August 6, 2025 • 37 mins
Recruiting has always been a race between speed and quality. But what if you could have both?
In this episode, we’re joined by David Paffenholz, co-founder of Juicebox AI, to explore how AI is transforming talent sourcing as we know it. Born out of his frustration with traditional recruiting, Juicebox lets you search through 800 million profiles in minutes, surfacing candidates traditional keyword matching would miss entirely.
David breaks down how Juicebox uses large language models (like ChatGPT) to go beyond keywords and actually understand candidates' context, skills, and potential. We dive into the platform’s “Juicebox agent,” a sourcing assistant that never sleeps, soon able to launch searches automatically based on your ATS activity.
We also discuss market trends, including the BLS's job count correction, return-to-office dynamics, and how AI tools like Juicebox can provide staffing firms with a competitive edge without compromising on quality.
If you’re ready to spend less time sourcing and more time hiring, this is the episode for you.
🎧 Try Juicebox for free or book a demo at juiceboxai.com
Additional Resources:
1. Innovative Sourcing Techniques for Recruiters
2. Can You Trust AI to Handle Recruitment?
3. How Is Artificial Intelligence Changing the Recruiting Process?
🧠 WANT TO LEARN MORE? Be sure to subscribe and check out 4 Corner Resources at https://www.4cornerresources.com/
👋 FOLLOW PETE NEWSOME ONLINE:
LinkedIn: https://www.linkedin.com/in/petenewsome/
Blog Articles: https://www.4cornerresources.com/blog/
👋 FOLLOW DAVID PAFFENHOLZ ONLINE:
LinkedIn: https://www.linkedin.com/in/david-paffenholz/
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Pete Newsome (00:00):
You're listening to the Hire Calling Podcast,
your source for all thingshiring, staffing and recruiting.
My guest today is DavidPaffenholz from Juicebox.
David, how are you today?
I'm doing well.
Thanks for having me.
It is great to have you.
I've been trying for a while.
You are a busy guy to get ahold of.
That must mean business isgoing pretty well.
David Paffenholz (00:18):
It's been an exciting time at the company,
lots of new things happening.
Pete Newsome (00:32):
Team is growing, customer base is growing and
super excited to be on todayAwesome Well.
David Paffenholz (00:33):
I'm happy to have you just quickly set the
stage and describe Juicebox in anutshell, if you could.
For sure, juicebox is an AIrecruiting platform.
We help our customers find thebest talent by aggregating a
bunch of different data sourcesand then using natural language
search to surface the besttalent.
We then have some additionaloutreach workflows, including
emailing, managing thosecontacts and more, and so I
think of us as an all-in-one,top-of-funnel AI hiring solution
(00:54):
.
Pete Newsome (00:55):
Excellent.
So, needless to say, ai alittle bit of a hot topic these
days, but we'll get into that ina minute too.
First, tell me about yourbackground a little bit.
How did you get to where youare now?
David Paffenholz (01:06):
For sure.
So, as one might be able totell by the last name, I'm
originally from Germany.
I came to the US first forcollege and then stuck around
since I started the business,together with my co-founder, in
2022.
And so this was shortly beforeChatGPT came out and everyone
started talking about AI, andour kind of initial vision, or
(01:26):
what we wanted to focus on, isbuilding tooling to make the
recruiting process easier,particularly on the search side,
and so that's where we started.
The reason we were passionateor interested in that space is
because both my co-founder and Ihad always found employment
through kind of irregularmethods.
We were basically cold,reaching out to jobs or roles we
thought were interesting ortrying to create those roles for
(01:47):
ourselves, and so it made usbelieve that there's still a lot
of friction in that hiringprocess and still a lot that can
be solved, which also made usinterested in building the
product.
Pete Newsome (01:57):
I love that, and you know there's I think there's
not many people who are youngand coming out of school finding
the job search process easythese days.
So not a unique situation, butyou chose to tackle it in a
unique way, which I love Notjust complaining about it, but
doing something about it.
So good for you for reachingthat conclusion.
(02:19):
Let's talk about the state ofthe market today a little bit.
I don't know if you saw thatthe Bureau of Labor Statistics
just happened to put out theiremployment report this morning.
Did you see the numbers?
David Paffenholz (02:31):
Yeah, so there was like a large correction
shift.
Pete Newsome (02:34):
Yeah, a massive one, right that almost 300,000
jobs were overreported over thepast two months, and so a lot of
us in this industry and a lotof job seekers are feeling that
the market is not easy.
Right, the government for thelast couple of months, has told
us that it's easier than we'refeeling, and this told a pretty
(02:54):
big story.
I think that the job market isnot growing as well as we'd like
it to, and it's really hard tofind a job.
It's really competitive,recruiting is tough.
So that's my perspective on themarket right now.
What are you seeing?
David Paffenholz (03:08):
Yeah, I think there's kind of very different
scenarios in differentsubsectors of the market, and so
in some spaces we're seeing themarket really heat up and kind
of becoming very competitive,and other sectors we're seeing
kind of a lot less of that.
And I think the labor marketcorrection is a good reflection
of that too, where thestatistics sometimes say one
(03:31):
thing, the reality on the groundsays another thing, and then
the reality on the ground alsochanges depending on the role,
the location and even thatspecific point in time.
And so we've seen quite somedrastic shifts amongst where our
customers are hiring, wherethey're focusing on their
searches, even what industriesthey're focusing on In that case
, of course from the recruitingagency perspective as well.
Pete Newsome (03:53):
You know, and I always feel like and I'm biased,
of course, because I own astaffing agency but I feel like
we have our finger on the pulseas an industry of what is
happening at any given time andthat's why, for me personally,
some of the numbers weresurprising.
Right, you look on LinkedIn, itdoesn't sound like it's a great
job market right now and when Italk to my peers around the
country, we've felt I've pickedup on some struggles that a lot
(04:16):
of companies are having thesedays.
But I wanna know if you couldshare which sectors do you see
they're doing well?
David Paffenholz (04:25):
Yeah, I'd say the one where we've seen the
strongest pull in is like thetech sector, especially here in
San Francisco.
We've seen like a lot of surgein hiring there, really across
roles, but particularly forin-person roles here in the
broader Bay Area, and so thatpart of the market seems to have
really been accelerating andthat's like roughly 20 market
(04:45):
seems to have really beenaccelerating, whereas and that's
like roughly 20 to 30% of ourcustomers so it's a sizable
chunk but it's not the majorityof our customer base and so
we've seen that accelerationthere, and then in other parts
of the market perhaps less soand perhaps less of that pull.
I think it's interesting too,because it's almost like an
inverse, where the tech sectoralmost has this opposite
(05:06):
relationship to the broaderhiring market and in times where
the tech sector is doing well,the broader economy might not be
, and vice versa, where we sawthe opposite two years ago.
And so for us as well, where weserve both customer groups, we
always see some discrepancy inthose as well.
Pete Newsome (05:24):
What are you seeing locally out there with
return to office?
It's kind of subsided.
I haven't heard as much as wehave for the past couple of
years here in Florida, where Iam.
I think it's business as usual.
For the most part, those whoare going back have already gone
back and there's a lot ofhybrid roles out there.
But what are you seeing?
David Paffenholz (05:42):
Yeah, I think most companies and most
companies have kind of madetheir decision, like, either
they are going to be, you know,fully in person they're hybrid
or they're fully remote, and atleast at this point they really
should have made that decisiontoo, because, you know, it's
been a while since the world hasstarted to normalize.
In our case, our team is inperson.
(06:03):
We see a lot of companies thatare kind of fully remote or have
chosen that hybrid path as well.
I will say that, from whatwe've seen, the in-person market
has just become a lot harder torecruit for because it is a
constraint on how manycandidates can take on that role
.
Pete Newsome (06:21):
Yeah, and I think so much of that comes to supply
and demand right Candidates whohave a rare skillset, which has
always been the case, but nowit's in many ways targeted at.
Are you going to work in theoffice or do you have the luxury
of having that skillset thatallow you to call your own shots
more and work at home?
Are you seeing that?
Because I really think supplyand demand is driving a lot of
(06:43):
it.
David Paffenholz (06:43):
Yes, yeah, definitely.
I also think there's like someinteresting parts where, like
roles that used to we used tothink are like in-person
mandatory because of COVID,we're like forced to shift to
remote, and then they've kind ofjust there's been this
consensus of like, you know,some of these roles can
completely be done remote, whereif we had just asked people
five, six years ago, is thateven possible for these roles, a
(07:05):
lot of people would have saidno, and so I think there's kind
of been that rethinking in themarket which has really changed
what people think about in termsof demand as well, of like what
roles are really needed to bein person versus not.
Pete Newsome (07:17):
So that's a perfect segue into what we're
really talking about in the bigshift is AI.
We're really talking about inthe big shift is AI.
What wasn't possible in thevery recent past is now a
reality that we're having tofigure out how to deal with
right, not just in the staffingindustry, but as a whole.
So what's your general take onAI and how it's impacting the
(07:41):
job market already, and then howit's going to impact it in the
future?
David Paffenholz (07:45):
Yeah, I think AI has had two big impacts on
the industry.
