prettymuch a goal-based world.
So we are moving into the worldof, like the same engine
orchestration engine that washard-coded to some extent and
then could do all thesecommunications and automations,
which is still great and added alot of efficiency.
Now we got to move to a worldwhere it's all goal-based and
agent-ticket nature and that'struly exciting and it's like
(15:31):
it's going to unleash a slew ofbusiness use cases that in the
past were very hard to stitchtogether for companies and
organizations.
And suddenly I think the eyesare opening and the art of
what's possible is coming true,right?
So what does that mean?
That basically means I can tellyou like.
So, for example, if you'rescreening a candidate, right In
(15:53):
the past it would be ask themquestion A, then ask B, then C,
then D, and there's a linearprogression of what you could do
right.
And if in the chat, thecandidate would say, hey, I have
a different question to ask you, the chatbot would just unwrap.
It's like, okay, what justhappened here and I don't
understand it and I'm going tohand you to a human or it's just
(16:16):
a failed chatbot experience.
Right Now, with chatbot or voice, we have this amazing rule base
hey, listen, you can have aconversation if you want,
especially voice if you want,for 30 minutes with this
candidate, but I want you to askthese five knockout questions
one way or the other, becausethat's the goal.
(16:37):
The ultimate goal is, is thiscandidate like an example in
light industry could be?
You know, if they can't standon their feet for more than
eight hours, we shouldn't reallybe wasting our time talking to
that candidate, right?
So now the candidate can spend15 minutes asking different
questions which might not berelated to that one that
screening knockout question butin the end, the AI is smart
(16:58):
enough to bring them back tothat question to ensure that
that knockout question is infact asked right.
So I think the goal-based worldis turning out to be super
exciting for us and there are somany use cases for us to unlock
that right there.
Speaker 1 (17:12):
Yeah, that's a lot of really cool things that the LLM
is able to do and have more ofthose interactive conversations
and loop back, as you mentionedtoo, which is really, really
cool, and I'm just looking atagain, like your website, and
there's just so many differentareas that we can dial into.
I'm curious on just to pivot alittle bit and talk about
(17:38):
sourcing.
I'm wondering if we could dialinto that for a minute.
I mean, I'm talking right nowbecause you're mentioning
accessing candidate pools thatare existent in CR CRMs.
You know there's some companiesdoing some innovative things on
the sourcing side.
Particularly, there's twoepisodes that folks should check
out Steve Vartel with Gem.
They're doing a lot of coolstuff with sourcing right now
Different functionality,different products and features
(18:00):
coming out and also the Workableepisode.
We had the CEO of Workable onthe show, nikos, and he was able
to break down different AIsolutions that they're
incorporating into their product.
Speaker 2 (18:21):
But yeah, I'm curious about, from a sourcing motion,
how you're thinking aboutleveraging LLMs to provide
better sourcing experiences andautomation and speed and quality
and that type of thing.
Yeah, yeah, I mean, that's somuch to unpack over there, but I
can start with a simple thing.
But first, before we getstarted on that, one caveat is
we do not bring a database toour customers, right?
So we assume that they havecertain sources of candidates
where they can access them.
The whole database world isriddled with issues and, I think
(18:42):
, ethical challenges, franklyspeaking, from where people have
sourced those databases.
So we try and stay out of that.
So what we see is one is openingup the aperture for the
organizations in terms of, youknow, giving candidates the
opportunity to apply in multipleways, right, so it could be QR
(19:03):
codes, text to apply career site.
They could be sitting to yourpoint, they could be sitting in
the CRM, they could be sittingin other databases.
There's also access to jobboard databases and things like
that, right?
So I think one layer one has tobuild as an HR tech vendor is
like how do you bring all thosedatabases together in a unified
search experience so when a jobdescription comes in, you
(19:25):
understand what it's looking for, then you can go and run a
really highly qualified searchon those multiple sources of
data.
So you have your passivecandidates that might be in any
of these data sets, you haveyour active candidates that are
applying inbound, and then youhave to mesh together and search
in that context and ifsearching is just not enough,
(19:49):
once you search, then you needto quickly reach out to them as
fast as humanly possible.
And that's where the automationlayer comes in right.
So all the way, extracting thejob description from the job
description, extracting likewhat is the semantic search that
needs to be run on thesedatabases, is the semantic
search that needs to be run onthese databases, and then from
(20:09):
there coming up with a list andhopefully that list at least is
ranked in some way or the otherright.
