#2: Using AI to Level Up Recruiting & Gain Insights into the Employee Experience

Welcome back to another episode of The Junction.

In today's episode, we're diving into the

realm of human resources and how artificial

intelligence is shaking things up.

So whether you are an HR pro wanting to stay ahead

or just curious like the rest of us about the future

of work, I think you're in for a treat. Chase.

Let's talk HR.

AI.

What are your initial thoughts?

I know you've got a thought for everything.

I think way too much about all this stuff. Actually.

I was up at 230 this morning because my daughter

came in and then I couldn't go back to bed.

So I was thinking about all these things

we were going to be talking about.

Yeah, but this is really one of the

most interesting use cases for us internally.

We have lots of people coming in the door, they

want to work here, we want to work with them.

But we've got a lot of processes, processes that go into

how this all works and if there is any way we

can speed it up, we are going to benefit. Right.

And we can get people that want to work here

and that we want to work with in the door

faster so we can do better and greater things.

This is a really interesting

area for automation and AI.

I remember in the beginning when we were two

dudes sitting in the office, we were like, well

we want it to be really personable.

I want to get on the phone and talk with these people.

I want to send them an email.

And when you start introducing automation, it starts

to feel a lot less personable, right?

Like if you got an automated email for everything and

your interview was automated and you're talking to a bot,

that's not a place I want to work. Right.

So you have to rope in automation and this idea

of AI and the recruiting process or across the entire

human resource set of processes in a way that keeps

it personable because it's not human resources without humans.

Right.

It's not like AI resources.

Let's hope not.

There might be a department like that one day.

Yeah, I think as with all of these things,

we have to look at it from A.

What are the repetitive tasks that a human is doing

and their time is better spent with the candidate the

more that you can automate out the tasks that it

takes them to get to the candidate.

Because again, I know HR is there's a bunch

of different functions that HR manages at a company.

We're jumping right into recruiting.

We can talk a little bit too about

how we're using it, even from A.

Let's say you've got employees moving on to

their next role outside of the company, right.

So exit interviews, there's probably some valuable insights

we can use AI for there too.

But can we also enhance the employee experience?

What can we do once they get onboarded are there

layers of automation and AI that we can leverage without

replacing the personal touch, but something that gets them onboarded

quicker and allows us to spend the hiring manager and

the team to spend more time with the candidate so

that they can start adding value faster?

Yeah, that's one area that we've focused in

on is giving the people that are talking

to prospects more data up front.

Nothing's more annoying for the prospect than to

be asked the same questions over and over

again, like, oh, where did you grow up?

And then you get asked five times, well,

why didn't they just share that information internally?

Sure.

One of the challenges that we've don't they have a CRM.

Yeah.

Don't they take notes? Exactly. Right.

Don't they log that or talk to each other or something?

But when you deal with an increasing amount

of applications, you just start running out of

time to log that down right.

Or transfer that knowledge or put

it anywhere for anybody to know.

And the only place it ends up being

is in the recording of the call.

So what we've started looking into is, well, let's

take that information that's being passed along in that

call and have like an automated notetaker, right.

This idea of, well, they said that

they live in College Station, Texas.

Well, let's put that down, let's get the AI

to grab that and put it somewhere right.

And alleviate that human from having to keep

track of like, okay, did we ask this?

Okay, did we ask this?

How many times did we ask that?

It doesn't matter because now we have that information

up front for that next interviewer to just pick

up kind of where they left off. Yeah. Curious.

Have we done any assessment or thought about use

cases before they get the candidate in the door?

So kind of like looking at resumes

in mass or something like that.

I don't want to get too far down the

rabbit hole of that because obviously we already have

automations, actually that are helping us do this today,

which I think are pretty powerful.

Maybe we should spend a couple

of minutes actually talking about that.

When somebody comes to the website, they look

at all these roles, they apply, what happens?

We start tracking as much as we can from that point on.

It actually syncs into HubSpot.

HubSpot then syncs over to Salesforce,

salesforce then syncs over to Slack.

We try to pop that information up to all of

those interviewers, the people that are going to be involved

in that process where we could be doing more.

And this starts to get on the

ethical we're on the boundary of ethics. Right.

HubSpot already tracks all the pages that you go to.