I'd say like one maybe on thelabor market more generally and
then the second specifically onlike staffing and recruiting.
And so I think in the labormarket kind of more broadly, it
means that the output of everysingle role I think the
expectations have increased andthe ceiling of what is possible
(08:09):
has increased because people arestarting to use AI tools,
starting to be more efficient inparts of their workflow and
achieving more.
But that then also means thatin some cases, the expectations
for those roles have gone upbecause the kind of normal is
now at a different rate.
I think overall that's a greatthing because productivity
increases, the economy growsfaster and we're able to do much
(08:29):
more.
At the same time, it also makesit particularly tricky for roles
where maybe that AI adoption isa bit slower or employees don't
provide the opportunities toadopt AI or the tooling for it.
I do think in most sectorsthere is now at least one AI
tool that can really simplifyworkflow, if not multiple, I
think, in recruiting.
(08:50):
You know, despite us being inthe AI recruiting space, there's
also a lot of other goodcompanies in the AI recruiting
space and I'd say there's likethere's a good amount of
different solutions to evaluateand to test from, and so I think
, yeah, there's there's a lot,right to say the least, and it's
evolving rapidly.
Pete Newsome (09:07):
That's one of the things that I struggle with.
One of the reasons I wanted tospeak with you today, quite
frankly, is to share a messageof what's even out there,
Because when I speak to my peersaround the country, they
struggle to keep up.
But I'm curious about adoption,how your perspective may differ
than mine.
Have you seen any trends amongcompany size, you know being
(09:31):
faster or slower to adopt, or isthere generational differences
that you see on?
You know, if you had to put anycategories in place, can you do
that?
Or is it just really all overthe place?
David Paffenholz (09:44):
Yeah, so what we've seen that gets me really
excited and that I think is kindof unique to this time as well,
is that individual users inmany cases, or like individual
team members, can find AIsolutions and start adopting
them for their own workflows.
If they find that it works forthem, it can kind of spread
within an organization withoutneeding to be like centrally
(10:05):
purchased solution, and so Ithink a lot of AI tools have
started positioning themselvesthat way too.
It's on like the simplest level, say a chat GPT where everyone
can sign up, create their ownaccount and start using it in
different ways, but then evenworkflow specific tools where an
individual user can sign up,try using it for free and then
might spread adoption acrosstheir team.
For us, almost all of ourcustomers start off by having
(10:28):
one or two users internallyhappen to come across the
product or maybe have heardabout the product, test it out
and see if it works for theirworkflow before showing it to
the wider team or looping intheir manager to do a kind of
more traditional evaluation, andthat kind of bottoms up
discovery I think is reallypowerful, because it also sets
the bar a bit higher.
For, you know, does this reallywork, or is this just a really
(10:49):
good sales pitch?
Pete Newsome (10:51):
I really like that and I hadn't thought of that
before what you just said before.
But as I'm listening to you, Iunderstand exactly what you mean
and I've seen that over andover and that is pretty unique
in terms of technology adoptionin my lifetime probably everyone
else's where you really do haveaccess at an individual level
(11:12):
without having the company haveto make the decision as a whole,
and what a great thing I mean.
As an employee.
Over the years I've had lots ofideas where, when I was an
employee which was early in mycareer I would often get
frustrated by an inability tohave the company adopt new
things right.
I've worked for two largeemployers prior to starting Four
(11:33):
Corner and they've moved veryslowly.
I don't think you can do thatanymore and I think companies
have to listen to theiremployees and would want to
listen to their employees whohave the willingness to go and
seek things out on their own.
David Paffenholz (11:45):
Yup, yeah.
And frankly, it's also a greatway where you know if companies
notice that, say, two employeesare really outperforming or
suddenly producing a lot moreoutcomes and output, and you
know.
Then one asks, hey, like whatare you up to?
Or what are you doing, hasanything changed?
And maybe they're using newtooling or they're trying new
workflows.
That can be a really coolevidence point as well, where it
(12:08):
almost becomes a no brainer forthe organization to at least
consider those resources orthink about what a wider
adoption could look like too.
Pete Newsome (12:15):
And yeah, so true, and I love it, and you know as
well as I do as a businessleader, there is nothing better
than an employee coming to youwith a solution to a problem you
didn't even ask him to solveright.
David Paffenholz (12:26):
I mean, how great is that it doesn't get me
better?
Yup yeah, Especially when it'slike a problem that you know,
it's always kind of being in theback of one's mind and like
never really had the chance totackle it or think about what
one can do about it, and thenjust seeing that solution in
front of you, it can be a prettynice feeling.
Pete Newsome (12:42):
A hundred percent, all right, man.
Well, let's get into juice boxspecifically now.
So I can tell you that theentirety of the time I've owned
a staffing company which iswe're in our 20th year we've
always tried to find a betterway to do things right, gain
efficiency, gain speed, gainthoroughness, and there's been
(13:03):
very little potential to do thatup until recently.
So I couldn't be more excitedabout solutions like yours, and
I know that as an industry, wedesperately need new solutions
to improve.
So just break it down from thebeginning what is the premise in
terms of what is the biggestproblem Juicebox solves?
(13:26):
And then let's get into somedetails For sure.
David Paffenholz (13:29):
So I'll start off by kind of describing what
we see as the status quo, orwhat our users typically do
before switching to Juicebox.
Then I'll contrast that withwhat Juicebox does.
And so currently, what we see alot of our customers do before
adopting Juicebox is when doinga candidate search.
They have a pretty manualprocess.
There's usually two places theygo.
One is their existing ATS ordatabase where they look through
(13:50):
any candidates that they mightalready know that could be a
good fit.
Usually involves perhapssetting a filter, running a
search depending on the softwarethat they use, and then kind of
manually reviewing thoseprofiles In some cases pretty
long lists, and ends up beingfairly time intensive, Also
usually not the best interfaceto do it with.
And then the second thing thatwe see our customers do is going
out and finding net newprofiles for those roles, and so
(14:12):
you know, be that sourcingexternally through a channel
like Indeed, LinkedIn or others,where we kind of manually also
enter those filters and thenstart reviewing, and then
finally we'll want to engagethose candidates, be that
through emailing, phone calls,etc.
That whole process ends upbeing a large share of recruiter
time spent.
(14:33):
That is exactly the workflowthat Juicebox aims to solve, and
so we do that in a fewdifferent ways.
The main one is we combine yourexisting data with our larger
data set, where we have over 800million profiles globally.
We've indexed them, added abunch of additional data into
them and provide one unifiedsearch.
And so, rather than having togo into those different systems,
(14:54):
describe who you might belooking for the role you're
hiring for, you can just typethat into Juicebox or put in the
job description.
We'll automatically go out andsearch for both the best
profiles who are already in yourdatabase as well as net new
profiles that you shouldprobably consider for the role.
From there, we'll then rankthem for you, and so we'll go
skill by skill.
Do they match what you'relooking for, criteria by
(15:15):
criteria, and present it to youvery similar to how a human
researcher or sourcer might.
They'll go profile by profileand say you know, this criteria
is met, but maybe not this oneor this profile could be a good
fit for this reason.
We'll then present those topmatches to you and let you reach
out to them in one click,either through an email sequence
or, if you prefer, to reach outby phone.
We'll provide that data too,and so that whole process,
(15:38):
instead of taking 10 plus hours,can go as fast as 15 minutes,
going from initial search tohaving that output and then
being able to take action fromit.
Pete Newsome (15:48):
And everything you described is for a human to do
is what I consider to below-quality time.
It is necessary, but we wantour recruiters engaged with
candidates directly right, wewant them connected and
everything that leads up to thatagain necessary.
But you'd like a better way todo it, a more efficient way.
So tell me about the 800.
(16:08):
I'm intrigued by that 800million.
Are those existing in yourdatabase, or do you have access
to them in a database likeLinkedIn?
I mean, where do thosecandidate profiles reside?
David Paffenholz (16:21):
Yep.
So that's aggregated frompublic profile data.
So, be that profiles thatcandidates may have on
professional networking sites,be that profiles they may have
on social media sites like aFacebook or Twitter, be that
other information they mighthave on a company website or on
a personal website, depending onwhat's available, it varies a
(16:41):
bit from from role to role aswell.
So, for example, for the healthcare sector, there's some other
data sources available, or evenpublic licensing information
that might not be available forother sectors, and so we kind of
see what's out there, dependingon the sector, aggregate that
data and then put that into amore what we call unified
profile, basically justcombining those different data
sources into one source of truth.
Pete Newsome (17:02):
Okay, now is a tool, so do you have better can
it profiles or more can itprofiles in?
Do you have a strength area inparticular, or are you pretty
agnostic when it comes to theposition types?
David Paffenholz (17:13):
We're pretty agnostic.
I'd say, geographically we'restrongest in North America and
it's where the majority of ourcustomers are, but then within
that we're very role agnostic.
Pete Newsome (17:25):
Okay, nice.
So give me customer sizetypically.
Do you have an ideal customersize?
Do you have an ideal number ofpositions that you really see
where juice box is mosteffective?
Paint that picture for me.
David Paffenholz (17:40):
Yeah, it's become quite broad.