So you might have 100, 200candidates that match the
profile but you're looking fornow.
Once you have that list, thenyou need to reach out to them
within seconds.
So that's where the whole flowruns.
So you use an app to create ajob description.
From the job description youextract the search.
(20:30):
From the search you search onall these databases.
That whole flow is just anagentic flow, right.
And then you start thecommunication flow.
So now you start reaching outto those candidates.
Maybe you need to reach out to100.
So you start going one by oneand you start having either chat
conversations or voiceconversations.
Now what we've understood fromchat is chat is great and it
(20:53):
serves this purpose.
Not everyone wants to be on thephone and have a conversation
in public or even at home, butthe chat responses are pretty
short in nature, so the amountof information you can actually
get out of the candidate islimited.
So we've had a very successfulchatbot.
There are other chatbotcompanies.
(21:14):
It's great, but text is limitedand voice definitely leads to a
richer conversation and you canbuild a much richer profile
about the candidate through avoice conversation.
So anyway, but going back tosort of these knockout screening
questions, then those happenand then you have kind of create
a shortlist that yourrecruiters can work off of.
(21:35):
So hopefully, you know I'veexplained at least some part of
the sourcing side that is, yeah,that can basically become super
fast and agentic in nature.
Speaker 1 (21:45):
But yeah, that was what I'm curious about, because
that's I'm seeing a lot of usecases around sourcing as well,
and then going into thispre-screening use case, which I
think is really interesting,right, and so I'm curious,
though, from when you're gettinginto more so the pre-screening
where the AI agent is reachingout to folks.
What about from a consentperspective?
Is there anything that folkshave to consent in?
(22:05):
And then, like, how, how do youensure that you're able to do
that for your customers so thatyour agent can actually do the
outreach?
Speaker 2 (22:17):
Yeah, so, so, absolutely so.
These databases, hopefully.
Yeah, the only way we work withthem is you know if the client
has gotten the permission toreach out to that.
So, and we follow GDPR and allthose compliance and depending
even on sometimes which stateyou're dealing with, which
country you're dealing with, soit's all localized and we comply
with that.
And voice is also a newterritory.
Like just because you gotconsent for text doesn't really
(22:39):
mean that you got consent forvoice.
So there are nuances there.
But I think even there, likethe way it's pretty nice, like
the orchestration layer allowsyou to text somebody and say,
hey, I'm going to call you in afew minutes, would that be okay
to have a conversation, and thenthe person gives you consent
and then you have a conversation, right?
So at all points we areobviously following the opt-in
(23:01):
and opt-out criteria, which iseither maintained by the ATS or
maintained by us.
And it's bidirectional sync atany given time.
Of course we don't go blind, wedon't go and get into a
database where they have noconsent and we start spamming.
That's the worst experiencepossible.
Yeah, yeah, yeah.
Speaker 1 (23:20):
But you also mentioned for outbound.
So I guess for is there a usecase for when folks are like
recruiters are doing outreach tocandidates, like, let's say,
doing like human touch, likereach, linkedin and mails, to
then run into prescript Like howdoes how does the workflow work
if that's like it's newoutbound source candidates?
Speaker 2 (23:41):
Yeah, so you know.
So there is the.
As you know, all the CRMs andADSs have come.
You know, those are 20 year oldtechnologies, 25 year old
technologies.
They've come into the worldwith a view of, like it's a
recruiter or a sourcer use case,right, that's the insertion
point.
I want recruiters to come inand do a take certain action,
sources to come in and takecertain action.
We came in from an automationfirst world so we believe in the
(24:08):
future.
Like recruiters don't need tohang out for hours in the CRM or
the ATS.
Either way, they don't likethose, right?
So everything that I can dowhether I'm negotiating, whether
I am selling them the positionof what the value proposition of
this job is, and the company itcan all be done offline without
the ATS or the CRM involved.
Right, and now the AI is justhelping record all those calls
(24:29):
and put them out right.
So, yeah, so our approach isdifferent, you know.
So we do allow for if therecruiter or the sourcer comes
in and has to build a talentpool first manually, they can do
it with the platform.
Ideally, we just like to buildeverything that is just fully
(24:49):
automated, right?
The future is pretty simple.
You have your annual operatingplan and your annual hiring plan
uploaded in the system.
Adss have that capability.
From that you extract all theroles people are hiring for.
You know exactly the roles thatare coming up.