So it would be really interesting for us, and we've

been looking into this like, well, Mel just applied.

Well, how much research did she do on us?

Did she go to like one page.

Did she go to any pages at all?

Or is she downloading stuff?

Is she looking at all of our guides?

That's really telling that

somebody's really actively engaged.

And if we can pull that information up to the

individuals that are on the phone with them, they're going

to have this sense of excitement even without ever talking

to them, that this person's really engaged.

Or they didn't do anything at like

they're just looking for a gig. Right? Yeah.

So in HubSpot, I can tell you from

the marketing seat, I can definitely see when

someone visits a number of pages. Right.

But then being able to pull that into a

summarized view for the hiring manager so they know.

Yeah, they went and visited ten different Ven at work

blogs to see what it's like to work at Venn.

There's obviously some if they're consuming content, let's

say on YouTube, there's not some connectivity between

YouTube and HubSpot in that way.

Usually what I do when a candidate comes in

the office and I have the opportunity to meet

them, I ask, How'd you find us? Sure.

I can go look at your HubSpot record and see

if you came in through organic search or organic social

or word of mouth, if you self report.

Because again, we can't trust all your tech. Right.

It gets you so far, but

you still have to critically think.

So that's one of the things I like to ask.

And then especially if I'm interviewing someone, even

when they come in the door, let's say

we hired them, we do video interviews with

new hires, ask them what their experience was.

Like, I even come in with the assumption, like, I

don't know if you've watched any of these videos, right.

And I kind of tee it up a little ignorant, right?

I'll go, oh yeah.

I'm very know it's not always the case.

Not everyone pours over the

YouTube channel, but they should. Like I did. Yeah.

Like I was like, please, I love

this place, can I work there?

Well, then you started generating those videos, right?

Yeah, with Randall and Co.

And now people are watching your videos.

Yeah, that's pretty great.

But I'm very interested in this from so again, kind of

taking it back to the marketing side of the house.

We work with our HR functions a lot around how can

we get more information into the recruit's hands before they get

on that first phone screen or before they apply. Right.

You would like to think that people are out there

doing their research and being intentional, but obviously there's also

just a lot of folks looking for roles.

And if they're applying for tons of different

jobs, we also can't make the assumption that

until they receive an invitation to actually proceed,

that they're doing that level of research on

every company they're applying for.

So what I've been doing is working with HR

to understand what are the most common questions you're

answering on the phone every single time.

How can we obviously deliver that in different

forms on the website they're applying and we

get all this cool automation pumped into Slack

and we see, know, LinkedIn profile right there.

I can go view the resume, but we're looking

at, okay, here's the email they're going to get.

We've actually plugged in a

personalized video from Scott.

It says, hey, yeah, I realize it's

an automated email, we're an automation company,

but here's my personal touch, right?

So by even just adding that layer of automation,

now our HR team or coordinators don't have to

go kind of sift through and say, oh, did

I send that email to that candidate that I

think we're going to want to go phone screen?

And then in talking with her, I asked,

so, okay, beyond the questions you're answering, can

you send me a couple of transcripts?

Like, what are those calls like?

So then I popped those into OpenAI.

Yeah, I started asking questions like, why

would someone want to work for Venn?

And then I was like, wait, hold on a second.

Why would a senior salesforce consultant want to

work for a company like Venn Technology?

And I was so impressed with

the answer, it brought me back.

Obviously, you can get specific.

Give me five reasons, right?

I just left it open ended, listed out like

ten reasons that were all very on brand, frankly. Right?

And I want to add here one of my favorite

things about my experience in using OpenAI has been the

transcript use case because it's so different than going out

and asking it questions and having it kind of come

up with stuff like sort of the Google search, right?

And that's okay.

If you do that, I'm just jazzed on now.

I don't have to go listen to

a 30 minutes zoom call, right?

It can pull out.

I can say pull out quotes.

I just uploaded all of our video transcripts

from this video series we do called Hired. Yeah.

And in that video series we ask

like, how was your onboarding experience?

Do you feel welcome?

What's the culture like?

Do you have the resources to do your job?

I took, I don't know, like six, eight transcripts,

put them in there, asked OpenAI to come back

with what are all the direct quotes?

And next thing you know, I'm like sending that over to

someone else on my team saying, hey, here's 38 direct quotes

for you to use on Social for the year.