So we now work with over 2,500customers, ranging from
individual recruiters who maybejust set up their own agency or
started freelance recruiting,all the way to some pretty large
enterprises where they have,say, a 100-person recruiting
team, or larger recruitingagencies where they have
similarly sized recruiting teams.
I'd say the sweet spot to getstarted is once there's at least
(18:04):
five recruiters on the team,there's proper processes in
place and the baseline ofprocesses already exist.
That baseline becomes even moreimportant when adding in the AI
layer on top, because ifthere's not a consistent
internal workflow, then it's tooearly to add the AI or
automation on top, because itmight just lead to more
confusion than benefit.
And so that's usually what Irecommend is that, like a
(18:26):
baseline to get started is youknow, are those workflows
defined, are the processesdefined, and then are we ready
to automate those?
And oftentimes we see thataround like the five recruiter
mark, though it can vary too.
Pete Newsome (18:37):
So more effective if there's a certain level of
maturity in the process andexperience.
More effective if there's acertain level of maturity in the
process and experience, whichmakes sense.
So I do my search right and Iget to return profiles.
Do I download them into my ATS?
For the ones I want to contact?
Do they stay within Juicebox?
How does that process work?
David Paffenholz (18:57):
Yeah, so that full process things with your
ATS.
So, as you've discovered thosetop profiles, let's say there's
maybe top 50 that were matchedfor you and 20 of those already
exist in your ATS.
So we'll flag that for you andthen the remaining 30 might be
net new, and so you can thentake the 30 net new ones, save
those into your ATS as well andthen reach out to all 50 of them
and make sure that that data islogged there too, and so all of
(19:20):
that data ends up living inyour ATS2.
So you still have that as yoursource of truth, while also
having it available in Juiceboxas you're doing new searches.
There is one additional elementto the platform which I didn't
mention earlier, which is theJuicebox agent.
The Juicebox agent kind ofworks on its own and it starts
doing its own searches and itsown outreach without necessarily
requiring you to review eachprofile.
(19:41):
And the reason I'm bringingthat up is because it can work
in parallel with a humanrecruiter, where the agent goes
out and does some of its work onthe side and then just presents
you those results along the waytoo.
So while you say found thosetop 50 profiles, say the agent
found an additional 10 or 20that you might also want to
consider.
Pete Newsome (20:00):
Now, is that a longer process?
Is that why it's an agentversus the automatic search, or
what's it differentiate that for?
David Paffenholz (20:09):
Yep, yeah, so it runs asynchronously.
Usually it'll get you the firstresults after five minutes or
so, but then every day it'llpresent you with some additional
results, and so it'll send youan email saying hey, you know
the search we started yesterday.
Here's another batch of 30profiles that either the agent
has already reached out to or ispending a review to reach out.
Pete Newsome (20:28):
Yeah, and I've noticed that that used to be a
very prevalent tool inCareerBuilder and.
Monster back in their heydays.
And those companies I thinkthey've now declared bankruptcy
and have been bought by someoneelse, which is wild.
Right, because at one pointeach of those were the 800-pound
gorilla in the space.
I don't see that feature reallyexisting anymore, what I would
(20:53):
call an alert.
Right, I want to see everyresume that shows up or someone
posts that's new with thisspecific skill set.
Right, I define my criteria,whatever it is that's really
powerful.
I mean, is there anyone elsedoing that today that you know
of?
David Paffenholz (21:10):
Not that I directly know of.
I think the underlying reasonit's now possible is because the
way the instant search works isit basically matches or looks
for the filtering, so thisperson matches the criteria or
it clearly has what we'relooking for.
But then what the agent does isit basically matches or looks
for the filtering.
So this person matches thecriteria or it clearly has what
we're looking for, but then whatthe agent does is it kind of
goes out and it'll review, liketens of thousands, hundreds of
thousands of additional profilesto look for potential signals
(21:33):
or reasons they could be a goodfit, and so that process.
One, it takes some time and soit has to run.
It's like it would be a baduser experience to have to wait
for it in real time to happen.
But then, two, it reliesheavily on large language models
similar to ChatGPT, to do thatprocess and to take that time,
and so it's really a kind of anew capability that we've had to
(21:54):
do that in a really good wayand present those profiles.
And my prediction is thatthat's going to become more and
more common again, where moreplatforms are going to offer
that type of automation andreminder.
Pete Newsome (22:07):
It's a really powerful tool to have and one
that I miss.
I wish my team had access to ittoday, and do you envision that
that's something you just keeprunning indefinitely, if it's a
position where you kind of wantevergreen candidates to continue
to be accessed?
David Paffenholz (22:25):
That's right.
Yeah, you can continue runningthe agent indefinitely if you
need the evergreen candidates.
And then the second part andthis isn't live yet, but coming
pretty soon is, as soon as wedetect a new role or job
description in your ATS, we'reready to go and launch an agent
for you.
And so as soon as that happens,it's already out there working
and trying to find some initialcandidates so that when you then
(22:46):
start working on the rolethere's already a baseline
that's happened for you and youcan kind of hit the ground
running rather than having tostart from scratch.
And so you know we're trying tobe as proactive as possible
with the agents to hopefullyhave that kind of user delight
from the first minute thatthey're working on the role.
Pete Newsome (23:02):
That's awesome.
I love it.
Where do you see the biggesttime savings as a whole so far
that recruiters are gaining fromusing Juicebox?
David Paffenholz (23:09):
I think the profile review is probably the
biggest one.
We've set up the search.
Now we have the list ofprofiles, but we still have to
click through every single oneof them.
That part is where we've seenthe most time used to be spent,
and perhaps also the most timethat the AI is able to add a lot
of value in, because it can doa very similar process at a lot
(23:31):
larger scale, and so there, Ithink, is where we're able to
save the most hours perrecruiter to this, but could you
define why Juicebox is betterat ranking those candidates?
Pete Newsome (23:51):
The ranking systems existed in ATS for years
, so is there something thatJuicebox is able to do to take
that to a greater level?
David Paffenholz (23:57):
Yeah, so basically all platforms that
have existed before 2022 usedsome form of machine learning or
keyword predictions,essentially to rank those
profiles.
So how many keywords match whatwe're looking for, or how
likely is that keyword to matchwhat we're looking for?
And then in some cases theywould go a bit beyond that, but
it would still be keyword basedand maybe including, like
(24:18):
similar keywords or things likethat.
That, for a long time, was likethe best in class approach, but
it's also inherently limitedbecause you can't go further
than that or like there's apretty clearly defined ranking
that you can do and there's nota ton of creativity that can go
into it.
What changed in 2022 whenChatGPT launched is that we
(24:38):
could do kind of fullylanguage-based inferences and so
we can look at the full profile, actually read the full context
of these are the roles theperson has worked in.
This is what the company doesthat the person worked at, and
then make a prediction of do wethink they might have the skill
that we're looking for?
Do we think they might haveexperience in this role that
we're looking for and that mightmean that there's zero
(25:00):
overlapping keywords or evenzero similar keywords, and so
they would never appear in thattraditional ranking approach
where it's like more similaritysearch based, but they would
appear on a ranking powered byjuice box or other large
language model powered platforms.
Because it's reading the fullcontext, is truly trying to
understand the profile and therequirement and then make that
judgment.
Could this person be a good fit?
Pete Newsome (25:21):
And no one trusted those rankings previously
anyway.
David Paffenholz (25:24):
Yeah, they weren't good and they weren't
explainable either.
Right, it was like you know.
Here's the word and that's whywe think it's there, whereas now
we can actually get a smallparagraph from the AI saying you
know, this profile spent fouryears working at this company,
which seems pretty similar tothe company you're recruiting
for.
Plus, they had a similar role.
Pete Newsome (25:46):
We think they're a great fit.
David Paffenholz (25:46):
That's awesome .
What?
What kind of feedback are yougetting from the recruiters
who've been using it for a while?
The we've we've been fortunateto see a lot of growth in our
customer base, so over 40percent of customers have
expanded their plans um toinclude additional team members
or or kind of grow their usageeven further, which has been
really exciting to see.
I think that's like the bestway of getting validation is
like do they they want to?
You know, not only continueworking with us, but grow their
(26:07):
use of Juicebox too.
There's also still areas wehave to work on.
You know, there's alwayslimitations to what AI can do
and how smart it gets and howfast it is, and so I think
there's also still a lot ofroadmap items that we're pretty
excited about for the future.
Pete Newsome (26:24):
I have no doubt about that Any, in particular
that you say, hey, this isreally going to be a game
changer when we get there, butthe technology is not quite
ready.
David Paffenholz (26:33):
One that we're kind of getting ready for is
compensation data, so showingyou person level predictions of
their current compensation andthen letting you filter based on
that as well, and so that youcan get very comprehensive but
also specific compensation dataon your talent pools and use
that as a criteria in yoursearches, even for candidates
(26:54):
that you've never interactedwith previously.
So that's been a really biglift, you know, one aggregating
that data and then two makingpredictions for the cases where
we don't have the data, and wethink it's going to be the most
accurate form that we've seen ofit so far, but I don't want to
make too many promises until itgoes live.
Pete Newsome (27:11):
No worries, I like it, we'll hold you to it, I'll
get you back here and we'll talkabout it.