You potentially know the rolesin which month they need to hire
.
You have the data to tell youlike how, what kind of lead
times you need.
You automatically create thosetalent pools.
A recruiter doesn't need to doanything and you start building
(25:17):
those talent pools and startmessaging and asking those
people for having conversations.
You're having voiceconversations, chatbot
conversations.
Hey, I know you were a silvermedalist six months back.
Where are you today?
What's your job title?
Would you be looking for anopportunity in the next six
months?
And you know which candidatesare warm, which candidates are
hot, which candidates areactually cold.
They don't want you to talk tothem again, you know.
So that's where all of this isheaded, and I think there's a
(25:41):
mix of AI and human power willalways remain, but I think the
human, the amount that will behuman powered, will keep
diminishing over time.
At least, that's ourperspective.
Speaker 1 (25:54):
Okay, yeah, for sure.
What about?
So can we dial in on AIevaluation and how you're
thinking about that?
So you have pre-screening rightFor just like the top knockout
questions, right, which is afairly straightforward use case,
and then you start to get alittle bit down funnel and I'm
wondering if that's a use caseyou're thinking about.
(26:17):
I think we're seeing probablymore demand top of funnel or you
probably have more datasurrounding this than I do, but
my assumption is top of funnelbecause it's higher volume.
It's just a lot more work,right.
And then when you get likeparticularly for professional
service knowledge work, by thetime you get to like a second or
a third round, potentially peropening you have fewer
(26:37):
candidates.
So maybe there's potentially,arguably maybe, a little less
demand in the market currentlyfor down funnel evaluations, and
maybe there's also a trustaspect and maybe there's also
concerns around potentialupcoming regulation or whatever
else.
I don't know.
I don't know.
I mean, I guess I'm justthinking like how are you
thinking about like deeperevaluation use cases and is that
(27:01):
something, an area you're goingto play?
Speaker 2 (27:02):
yeah, no, I think I think you kind of nailed it like
.
The first part is top of thefunnel, is a very much an
efficiency use case, right, and?
And also a use case where theTA leaders are realizing my
recruiters are getting boggeddown by the sheer scale of
candidates that are coming in.
There's just no way, and Ithink if you're a TA leader that
has not made an investment oris not thinking of making an
(27:23):
investment at the top of thefunnel to do something about it,
I think you're failing yourorganization, right.
So you need to like, reallybring some of these technologies
, because they give youefficiency, they give you scale,
they reduce the noise, so, butthere's no value, not
necessarily like code evaluationhappening right.
And, by the way, this is validnot just for professional roles,
even for high volume roles.
(27:43):
In high volume roles, let metell you, like, if you're a
large BPO center, you arespending tens of millions of
dollars on assessment toolscenter.
You are spending tens ofmillions of dollars on
assessment tools.
They don't want the noise tohit the assessment tool, they
don't want fraudulent applicantsto hit the assessment tool.
They want to keep screening,keep knocking out people as much
(28:04):
as they can to only have aselect list of candidates.
At the end of the day, inprofessional it might be I just
need 10, like, give me reallywell screened folks that could
match this role and then I willgo super deep with them.
That's why I also want myhiring managers to spend a lot
more time right.
So, yeah, so to your point.
(28:24):
There is this screeninghappening, there is the noise
reduction happening at the top,and companies are investing
massively.
But they're also trying toavoid costs that come later,
either cost to the hiringmanager, cost for assessments,
cost for background checks, andif they are low margin
businesses that need that rapidhiring, they really want to
(28:46):
reduce any of this OPEX thatthey can.
So it's a real ROI pitch forthem.
On the evaluation scale at thebottom, towards sort of the as
you get closer to the hiring,yes, I think interview tools
that are at least able totranscribe what happened in the
interview and very quickly sortof compare those to the
scorecards that people mighthave to use to evaluate.
(29:09):
You know the final.
Maybe it's the silver medalist,the final person they help
picking, but at the end of theday it's still core human in
nature, according to us, and itwill probably remain there, at
least in knowledge, work andprofessional, for a while,
that's not going away.
Yeah, I mean, that's a goodthing, and I think that's a good
thing.
(29:29):
You want it that way.
Speaker 1 (29:31):
Yeah, yeah, I think so.
I mean, I think that's theconclusion that I've come to as
well just from hosting thisseries, talking with a lot of
brilliant CEOs such as yourself,right, learning from folks who
have a lot of access to data andare constantly having these
very high level strategicconversations with customers,
and so I think that that makes alot of sense for me.