That's powerful to go mine those insights.

And you're not sitting there and then you're copying

and pasting it from a transcript into a Google

Doc and then it needs to go somewhere else.

What do you think people are doing now? Right?

Like, maybe they have no idea that

OpenAI can do any of this.

Are they just not doing it?

Are they doing it by hand?

Are they wishing that they could do, like,

where are these people with these similar ideas?

Like, man, if I could only do this.

Or they're like, there's no way that'll ever happen.

Let's just move on.

Yeah, I mean, I think they're either

not doing it or it's manual, right?

So before I started playing with Was, I

was taking the transcript, listening back, and then

drafting myself or someone on my team is

then drafting a blog post based on that.

And then we're copying and pasting some quotes.

We're putting it in a Google Sheet, and then

we're going and putting that into a design tool,

which actually we've started to learn that.

So canva is one of the tools that we have in

our mix, and they're rolling out AI features all the time.

There's this way that you can do, like, a

bulk en masse, give it here's the quotes.

Here's my template, and it can go

auto generate all of your content.

And we're not replacing the designer. Right?

We're still designing the template and the format, but the

time that it takes to individually do each one, you

still have to go through and clean some stuff up.

But yeah, I don't know your original question.

I just don't think maybe people aren't it's too

laborious to think about going and getting those insights.

Like, why would someone want to work for Venn?

And anyone you ask, you might get a different answer.

If you pump in a number of, like, I know we

say, how many transcripts do you think we really need?

Would we need, like, 101,000?

We don't have that many employees.

We really only have the employees

that we have here today.

So we get to ask 25 people, right?

What do you like about working at Venn? Right?

But where am I capturing that information?

And, oh, by the way, I got to make sure I

get your role, and then I'm having to put that probably

in a Google Sheet, and then I'm like, well, all the

salesforce consultants at Venn like, it because of this reason, right?

It's an excessive amount of time to generate

maybe content that I won't say is extremely

valuable, but valuable enough to if we could

automate it, then we could really benefit.

What do you think about the idea of looking

at these transcripts and, of course, getting insights, like,

what did they say word for word?

But what if you start asking your questions

of, like, well, are they a good fit?

How did they feel about you?

Kind of the sentiment, the questions, right?

Is there ethical concerns in your mind

of, like, okay, Mel came in, right?

And, oh, she she didn't have a great interview, and

this is the do we should we check what the

AI says, and now we got to go listen to

the recording and agree with the AI?

Or do we take the AI as it says?

Where are you on that?

So we've done this a couple of times

with some of our sales calls, right.

And just trying to gauge sentiment, replicating some of

the functionality of some tools out there like Gong.

I do think that I know myself as an interviewer or

I'm asking questions and of course we all like to think

that we're listening in our mind as a still trap, but

it is a good gut check to kind of like especially

if we're naturally more on the optimistic side.

I'd say you probably fall into that category of never.

You get so excited.

We are so excited to have someone in and then you can

kind of have your rose colored glasses on a little bit.

I'm not saying that we should take it at its

word, but in general, I think it's another data point.

I don't think you can ignore it

because there's words that people use and

there's connotations for those words.

It's also probably, oh, did I ask that question?

Did I vet them for this? Right?

Do you think that someone at this

skill level can go do this? I mean, I don't know.

Can you just start asking it questions?

I haven't used it since interviewing someone.

I'm just kind of like spitballing with you here.

If I was to go hire for someone,

I would absolutely love to use it for

that, to give me some additional insights.

Well, now I'm even thinking, while we're spitballing,

what if I start grading you the interviewer?

How good are you as an interviewer?

Is that like an ethical issue that we're

looking back at our own employees to determine?

Well, Mel isn't that great, but with

this one candidate, she really like, is

that something that we shouldn't be doing?

Because now we're kind of relying on this third party,

if you will, to tell us about our own employee.

I would argue tools like Gong

are probably already doing that.

There are probably people doing that, right?

Because and you can use that as a training tool, right.

So, hey, I've noticed, especially as we start to grow

and scale and you're building out full fledged teams and

you're hiring for the same role, let's say we're always

going to get if it's like a BDR.