Well, what about contact info?
I mean, that's always achallenge.
You have sites like Indeed thatwon't give it to you, and
LinkedIn doesn't have phonenumbers.
So is there a percentage youcould assign to how frequently
the candidate profile comes backwith contact info?
David Paffenholz (27:33):
Yep.
So we're pretty transparent onhow we do the contact info.
So we essentially work with asmany contact data providers as
possible.
We then compare them againsteach other for each profile, and
we'll show you the one that hasthe highest verification rate.
Because we do this at prettymassive scale, we're able to
negotiate good rates with thecontact data partners, and so we
can do so in a way that's stilleconomical, where it wouldn't
be for each firm to get allthose different contact data
(27:55):
providers, but it does makesense for us to do it on behalf
of our customers, and so thatincludes firms like ContactOut,
RocketReach and more.
In each of those cases, we'llpull the phone number and email,
compare them, verify them andthen show them to the user.
Pete Newsome (28:09):
So that's great.
So your customers get thebenefits of your scale you have
as a whole Makes perfect sense,because that's a constant
challenge and I would expectthat's a pretty common question
you get asked coming in byanyone who's considering
Juicebox.
David Paffenholz (28:23):
Yeah, and it's also like I think it's a nice,
it kind of almost makes it a nobrainer on the contact data side
because we can prettyconfidently say like oftentimes
our customers have one of thosecontact data providers or maybe
two, and you know we can showyou the list of contact data
providers that we work withwhere we can be pretty confident
that we'll get you bettercoverage and also that data
(28:43):
continues to improve.
So we learn from you, know anyemails that may have bounced or
when we know that a kind ofcontact data has changed and we
try to dynamically update thatdata set as well that's great.
Pete Newsome (28:53):
Um yeah, since you mentioned cost, I have to ask
about that too.
Are there?
You know, we?
We pay, we as a, as an industry, right staffing companies like
mine pay way too much money, Ithink, to Indeed and LinkedIn,
as we mentioned.
It used to be CareerBuilder andMonster that continues to
evolve.
Job boards just eat into ourprofits in a major way.
(29:16):
Do you see that as a potentialsavings?
Do you think we still need torely on those, even though
Juicebox has efficiencies wegain?
Can we gain some cost savingstoo?
David Paffenholz (29:27):
Yeah.
So, in short, yes, most of ourcustomers find cost savings by
using Juicebox.
That being said, I thinkoftentimes the cost savings even
get outweighed by the timesavings and the additional
productivity that comes fromusing Juicebox, where we're just
closing more roles, placingmore candidates, and the
business has a better outcomebecause it's able to operate
(29:48):
more efficiently.
And so, while we often doprovide cost savings compared to
existing solutions, thebenchmark we try to hold
ourselves to is are we makingthe business operate more
effectively because of oursolution, and do they also see
the ROI kind of on that purebasis too?
Pete Newsome (30:04):
That makes sense.
I'm not surprised that youanswered that way.
By the way, I would have lovedit if you said, yes, you'll get
be able to get rid of Indeedcompletely one day.
One day.
I look forward to that coming.
But no, that makes sense.
And that efficiency iseverything right, Because it's
not just in terms of internalresources and making the
recruiter's lives easier.
(30:25):
It's a competitive advantage ifyou can respond faster.
But we know that, and that'salways a balance With a company
like mine.
We're high touch, we are verythorough.
That's one of the things we'reknown for.
But my recruiters areconstantly looking at Canada
when it's a race, when it'scompetitive, like hey, we've got
to do all these things whenother companies are cutting
(30:47):
corners.
So I think you give the goodguys an advantage by creating
those efficiencies that the badcompanies benefit from otherwise
.
David Paffenholz (30:57):
Yeah, it's interesting To me that's one of
the most fascinating thingsabout the recruiting industry is
that it's one of the very fewindustries that is truly zero
sum, where a candidate can onlygo to one company at a time and
only one firm is going to beplacing that candidate, and
similarly for the open positionsthat a company has, and I think
that results in these kind ofvery interesting competitive
(31:18):
dynamics, which then in turnlead to some companies choosing
a route of.
You know, we're going to buildour reputation, we're going to
become long-term partners, whichoften is the most sustainable
route, but then there's alsoalways players that take a
different route, that maybeplayed a bit fast and loose and
get the temporary growth fromthat as well, and so I think
that'll continue to be the casetoo.
Pete Newsome (31:36):
And sometimes it's not even temporary, right?
I mean, the world of MSPs havemade, I would say put less value
on quality and more on justvolume and quantity, and no
knock against companies whooperate that way.
Good for them, right?
(31:57):
That industry exists, we knowit.
But if you are on the otherside of that equation, you can't
compete in that space, and it'sone of the things I've looked
for AI to potentially do for usand I've been excited about is
how can we bridge that gap,right?
How could we not be sodisadvantaged by doing what I
consider to be the right things,which everyone ultimately wants
(32:19):
, right?
But I think AI is going toclose that gap a lot.
So everything you're saying isa big piece of that.
David Paffenholz (32:29):
Yep, yeah, and hopefully I like.
The best part of the AIsolutions is that the teams who
adopt them kind of seethemselves winning more too,
because they're able to operatebetter and faster and keeping
that bar of consistency too.
So, yeah, we'll see how itplays out.
Pete Newsome (32:49):
Now, before I let you go cause I'm watching the
clock here I told you in advancebecause I'm watching the clock
here, I told you in advance howlong we'd talk, but that, as far
as the candidate experience youmentioned, that's how you
initially came to be here.
Is there anything that you'vebeen able to point to to say wow
, it really is.
I mean, it clearly benefit torecruiting firms and individual
recruiters, but how about forthe candidates themselves?
David Paffenholz (33:10):
Yeah, I think the underlying thing that
interests me the most is like dosoftware solutions like ours
actually help find candidates ormatch candidates that we're
pretty certain wouldn't havebeen found otherwise?
And I think that's the mostinteresting part, even if the
candidate isn't necessarilyaware of that in that
circumstance.
Because if we think of, like,the broader labor market even
going back to what you mentionedin the beginning, of, like, um,
(33:31):
the, the bro of laborstatistics is are we helping
make that entire job matchingfunction actually be more
efficient?
Are we creating moreopportunities and more matches
than than was possible otherwise?
And so I think an individualcandidate probably doesn't know
that because they have no way ofknowing, um, if they would have
been reached out to otherwise.
But on aggregate, there'll bemore placements being made,
(33:55):
there'll be more people beinghired, which I think is the
coolest kind of KPI behind allof this.
Now, at the same time, there'salso kind of more micro,
individual things that I canmake for a better candidate
experience.
That can be as simple asreminding a recruiter to respond
to a candidate if we know thatsomething has gone unanswered
for, say, 24 hours, to justensure that those best practices
(34:19):
are easier kept because thesoftware tries to be a little
bit proactive about that too,and so we try to think about
those little things to kind ofencourage a good experience too.
Pete Newsome (34:28):
Yeah, and those little things add up to, um, you
know, making or breaking theexperience as a whole.
No, no doubt.
And you know I have, um, youknow, four kids.
My youngest is, uh, is still inhigh school, but I have three
others that are in the world ofhaving to be involved in
professional jobs and, and, andI see, through their friends and
them, how difficult it is as acandidate.
(34:51):
In addition, you know, causeI'm I'm kind of in the middle of
being in staffing, but now thatI'm seeing it through
candidates eyes so that are sopersonal to me, um, I, I
realized how awful it really isand I know what I'm doing.
Right, I can give them goodguidance, and it is a bad
situation.
So anything you can do to, um,to enhance that overall
(35:15):
experience, man, it makes a bigdifference to the individuals
out there.
David Paffenholz (35:18):
Yeah, yeah, I agree, and it's also like it
feels like it's clear to me thatin like 10 years things will
continue to get even better, butthen it's also like it has to
happen in the in the meantime.
Pete Newsome (35:29):
Absolutely so.
Anything else about juice box Ididn't get to ask that you'd
want to highlight because I knowif what you've said and talked
about before we startedrecording so many of my peers
and staffing are looking for.
You know better ways to dothings and you're certainly
offering that.
I want to make sure we didn'tmiss anything obvious.
David Paffenholz (35:46):
No, I think I think we covered most of it.
The only thing that I'demphasize or encourage is kind
of what I tried to mention onlike the you know, how do people
discover AI software piece ofallowing employees or team
members to bring up things andtest out new things, and so I
think, even when evaluating asolution like ours, be that
(36:09):
Juicebox itself or othercompanies in the space that
would be my main word ofencouragement is letting the
people doing the work test itout and see what they think, and
trialing that, because a lot ofthat software is meant to be
trialed, meant to be used day today, rather than just being,
say, purchased top down and thenimplemented that way, and so
(36:29):
that's the one thing I alwayslike to encourage, because I
think it ends up leading tobetter decisions too.
Pete Newsome (36:34):
Perfect, and to that point, david, if someone
wants to try Juicebox, what dothey need to do next?
David Paffenholz (36:39):
Just head to juiceboxai, and then you can
either try it directly for free,or you can book a demo with our
team, and we'd be happy to walkyou through it too.