(29:54):
Another question that I have isyou know, you have companies
like BrightHire and Pillar,companies that are doing like I
suppose at this point they'recalling it interview
intelligence, essentiallyjoining, co-piloting, zoom
interviews, which I'm sure, iftheir CEOs heard me explain it
that way, maybe I should let mejust, it's a lot more than that,
right, but basically they'rejoining recruiters on the
(30:16):
interviews.
They have all the rolerequirements on the back end and
they're essentially they haveall the custom questions and
packaging the data.
To summarize, you know how well, how many of the questions that
folks answer and hey, did youforget to ask some?
So they push that along to thenext interview.
They're, of course, generatingjob descriptions and doing some
of those generation tasks toofor customer questions, but
(30:38):
they're really ensuring thatcandidates are going through a
consistent, thorough,comprehensive interview process.
Is that an area that your teamis building in that direction,
or are you staying away fromthat?
Speaker 2 (30:52):
Yeah, we are staying away from that, but we like
aspects of that.
We help enable that in the way.
We have a large number ofcustomers that have bought our
interview automation schedulingproduct, which is a very big
need of the hour, right.
So you're talking about theactual interview process and the
interview itself.
I'm talking about the processaround the interview, which is
(31:15):
key because hiring managers arefrustrated, recruiters are
frustrated, candidates arefrustrated Everybody's
scheduling, rescheduling hiringmanager are not prepared and
they're coming into theseinterviews right.
So that's a big problem, hugepain point.
We have some of the largesthealthcare hospital systems that
we sell that product to.
We have, I think, the fourthlargest, the fifth largest
automaker that is using us.
(31:35):
So tons of examples.
The example is as simple asthat.
I have 20,000 hiring managers.
Please fix interview automationscheduling for them right.
So it's such a natural extensionof an automation platform.
Right Now.
Your point where you're comingto is we've seen this work like
(31:56):
consistency is required Beyondconsistency.
That's why chatbots and voiceAI is great.
It's asking the exact samequestion over and over again to
tens of thousands of candidatesfor that particular role.
That's great.
It's removing actually biasfrom that screening right that
used to exist.
So same thing is absolutelytrue, probably for the interview
(32:16):
platforms, like at leastthey're helping the hiring
managers ask the same questionsand be consistent, I think.
Number two instead of talkingabout the candidates, it's
important to talk about thehiring managers.
We know so many hiring managerseven probably me we have gaps
when we interview candidates.
We are probably not asking thequestions the right way.
Speaker 1 (32:34):
There's a lot of training that we can provide to
these hiring managers,especially the first-time hiring
managers, and if there's an AItool, ai co-pilot that's helping
me train my own internalmanager teams to ask better
questions, to put the best footforward.
That's a great win, I think.
Yeah, I mean, I think it's likealmost to some extent there,
and that's a whole other side ofthe product too that I'm glad
(32:56):
you're bringing up, becauseBrightHire and Pillar are also
focused on essentially likeevaluating what the hiring
managers are doing.
I think even BrightHire has afair amount of this
functionality where they'reactually, I think, data like,
even even like stuff like didyou start the interview on time?
Did you ask questions the rightway?
Did you spend five minutestalking about the football game
you watched on Sunday and notreally dialing into the
(33:16):
screening questions?
Right, just helping them stayorganized and producing a score?
Almost like, maybe like a gongright On the sales side they're
doing some of that kind ofanalytics functionality on the
call itself, which I think ispretty cool too.
But yeah, it seems like they'vegone like the interview
intelligence, like co-pilot,packaging data, showing the
candidates answers, responses,how relevant it is to the role,
(33:38):
requirements, and then alsoproviding analytics and data on
how the hiring team is doinginterviewing.
You know cause I?
You know it is havingconsistent interviews.
Having hiring managers performat a high level when it comes to
hiring and interviewing ischallenging, we interviewers.
Having hiring managers performat a high level when it comes to
hiring and interviewing ischallenging.
We had Daniel Chait on the showa lot, the co-founder and CEO
of Greenhouse.
There's something funny healways likes to say is hiring
managers struggle with twothings hiring and managing Right
(34:01):
.
Speaker 2 (34:01):
So you know, yeah, no , no, absolutely.
Speaker 1 (34:04):
And the same thing.
Speaker 2 (34:05):
Like you know, when we have the text message, we
have a whole text messagingsuite.