Pete Newsome (36:48):
Perfect, awesome.
All right, man.
Well, thank you so much.
I look forward to followingeverything Juicebox is doing.
I look forward to trying it outas well, which we're going to
do.
I'm convinced that this issomething my team absolutely
needs to look at, and I'll getyou on in another year and we'll
see what's changed since then.
Is that fair?
David Paffenholz (37:08):
Sounds like a plan.
Thanks so much for having me.
This was fun.
Pete Newsome (37:11):
Awesome.
Thank you so much, David.
Have a good rest of the day.
David Paffenholz (37:13):
You too Bye.
You're listening to the Hire Calling Podcast,
your source for all thingshiring, staffing and recruiting.
My guest today is DavidPaffenholz from Juicebox.
David, how are you today?
I'm doing well.
Thanks for having me.
It is great to have you.
I've been trying for a while.
You are a busy guy to get ahold of.
That must mean business isgoing pretty well.
David Paffenholz (00:18):
It's been an exciting time at the company,
lots of new things happening.
Pete Newsome (00:32):
Team is growing, customer base is growing and
super excited to be on todayAwesome Well.
David Paffenholz (00:33):
I'm happy to have you just quickly set the
stage and describe Juicebox in anutshell, if you could.
For sure, juicebox is an AIrecruiting platform.
We help our customers find thebest talent by aggregating a
bunch of different data sourcesand then using natural language
search to surface the besttalent.
We then have some additionaloutreach workflows, including
emailing, managing thosecontacts and more, and so I
think of us as an all-in-one,top-of-funnel AI hiring solution
(00:54):
.
Pete Newsome (00:55):
Excellent.
So, needless to say, ai alittle bit of a hot topic these
days, but we'll get into that ina minute too.
First, tell me about yourbackground a little bit.
How did you get to where youare now?
David Paffenholz (01:06):
For sure.
So, as one might be able totell by the last name, I'm
originally from Germany.
I came to the US first forcollege and then stuck around
since I started the business,together with my co-founder, in
2022.
And so this was shortly beforeChatGPT came out and everyone
started talking about AI, andour kind of initial vision, or
(01:26):
what we wanted to focus on, isbuilding tooling to make the
recruiting process easier,particularly on the search side,
and so that's where we started.
The reason we were passionateor interested in that space is
because both my co-founder and Ihad always found employment
through kind of irregularmethods.
We were basically cold,reaching out to jobs or roles we
thought were interesting ortrying to create those roles for
(01:47):
ourselves, and so it made usbelieve that there's still a lot
of friction in that hiringprocess and still a lot that can
be solved, which also made usinterested in building the
product.
Pete Newsome (01:57):
I love that, and you know there's I think there's
not many people who are youngand coming out of school finding
the job search process easythese days.
So not a unique situation, butyou chose to tackle it in a
unique way, which I love Notjust complaining about it, but
doing something about it.
So good for you for reachingthat conclusion.
(02:19):
Let's talk about the state ofthe market today a little bit.
I don't know if you saw thatthe Bureau of Labor Statistics
just happened to put out theiremployment report this morning.
Did you see the numbers?
David Paffenholz (02:31):
Yeah, so there was like a large correction
shift.
Pete Newsome (02:34):
Yeah, a massive one, right that almost 300,000
jobs were overreported over thepast two months, and so a lot of
us in this industry and a lotof job seekers are feeling that
the market is not easy.
Right, the government for thelast couple of months, has told
us that it's easier than we'refeeling, and this told a pretty
(02:54):
big story.
I think that the job market isnot growing as well as we'd like
it to, and it's really hard tofind a job.
It's really competitive,recruiting is tough.
So that's my perspective on themarket right now.
What are you seeing?
David Paffenholz (03:08):
Yeah, I think there's kind of very different
scenarios in differentsubsectors of the market, and so
in some spaces we're seeing themarket really heat up and kind
of becoming very competitive,and other sectors we're seeing
kind of a lot less of that.
And I think the labor marketcorrection is a good reflection
of that too, where thestatistics sometimes say one
(03:31):
thing, the reality on the groundsays another thing, and then
the reality on the ground alsochanges depending on the role,
the location and even thatspecific point in time.
And so we've seen quite somedrastic shifts amongst where our
customers are hiring, wherethey're focusing on their
searches, even what industriesthey're focusing on In that case
, of course from the recruitingagency perspective as well.
Pete Newsome (03:53):
You know, and I always feel like and I'm biased,
of course, because I own astaffing agency but I feel like
we have our finger on the pulseas an industry of what is
happening at any given time andthat's why, for me personally,
some of the numbers weresurprising.
Right, you look on LinkedIn, itdoesn't sound like it's a great
job market right now and when Italk to my peers around the
country, we've felt I've pickedup on some struggles that a lot
(04:16):
of companies are having thesedays.
But I wanna know if you couldshare which sectors do you see
they're doing well?
David Paffenholz (04:25):
Yeah, I'd say the one where we've seen the
strongest pull in is like thetech sector, especially here in
San Francisco.
We've seen like a lot of surgein hiring there, really across
roles, but particularly forin-person roles here in the
broader Bay Area, and so thatpart of the market seems to have
really been accelerating andthat's like roughly 20 market
(04:45):
seems to have really beenaccelerating, whereas and that's
like roughly 20 to 30% of ourcustomers so it's a sizable
chunk but it's not the majorityof our customer base and so
we've seen that accelerationthere, and then in other parts
of the market perhaps less soand perhaps less of that pull.
I think it's interesting too,because it's almost like an
inverse, where the tech sectoralmost has this opposite
(05:06):
relationship to the broaderhiring market and in times where
the tech sector is doing well,the broader economy might not be
, and vice versa, where we sawthe opposite two years ago.
And so for us as well, where weserve both customer groups, we
always see some discrepancy inthose as well.
Pete Newsome (05:24):
What are you seeing locally out there with
return to office?
It's kind of subsided.
I haven't heard as much as wehave for the past couple of
years here in Florida, where Iam.
I think it's business as usual.
For the most part, those whoare going back have already gone
back and there's a lot ofhybrid roles out there.
But what are you seeing?
David Paffenholz (05:42):
Yeah, I think most companies and most
companies have kind of madetheir decision, like, either
they are going to be, you know,fully in person they're hybrid
or they're fully remote, and atleast at this point they really
should have made that decisiontoo, because, you know, it's
been a while since the world hasstarted to normalize.
In our case, our team is inperson.
(06:03):
We see a lot of companies thatare kind of fully remote or have
chosen that hybrid path as well.
I will say that, from whatwe've seen, the in-person market
has just become a lot harder torecruit for because it is a
constraint on how manycandidates can take on that role
.
Pete Newsome (06:21):
Yeah, and I think so much of that comes to supply
and demand right Candidates whohave a rare skillset, which has
always been the case, but nowit's in many ways targeted at.
Are you going to work in theoffice or do you have the luxury
of having that skillset thatallow you to call your own shots
more and work at home?
Are you seeing that?
Because I really think supplyand demand is driving a lot of
(06:43):
it.
David Paffenholz (06:43):
Yes, yeah, definitely.
I also think there's like someinteresting parts where, like
roles that used to we used tothink are like in-person
mandatory because of COVID,we're like forced to shift to
remote, and then they've kind ofjust there's been this
consensus of like, you know,some of these roles can
completely be done remote, whereif we had just asked people
five, six years ago, is thateven possible for these roles, a
(07:05):
lot of people would have saidno, and so I think there's kind
of been that rethinking in themarket which has really changed
what people think about in termsof demand as well, of like what
roles are really needed to bein person versus not.
Pete Newsome (07:17):
So that's a perfect segue into what we're
really talking about in the bigshift is AI.
We're really talking about inthe big shift is AI.
What wasn't possible in thevery recent past is now a
reality that we're having tofigure out how to deal with
right, not just in the staffingindustry, but as a whole.
So what's your general take onAI and how it's impacting the
(07:41):
job market already, and then howit's going to impact it in the
future?
David Paffenholz (07:45):
Yeah, I think AI has had two big impacts on
the industry.
I'd say like one maybe on thelabor market more generally and
then the second specifically onlike staffing and recruiting.
And so I think in the labormarket kind of more broadly, it
means that the output of everysingle role I think the
expectations have increased andthe ceiling of what is possible
(08:09):
has increased because people arestarting to use AI tools,
starting to be more efficient inparts of their workflow and
achieving more.
But that then also means thatin some cases, the expectations
for those roles have gone upbecause the kind of normal is
now at a different rate.
I think overall that's a greatthing because productivity
increases, the economy growsfaster and we're able to do much
(08:29):
more.
At the same time, it also makesit particularly tricky for roles
where maybe that AI adoption isa bit slower or employees don't
provide the opportunities toadopt AI or the tooling for it.
I do think in most sectorsthere is now at least one AI
tool that can really simplifyworkflow, if not multiple, I
think, in recruiting.
(08:50):
You know, despite us being inthe AI recruiting space, there's
also a lot of other goodcompanies in the AI recruiting
space and I'd say there's likethere's a good amount of
different solutions to evaluateand to test from, and so I think
, yeah, there's there's a lot,right to say the least, and it's
evolving rapidly.