Right, as I told you, therecruiters have.
We try and measure the samething.
We try and measure how manymessages they send, what kind of
speed do they have?
How fast do they respond tocandidates?
Which recruiters have betterresponse rates?
And we are also trying to justgive these sort of signals back
(34:26):
to the TA teams.
And now the teams are like okay,it's not the recruiter that is
pointing the finger.
Now it's like I actually have atranscribed chat of yours in an
interview of Mr Hiring Managerand here are the things that
went wrong.
Right, so I would like you toimprove, because that's leading
to a poor candidate experiencefor us.
Right, so that loop, like ourautomation, can measure the
(34:49):
candidate experience at anypoint point before the interview
, before the screen, after thescreen, after the interview,
after the disposition, the finaldecision was made and we can
give that intelligence back.
Right, so it's super importantand I think LLMs and AI is
really going to help change thegame, as long as they're not
making the final decision, weare all good, but yeah, you know
(35:11):
, I think it's a truly excitingtime.
Speaker 1 (35:15):
Yeah, it's a lot of really cool, really cool stuff
happening for sure.
What about, from a regulationperspective, for companies that
are a little bit concerned aboutincorporating AI into their
workflow?
Right, I mean, like right now,for instance, like one thing
I've brought up several times onthe show is, like workday right
now there's a class actionlawsuit for their AI system,
(35:37):
potentially like discriminatingfolks over several different
reasons, right, but you know,people are, I think some folks
might be a little bit concernedabout applying AI technology and
I'm wondering, like you know, Ithink evaluation is a key
sticking point on AI making afinal decision, or, and I'm
curious to get your thoughts onthat I mean, I think, honestly,
maybe there's nuance to thedefinition of what we consider
(35:59):
evaluating Right and, and youknow, is it considered?
For instance, like, is itconsidered?
Maybe it's it's more soevaluating if, if ai is
generating role requirements,generating a jd, and then
basically deciding what the rolerequirements are and then
evaluating against those, maybe,maybe that's what evaluating is
(36:22):
.
Maybe, if a hiring team putstogether the role requirements
and then essentially, the is itand the AI is evaluating or is
interviewing for that, Is thatreally evaluating or is that
just matching Right, like, andso I'm just wondering, like, how
do you think like the nuancesis like where?
Where do you think regulationis going to kind of fall in this
in the evaluation AIdecision-making piece?
Speaker 2 (36:43):
Well, yeah, yeah, I'm no EUOC expert or regulation
expert, so I won't, I won't makeany statements around that, but
yeah, I mean, listen, it'sfairly complicated.
It does to your point.
It does start from the JD.
Everything falls flat from theJD, the noise starts from the JD
(37:04):
.
You just copy something fromthe job description, from Google
or some past job description,and then you just run with it.
Right, that's how the screeningquestions are generated.
That's how the screeningquestions get on a chatbot,
that's how they get on a voiceAI.
You know, if you made themistake in the job description,
everything flows from there.
Right, that is the freakingreality.
Right, and nobody wants toaccept it, right, but that is
(37:24):
the reality.
So you know, the kind of thingsthat we are experimenting is
like imagine a hiring managerable to talk to a voice agent
and tell them exactly whothey're looking for, the core
skills I'm looking for.
Hey, it is still a forkliftoperator, or it is still a
cashier at a franchise.
But like, this is what makesthis particular role special, or
(37:45):
this is what makes thislocation special, you know, and
let me tell you more about it.
And can the voice agent capturethat and really humanize the
job description where, yes,there are certain knockout
things like you have to be here,but here's where I can give you
more context around that role,even though you know I'm Bath
and Body Works and you'reLululemon and you're Nike store,
(38:05):
but it's still a store job.
But here's what makes us trulydifferentiated, right?
So we are going to try and pushthe boundaries on that front,
but I did want to make the pointlike, job descriptions are the
worst and they start the wholeproblem chain, right, but there
is bias, potential bias.
So what are we measuring itagainst?
I think the regulations need tounderstand we are measuring
(38:26):
against humans, and I think it'stoo much bent on.
Hey, ai is biased, but humansare not.
The reality is, when we takethe human data and we feed it
into AI, ai is horribly biased.
That's when we realize, oh shit, we shouldn't be talking about
which school they went to, whatwas their name, what was their
(38:47):
gender, what was their location.
We have to keep all the dataout before we train the models.
Right and well, that data lies.