Pete Newsome (09:07):
That's one of the things that I struggle with.
One of the reasons I wanted tospeak with you today, quite
frankly, is to share a messageof what's even out there,
Because when I speak to my peersaround the country, they
struggle to keep up.
But I'm curious about adoption,how your perspective may differ
than mine.
Have you seen any trends amongcompany size, you know being
(09:31):
faster or slower to adopt, or isthere generational differences
that you see on?
You know, if you had to put anycategories in place, can you do
that?
Or is it just really all overthe place?
David Paffenholz (09:44):
Yeah, so what we've seen that gets me really
excited and that I think is kindof unique to this time as well,
is that individual users inmany cases, or like individual
team members, can find AIsolutions and start adopting
them for their own workflows.
If they find that it works forthem, it can kind of spread
within an organization withoutneeding to be like centrally
(10:05):
purchased solution, and so Ithink a lot of AI tools have
started positioning themselvesthat way too.
It's on like the simplest level, say a chat GPT where everyone
can sign up, create their ownaccount and start using it in
different ways, but then evenworkflow specific tools where an
individual user can sign up,try using it for free and then
might spread adoption acrosstheir team.
For us, almost all of ourcustomers start off by having
(10:28):
one or two users internallyhappen to come across the
product or maybe have heardabout the product, test it out
and see if it works for theirworkflow before showing it to
the wider team or looping intheir manager to do a kind of
more traditional evaluation, andthat kind of bottoms up
discovery I think is reallypowerful, because it also sets
the bar a bit higher.
For, you know, does this reallywork, or is this just a really
(10:49):
good sales pitch?
Pete Newsome (10:51):
I really like that and I hadn't thought of that
before what you just said before.
But as I'm listening to you, Iunderstand exactly what you mean
and I've seen that over andover and that is pretty unique
in terms of technology adoptionin my lifetime probably everyone
else's where you really do haveaccess at an individual level
(11:12):
without having the company haveto make the decision as a whole,
and what a great thing I mean.
As an employee.
Over the years I've had lots ofideas where, when I was an
employee which was early in mycareer I would often get
frustrated by an inability tohave the company adopt new
things right.
I've worked for two largeemployers prior to starting Four
(11:33):
Corner and they've moved veryslowly.
I don't think you can do thatanymore and I think companies
have to listen to theiremployees and would want to
listen to their employees whohave the willingness to go and
seek things out on their own.
David Paffenholz (11:45):
Yup, yeah.
And frankly, it's also a greatway where you know if companies
notice that, say, two employeesare really outperforming or
suddenly producing a lot moreoutcomes and output, and you
know.
Then one asks, hey, like whatare you up to?
Or what are you doing, hasanything changed?
And maybe they're using newtooling or they're trying new
workflows.
That can be a really coolevidence point as well, where it
(12:08):
almost becomes a no brainer forthe organization to at least
consider those resources orthink about what a wider
adoption could look like too.
Pete Newsome (12:15):
And yeah, so true, and I love it, and you know as
well as I do as a businessleader, there is nothing better
than an employee coming to youwith a solution to a problem you
didn't even ask him to solveright.
David Paffenholz (12:26):
I mean, how great is that it doesn't get me
better?
Yup yeah, Especially when it'slike a problem that you know,
it's always kind of being in theback of one's mind and like
never really had the chance totackle it or think about what
one can do about it, and thenjust seeing that solution in
front of you, it can be a prettynice feeling.
Pete Newsome (12:42):
A hundred percent, all right, man.
Well, let's get into juice boxspecifically now.
So I can tell you that theentirety of the time I've owned
a staffing company which iswe're in our 20th year we've
always tried to find a betterway to do things right, gain
efficiency, gain speed, gainthoroughness, and there's been
(13:03):
very little potential to do thatup until recently.
So I couldn't be more excitedabout solutions like yours, and
I know that as an industry, wedesperately need new solutions
to improve.
So just break it down from thebeginning what is the premise in
terms of what is the biggestproblem Juicebox solves?
(13:26):
And then let's get into somedetails For sure.
David Paffenholz (13:29):
So I'll start off by kind of describing what
we see as the status quo, orwhat our users typically do
before switching to Juicebox.
Then I'll contrast that withwhat Juicebox does.
And so currently, what we see alot of our customers do before
adopting Juicebox is when doinga candidate search.
They have a pretty manualprocess.
There's usually two places theygo.
One is their existing ATS ordatabase where they look through
(13:50):
any candidates that they mightalready know that could be a
good fit.
Usually involves perhapssetting a filter, running a
search depending on the softwarethat they use, and then kind of
manually reviewing thoseprofiles In some cases pretty
long lists, and ends up beingfairly time intensive, Also
usually not the best interfaceto do it with.
And then the second thing thatwe see our customers do is going
out and finding net newprofiles for those roles, and so
(14:12):
you know, be that sourcingexternally through a channel
like Indeed, LinkedIn or others,where we kind of manually also
enter those filters and thenstart reviewing, and then
finally we'll want to engagethose candidates, be that
through emailing, phone calls,etc.
That whole process ends upbeing a large share of recruiter
time spent.
(14:33):
That is exactly the workflowthat Juicebox aims to solve, and
so we do that in a fewdifferent ways.
The main one is we combine yourexisting data with our larger
data set, where we have over 800million profiles globally.
We've indexed them, added abunch of additional data into
them and provide one unifiedsearch.
And so, rather than having togo into those different systems,
(14:54):
describe who you might belooking for the role you're
hiring for, you can just typethat into Juicebox or put in the
job description.
We'll automatically go out andsearch for both the best
profiles who are already in yourdatabase as well as net new
profiles that you shouldprobably consider for the role.
From there, we'll then rankthem for you, and so we'll go
skill by skill.
Do they match what you'relooking for, criteria by
(15:15):
criteria, and present it to youvery similar to how a human
researcher or sourcer might.
They'll go profile by profileand say you know, this criteria
is met, but maybe not this oneor this profile could be a good
fit for this reason.
We'll then present those topmatches to you and let you reach
out to them in one click,either through an email sequence
or, if you prefer, to reach outby phone.
We'll provide that data too,and so that whole process,
(15:38):
instead of taking 10 plus hours,can go as fast as 15 minutes,
going from initial search tohaving that output and then
being able to take action fromit.
Pete Newsome (15:48):
And everything you described is for a human to do
is what I consider to below-quality time.
It is necessary, but we wantour recruiters engaged with
candidates directly right, wewant them connected and
everything that leads up to thatagain necessary.
But you'd like a better way todo it, a more efficient way.
So tell me about the 800.
(16:08):
I'm intrigued by that 800million.
Are those existing in yourdatabase, or do you have access
to them in a database likeLinkedIn?
I mean, where do thosecandidate profiles reside?
David Paffenholz (16:21):
Yep.
So that's aggregated frompublic profile data.
So, be that profiles thatcandidates may have on
professional networking sites,be that profiles they may have
on social media sites like aFacebook or Twitter, be that
other information they mighthave on a company website or on
a personal website, depending onwhat's available, it varies a
(16:41):
bit from from role to role aswell.
So, for example, for the healthcare sector, there's some other
data sources available, or evenpublic licensing information
that might not be available forother sectors, and so we kind of
see what's out there, dependingon the sector, aggregate that
data and then put that into amore what we call unified
profile, basically justcombining those different data
sources into one source of truth.
Pete Newsome (17:02):
Okay, now is a tool, so do you have better can
it profiles or more can itprofiles in?
Do you have a strength area inparticular, or are you pretty
agnostic when it comes to theposition types?
David Paffenholz (17:13):
We're pretty agnostic.
I'd say, geographically we'restrongest in North America and
it's where the majority of ourcustomers are, but then within
that we're very role agnostic.
Pete Newsome (17:25):
Okay, nice.
So give me customer sizetypically.
Do you have an ideal customersize?
Do you have an ideal number ofpositions that you really see
where juice box is mosteffective?
Paint that picture for me.
David Paffenholz (17:40):
Yeah, it's become quite broad.
So we now work with over 2,500customers, ranging from
individual recruiters who maybejust set up their own agency or
started freelance recruiting,all the way to some pretty large
enterprises where they have,say, a 100-person recruiting
team, or larger recruitingagencies where they have
similarly sized recruiting teams.
I'd say the sweet spot to getstarted is once there's at least
(18:04):
five recruiters on the team,there's proper processes in
place and the baseline ofprocesses already exist.
That baseline becomes even moreimportant when adding in the AI
layer on top, because ifthere's not a consistent
internal workflow, then it's tooearly to add the AI or
automation on top, because itmight just lead to more
confusion than benefit.
And so that's usually what Irecommend is that, like a
(18:26):
baseline to get started is youknow, are those workflows
defined, are the processesdefined, and then are we ready
to automate those?
And oftentimes we see thataround like the five recruiter
mark, though it can vary too.
Pete Newsome (18:37):
So more effective if there's a certain level of
maturity in the process andexperience.
More effective if there's acertain level of maturity in the
process and experience, whichmakes sense.
So I do my search right and Iget to return profiles.
Do I download them into my ATS?
For the ones I want to contact?
Do they stay within Juicebox?
How does that process work?
David Paffenholz (18:57):
Yeah, so that full process things with your
ATS.
So, as you've discovered thosetop profiles, let's say there's
maybe top 50 that were matchedfor you and 20 of those already
exist in your ATS.
So we'll flag that for you andthen the remaining 30 might be
net new, and so you can thentake the 30 net new ones, save
those into your ATS as well andthen reach out to all 50 of them
and make sure that that data islogged there too, and so all of
(19:20):
that data ends up living inyour ATS2.
So you still have that as yoursource of truth, while also
having it available in Juiceboxas you're doing new searches.
There is one additional elementto the platform which I didn't
mention earlier, which is theJuicebox agent.
The Juicebox agent kind ofworks on its own and it starts
doing its own searches and itsown outreach without necessarily
requiring you to review eachprofile.
(19:41):
And the reason I'm bringingthat up is because it can work
in parallel with a humanrecruiter, where the agent goes
out and does some of its work onthe side and then just presents
you those results along the waytoo.
So while you say found thosetop 50 profiles, say the agent
found an additional 10 or 20that you might also want to
consider.
Pete Newsome (20:00):
Now, is that a longer process?
Is that why it's an agentversus the automatic search, or
what's it differentiate that for?
David Paffenholz (20:09):
Yep, yeah, so it runs asynchronously.
Usually it'll get you the firstresults after five minutes or
so, but then every day it'llpresent you with some additional
results, and so it'll send youan email saying hey, you know
the search we started yesterday.
Here's another batch of 30profiles that either the agent
has already reached out to or ispending a review to reach out.
Pete Newsome (20:28):
Yeah, and I've noticed that that used to be a
very prevalent tool inCareerBuilder and.
Monster back in their heydays.
And those companies I thinkthey've now declared bankruptcy
and have been bought by someoneelse, which is wild.
Right, because at one pointeach of those were the 800-pound
gorilla in the space.
I don't see that feature reallyexisting anymore, what I would
(20:53):
call an alert.
Right, I want to see everyresume that shows up or someone
posts that's new with thisspecific skill set.
Right, I define my criteria,whatever it is that's really
powerful.
I mean, is there anyone elsedoing that today that you know
of?
David Paffenholz (21:10):
Not that I directly know of.
I think the underlying reasonit's now possible is because the
way the instant search works isit basically matches or looks
for the filtering, so thisperson matches the criteria or
it clearly has what we'relooking for.
But then what the agent does isit basically matches or looks
for the filtering.
So this person matches thecriteria or it clearly has what
we're looking for, but then whatthe agent does is it kind of
goes out and it'll review, liketens of thousands, hundreds of
thousands of additional profilesto look for potential signals
(21:33):
or reasons they could be a goodfit, and so that process.
One, it takes some time and soit has to run.
It's like it would be a baduser experience to have to wait
for it in real time to happen.
But then, two, it reliesheavily on large language models
similar to ChatGPT, to do thatprocess and to take that time,
and so it's really a kind of anew capability that we've had to
(21:54):
do that in a really good wayand present those profiles.
And my prediction is thatthat's going to become more and
more common again, where moreplatforms are going to offer
that type of automation andreminder.
Pete Newsome (22:07):
It's a really powerful tool to have and one
that I miss.
I wish my team had access to ittoday, and do you envision that
that's something you just keeprunning indefinitely, if it's a
position where you kind of wantevergreen candidates to continue
to be accessed?
David Paffenholz (22:25):
That's right.
Yeah, you can continue runningthe agent indefinitely if you
need the evergreen candidates.
And then the second part andthis isn't live yet, but coming
pretty soon is, as soon as wedetect a new role or job
description in your ATS, we'reready to go and launch an agent
for you.
And so as soon as that happens,it's already out there working
and trying to find some initialcandidates so that when you then
(22:46):
start working on the rolethere's already a baseline
that's happened for you and youcan kind of hit the ground
running rather than having tostart from scratch.
And so you know we're trying tobe as proactive as possible
with the agents to hopefullyhave that kind of user delight
from the first minute thatthey're working on the role.
Pete Newsome (23:02):
That's awesome.
I love it.
Where do you see the biggesttime savings as a whole so far
that recruiters are gaining fromusing Juicebox?
David Paffenholz (23:09):
I think the profile review is probably the
biggest one.
We've set up the search.
Now we have the list ofprofiles, but we still have to
click through every single oneof them.
That part is where we've seenthe most time used to be spent,
and perhaps also the most timethat the AI is able to add a lot
of value in, because it can doa very similar process at a lot
(23:31):
larger scale, and so there, Ithink, is where we're able to
save the most hours perrecruiter to this, but could you
define why Juicebox is betterat ranking those candidates?
Pete Newsome (23:51):
The ranking systems existed in ATS for years
, so is there something thatJuicebox is able to do to take
that to a greater level?
David Paffenholz (23:57):
Yeah, so basically all platforms that
have existed before 2022 usedsome form of machine learning or
keyword predictions,essentially to rank those
profiles.
So how many keywords match whatwe're looking for, or how
likely is that keyword to matchwhat we're looking for?
And then in some cases theywould go a bit beyond that, but
it would still be keyword basedand maybe including, like
(24:18):
similar keywords or things likethat.
That, for a long time, was likethe best in class approach, but
it's also inherently limitedbecause you can't go further
than that or like there's apretty clearly defined ranking
that you can do and there's nota ton of creativity that can go
into it.
What changed in 2022 whenChatGPT launched is that we
(24:38):
could do kind of fullylanguage-based inferences and so
we can look at the full profile, actually read the full context
of these are the roles theperson has worked in.
This is what the company doesthat the person worked at, and
then make a prediction of do wethink they might have the skill
that we're looking for?
Do we think they might haveexperience in this role that
we're looking for and that mightmean that there's zero
(25:00):
overlapping keywords or evenzero similar keywords, and so
they would never appear in thattraditional ranking approach
where it's like more similaritysearch based, but they would
appear on a ranking powered byjuice box or other large
language model powered platforms.
Because it's reading the fullcontext, is truly trying to
understand the profile and therequirement and then make that
judgment.
Could this person be a good fit?
Pete Newsome (25:21):
And no one trusted those rankings previously
anyway.
David Paffenholz (25:24):
Yeah, they weren't good and they weren't
explainable either.
Right, it was like you know.
Here's the word and that's whywe think it's there, whereas now
we can actually get a smallparagraph from the AI saying you
know, this profile spent fouryears working at this company,
which seems pretty similar tothe company you're recruiting
for.
Plus, they had a similar role.
Pete Newsome (25:46):
We think they're a great fit.
David Paffenholz (25:46):
That's awesome .
What?
What kind of feedback are yougetting from the recruiters
who've been using it for a while?
The we've we've been fortunateto see a lot of growth in our
customer base, so over 40percent of customers have
expanded their plans um toinclude additional team members
or or kind of grow their usageeven further, which has been
really exciting to see.
I think that's like the bestway of getting validation is
like do they they want to?
You know, not only continueworking with us, but grow their
(26:07):
use of Juicebox too.
There's also still areas wehave to work on.
You know, there's alwayslimitations to what AI can do
and how smart it gets and howfast it is, and so I think
there's also still a lot ofroadmap items that we're pretty
excited about for the future.
Pete Newsome (26:24):
I have no doubt about that Any, in particular
that you say, hey, this isreally going to be a game
changer when we get there, butthe technology is not quite
ready.
David Paffenholz (26:33):
One that we're kind of getting ready for is
compensation data, so showingyou person level predictions of
their current compensation andthen letting you filter based on
that as well, and so that youcan get very comprehensive but
also specific compensation dataon your talent pools and use
that as a criteria in yoursearches, even for candidates
(26:54):
that you've never interactedwith previously.
So that's been a really biglift, you know, one aggregating
that data and then two makingpredictions for the cases where
we don't have the data, and wethink it's going to be the most
accurate form that we've seen ofit so far, but I don't want to
make too many promises until itgoes live.
Pete Newsome (27:11):
No worries, I like it, we'll hold you to it, I'll
get you back here and we'll talkabout it.
Well, what about contact info?
I mean, that's always achallenge.
You have sites like Indeed thatwon't give it to you, and
LinkedIn doesn't have phonenumbers.
So is there a percentage youcould assign to how frequently
the candidate profile comes backwith contact info?
David Paffenholz (27:33):
Yep.
So we're pretty transparent onhow we do the contact info.
So we essentially work with asmany contact data providers as
possible.
We then compare them againsteach other for each profile, and
we'll show you the one that hasthe highest verification rate.
Because we do this at prettymassive scale, we're able to
negotiate good rates with thecontact data partners, and so we
can do so in a way that's stilleconomical, where it wouldn't
be for each firm to get allthose different contact data
(27:55):
providers, but it does makesense for us to do it on behalf
of our customers, and so thatincludes firms like ContactOut,
RocketReach and more.
In each of those cases, we'llpull the phone number and email,
compare them, verify them andthen show them to the user.
Pete Newsome (28:09):
So that's great.
So your customers get thebenefits of your scale you have
as a whole Makes perfect sense,because that's a constant
challenge and I would expectthat's a pretty common question
you get asked coming in byanyone who's considering
Juicebox.