When, as soon as that resumehits a human, there's no way the
human's not looking at it,right?
So I think one is trulyfiguring out what has to be
screened out of these resumes.
(39:08):
The good news is for a lot ofjobs and I would say I don't
know whether you know the statslike literally 65% of the
world's workforce is not onLinkedIn.
How many of those peopleactually have a resume?
So what we are matching towardsand what we are training our
models towards is what I wouldcall progressive matching, and
(39:33):
what that means is there mightnot be a resume.
There's a good chance thatthere won't be a resume, right?
So I want to understand who youare.
I want to have a conversationbased on the skills that the job
is saying it absolutely musthave, but then I'm trying to
understand your preferences.
You know, can you work overtime?
Can you do the shift or not?
Can you work on the weekends?
Do you even have a transport toget to work?
(39:55):
I'm trying to understand who isAnil as a person, and even if
you tell me I've done customersupport work before, okay, but
were you in a situation wherethere was a blow up?
Can you talk to me more aboutthat?
So, trying to drill a littlebit deeper than just what a
resume might have, and even ifyou have a resume, a resume is
just another data point to use,but you've got to use the job
(40:19):
seekers' preferences, what theyexpect.
You've got to take into accountwhat the manager is saying and
then try and intersect it.
So the future, we believe, istry and remove the biases from
the training models as much asyou can.
Get yourself audited all thetime in real time.
Hopefully Make the AI moreexplainable.
If the AI matches a certainperson to a role, can it
(40:44):
actually explain why it did that?
And then, if you make itcompletely blind, will it match?
Will it match the same personagain?
You know that's where it getsvery tricky.
So I think explainability is ahuge sort of actually a killer
feature if you can get it goingin your product, and then you
need to make sure that thatexplanation satisfies the
auditors at some point when youaudit it right or can pass a
(41:08):
bias audit right.
So that's those are all theinvestments that at Sense we've
made to make sure that we aredoing it the right way and we
are following at least the localregulations.
I will say one last thing whichpeople miss.
We are so focused constantly oncandidates, inbound candidates
coming in, screening them.
It's not just screening anddispositioning them and saying
(41:31):
you're rejected, it's actuallyscreening them.
You know, it's not justscreening and dispositioning
them and saying you're rejected,it's actually screening them,
saying that you're not a goodfit.
But here are three other jobswhere you are a better fit for
and let me help you apply tothose jobs right, especially in
the high volume scenario,because that's truly possible.
But even if you're a technologycompany, maybe you apply for an
AI job and you're not a greatfit for the AI job.
(41:52):
You don't have that level ofexpertise that was required, but
maybe I can deflect you to aregular software engineer job
and get you hired there.
So the whole point of the systemthat we are trying to build is
and actually our data is proving70% of the candidates that
apply to a job are not a goodfit for that job.
(42:14):
They just think they are, theybelieve, they want to believe
that they are.
It's a aspirational apply, it'snot a true apply.
But if we can deflect them andput them in roles and jobs and
show them the jobs that theymight be a better fit at, we can
actually improve the conversionrates, we can improve the fill
rates and we can make both sideshappy.
(42:35):
Right, we can explain to themwhy.
Why are we taking the actionthat we are taking?
So that's our vision and that'swhy we think the candidate
experience also needs to change.
The good news is a lot of TAleaders understand that and they
know that that's what'sactually happening in the market
.
Speaker 1 (42:51):
Lot of TA leaders understand that and they know
that that's what's actuallyhappening in the market.
Yeah, wow, this has been anincredibly insightful
conversation.
I've certainly learned a wholelot.
I know our audience is going toreally enjoy this and I would
actually one day like tocontinue the conversation if you
want to come back on the showman.
You just shared so much greatinsight with us.
This has really been fantastic.
(43:11):
I think it's really incrediblewhat you're building over.
At Sense, it seems like areally impressive platform, a
lot of different functionalityto help out through the entire
workflow.
So I'm really excited tocontinue to follow what your
team is building and it soundsit just sounds all like really
impressive and incrediblyhelpful to your customers.
So, anyways, thank you forcoming on and sharing a little
bit about what you're doing.
Speaker 2 (43:32):
Likewise Thanks, james, and let's stay connected.
Speaker 1 (43:35):
Yeah, absolutely, and for everybody else tuning in.
Thank you so much for joiningus and we'll talk to you soon.
Take care.
Popular Podcasts
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!
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.