David Paffenholz (28:23):
Yeah, and it's also like I think it's a nice,
it kind of almost makes it a nobrainer on the contact data side
because we can prettyconfidently say like oftentimes
our customers have one of thosecontact data providers or maybe
two, and you know we can showyou the list of contact data
providers that we work withwhere we can be pretty confident
that we'll get you bettercoverage and also that data
(28:43):
continues to improve.
So we learn from you, know anyemails that may have bounced or
when we know that a kind ofcontact data has changed and we
try to dynamically update thatdata set as well that's great.
Pete Newsome (28:53):
Um yeah, since you mentioned cost, I have to ask
about that too.
Are there?
You know, we?
We pay, we as a, as an industry, right staffing companies like
mine pay way too much money, Ithink, to Indeed and LinkedIn,
as we mentioned.
It used to be CareerBuilder andMonster that continues to
evolve.
Job boards just eat into ourprofits in a major way.
(29:16):
Do you see that as a potentialsavings?
Do you think we still need torely on those, even though
Juicebox has efficiencies wegain?
Can we gain some cost savingstoo?
David Paffenholz (29:27):
Yeah.
So, in short, yes, most of ourcustomers find cost savings by
using Juicebox.
That being said, I thinkoftentimes the cost savings even
get outweighed by the timesavings and the additional
productivity that comes fromusing Juicebox, where we're just
closing more roles, placingmore candidates, and the
business has a better outcomebecause it's able to operate
(29:48):
more efficiently.
And so, while we often doprovide cost savings compared to
existing solutions, thebenchmark we try to hold
ourselves to is are we makingthe business operate more
effectively because of oursolution, and do they also see
the ROI kind of on that purebasis too?
Pete Newsome (30:04):
That makes sense.
I'm not surprised that youanswered that way.
By the way, I would have lovedit if you said, yes, you'll get
be able to get rid of Indeedcompletely one day.
One day.
I look forward to that coming.
But no, that makes sense.
And that efficiency iseverything right, Because it's
not just in terms of internalresources and making the
recruiter's lives easier.
(30:25):
It's a competitive advantage ifyou can respond faster.
But we know that, and that'salways a balance With a company
like mine.
We're high touch, we are verythorough.
That's one of the things we'reknown for.
But my recruiters areconstantly looking at Canada
when it's a race, when it'scompetitive, like hey, we've got
to do all these things whenother companies are cutting
(30:47):
corners.
So I think you give the goodguys an advantage by creating
those efficiencies that the badcompanies benefit from otherwise
.
David Paffenholz (30:57):
Yeah, it's interesting To me that's one of
the most fascinating thingsabout the recruiting industry is
that it's one of the very fewindustries that is truly zero
sum, where a candidate can onlygo to one company at a time and
only one firm is going to beplacing that candidate, and
similarly for the open positionsthat a company has, and I think
that results in these kind ofvery interesting competitive
(31:18):
dynamics, which then in turnlead to some companies choosing
a route of.
You know, we're going to buildour reputation, we're going to
become long-term partners, whichoften is the most sustainable
route, but then there's alsoalways players that take a
different route, that maybeplayed a bit fast and loose and
get the temporary growth fromthat as well, and so I think
that'll continue to be the casetoo.
Pete Newsome (31:36):
And sometimes it's not even temporary, right?
I mean, the world of MSPs havemade, I would say put less value
on quality and more on justvolume and quantity, and no
knock against companies whooperate that way.
Good for them, right?
(31:57):
That industry exists, we knowit.
But if you are on the otherside of that equation, you can't
compete in that space, and it'sone of the things I've looked
for AI to potentially do for usand I've been excited about is
how can we bridge that gap,right?
How could we not be sodisadvantaged by doing what I
consider to be the right things,which everyone ultimately wants
(32:19):
, right?
But I think AI is going toclose that gap a lot.
So everything you're saying isa big piece of that.
David Paffenholz (32:29):
Yep, yeah, and hopefully I like.
The best part of the AIsolutions is that the teams who
adopt them kind of seethemselves winning more too,
because they're able to operatebetter and faster and keeping
that bar of consistency too.
So, yeah, we'll see how itplays out.
Pete Newsome (32:49):
Now, before I let you go cause I'm watching the
clock here I told you in advancebecause I'm watching the clock
here, I told you in advance howlong we'd talk, but that, as far
as the candidate experience youmentioned, that's how you
initially came to be here.
Is there anything that you'vebeen able to point to to say wow
, it really is.
I mean, it clearly benefit torecruiting firms and individual
recruiters, but how about forthe candidates themselves?
David Paffenholz (33:10):
Yeah, I think the underlying thing that
interests me the most is like dosoftware solutions like ours
actually help find candidates ormatch candidates that we're
pretty certain wouldn't havebeen found otherwise?
And I think that's the mostinteresting part, even if the
candidate isn't necessarilyaware of that in that
circumstance.
Because if we think of, like,the broader labor market even
going back to what you mentionedin the beginning, of, like, um,
(33:31):
the, the bro of laborstatistics is are we helping
make that entire job matchingfunction actually be more
efficient?
Are we creating moreopportunities and more matches
than than was possible otherwise?
And so I think an individualcandidate probably doesn't know
that because they have no way ofknowing, um, if they would have
been reached out to otherwise.
But on aggregate, there'll bemore placements being made,
(33:55):
there'll be more people beinghired, which I think is the
coolest kind of KPI behind allof this.
Now, at the same time, there'salso kind of more micro,
individual things that I canmake for a better candidate
experience.
That can be as simple asreminding a recruiter to respond
to a candidate if we know thatsomething has gone unanswered
for, say, 24 hours, to justensure that those best practices
(34:19):
are easier kept because thesoftware tries to be a little
bit proactive about that too,and so we try to think about
those little things to kind ofencourage a good experience too.
Pete Newsome (34:28):
Yeah, and those little things add up to, um, you
know, making or breaking theexperience as a whole.
No, no doubt.
And you know I have, um, youknow, four kids.
My youngest is, uh, is still inhigh school, but I have three
others that are in the world ofhaving to be involved in
professional jobs and, and, andI see, through their friends and
them, how difficult it is as acandidate.
(34:51):
In addition, you know, causeI'm I'm kind of in the middle of
being in staffing, but now thatI'm seeing it through
candidates eyes so that are sopersonal to me, um, I, I
realized how awful it really isand I know what I'm doing.
Right, I can give them goodguidance, and it is a bad
situation.
So anything you can do to, um,to enhance that overall
(35:15):
experience, man, it makes a bigdifference to the individuals
out there.
David Paffenholz (35:18):
Yeah, yeah, I agree, and it's also like it
feels like it's clear to me thatin like 10 years things will
continue to get even better, butthen it's also like it has to
happen in the in the meantime.
Pete Newsome (35:29):
Absolutely so.
Anything else about juice box Ididn't get to ask that you'd
want to highlight because I knowif what you've said and talked
about before we startedrecording so many of my peers
and staffing are looking for.
You know better ways to dothings and you're certainly
offering that.
I want to make sure we didn'tmiss anything obvious.
David Paffenholz (35:46):
No, I think I think we covered most of it.
The only thing that I'demphasize or encourage is kind
of what I tried to mention onlike the you know, how do people
discover AI software piece ofallowing employees or team
members to bring up things andtest out new things, and so I
think, even when evaluating asolution like ours, be that
(36:09):
Juicebox itself or othercompanies in the space that
would be my main word ofencouragement is letting the
people doing the work test itout and see what they think, and
trialing that, because a lot ofthat software is meant to be
trialed, meant to be used day today, rather than just being,
say, purchased top down and thenimplemented that way, and so
(36:29):
that's the one thing I alwayslike to encourage, because I
think it ends up leading tobetter decisions too.
Pete Newsome (36:34):
Perfect, and to that point, david, if someone
wants to try Juicebox, what dothey need to do next?
David Paffenholz (36:39):
Just head to juiceboxai, and then you can
either try it directly for free,or you can book a demo with our
team, and we'd be happy to walkyou through it too.
Pete Newsome (36:48):
Perfect, awesome.
All right, man.
Well, thank you so much.
I look forward to followingeverything Juicebox is doing.
I look forward to trying it outas well, which we're going to
do.
I'm convinced that this issomething my team absolutely
needs to look at, and I'll getyou on in another year and we'll
see what's changed since then.
Is that fair?
David Paffenholz (37:08):
Sounds like a plan.
Thanks so much for having me.
This was fun.
Pete Newsome (37:11):
Awesome.
Thank you so much, David.
Have a good rest of the day.
David Paffenholz (37:13):
You too Bye.
Popular Podcasts
Fudd Around And Find Out
UConn basketball star Azzi Fudd brings her championship swag to iHeart Women’s Sports with Fudd Around and Find Out, a weekly podcast that takes fans along for the ride as Azzi spends her final year of college trying to reclaim the National Championship and prepare to be a first round WNBA draft pick. Ever wonder what it’s like to be a world-class athlete in the public spotlight while still managing schoolwork, friendships and family time? It’s time to Fudd Around and Find Out!
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.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.