And we've got kind of a methodology making

sure that as we continue to bring in

managers of that team, that they are asking

the questions that have led to successful hires.

So you still have to look at granted, when we

were brought in, we probably don't have transcripts of that.

No, too old, right?

I've been here almost three years, so we don't have the

data to go well, if we were going to go get

another Mel, here's the questions we got to ask.

Can we do that clone? That'd be great.

I'd love a clone.

So I do think that I hear what you're saying on

the ethical concerns, but I think it is very valuable from

an internal training tool yeah, the grading aspect, I don't know

how you start to set those thresholds right.

I think you need large volumes.

And then again, you have to be able to point

to who are our top performers in these roles, and

then what are the skills that they have today?

What did they have when they came in?

What types of questions?

Who did they meet with?

There's probably just a level of assessing and

just going door to door and asking those

questions before you can actually accurately say, that

was a B plus interview.

Well, one of the things that you're making me

realize is that there's kind of two underlying themes

here, is that we've got a lot of data

that amounts to a transcript, write a call recording.

And when you have a lot of those across a

vast amount of candidates, there's a lot of insights that

you can get, like maybe all of the transcripts for

this one particular role, we can now gain a ton

of insight even if we didn't hire right?

Is that where you're yeah, yeah, we didn't hire them.

But how knowledgeable were they?

Are we targeting the right people?

Can they answer the know if you throw out

a really difficult question about some Apex class in

salesforce, right, or some kind of Ruby method or

whatever challenging technical question and nobody can answer it,

of all the transcripts that you have, well, maybe

you're asking the wrong questions or maybe you're targeting

the wrong people, right?

And those are some of the insights that could

be really valuable, but they're really challenging to get

because you have to analyze so much data just

to determine or even think of these questions that

you should be asking of the AI.

So kind of bringing this back.

To be clear, we haven't operationalized any of this.

We are kind of doing these,

like, one off use cases, right?

But bringing it back to other functions

of the kind of the HR professional.

What are some of the other things that some

of these tools might be able to do?

Do you have any ideas around enhancing the

employee experience or getting them onboarded quicker?

We've kind of discussed a knowledge base of sorts.

One of the things that has been really interesting in

either the studies or the headlines is that people are

using one particular headline that I remember is in a

call center for this onboarding use case, they've seen a

lot of potential value in onboarding new people to get

them up to speed faster.

And I think it kind of related

to a chat bot that they trained.

They effectively gave the entire knowledge

base to a large language model.

And now this large language model looks

a lot like their best agents. Right.

They're the ones that wrote the

answers, and they know they're correct. Right.

And now this chat bot is just regurgitating stuff.

So as a new agent onboarding,

I just got asked a question? I have no idea.

Well, they're automatically feeding that

question into the AI.

The AI is pulling up the correct response

like, hey, you should respond like this and

here's the link to the answer.

It will speed up the onboarding

process to make your newbies right.

The people that are just coming in the door get much

closer to that expert in a much shorter amount of time.

I think where you're naturally going to have challenges

is where you don't have a knowledge base.

A lot of what we do is

nobody has ever done this before.

So it's not like there's some giant

knowledge base documentation, transcripts of recordings.

We've never done it before.

So it's really difficult to get that AI to

look at something and come up with right answers.

Well, I feel like there's a lot that we're

just barely scratching the surface on here, especially as

it pertains to recruiting on the backside of it. Right.

So now you have someone who has come and

they've worked and let's say they're moving on the

next thing and we're conducting an exit interview.

I think those are I've always kind of

been interested in that because you get sort

of assuming people are allowed kind of let

their hair down, they don't really have it.

They're on their next thing. Right.

So as long as they're being transparent and

open to the process, you can learn some

really unique eye opening things about your business

that maybe you didn't know before.

Because I want to say that folks are afraid to say it,

but that level of what will happen if I do say it. Yeah.

You're actively asking my feedback.

Yeah, your guards down.

So I'm interested in understanding not that

we've had a ton of these, right?

We've had like a handful.

But as we continue to grow and scale, like a

lot of the people listening out there growing, scaling companies.

Can you mine those exit interviews

for the same kinds of insights?

Are there things that we're missing as an organization

or things that we can be doing better? Right.

Like you could be doing this all throughout your entire

we do poll surveys, we ask folks, we check in,

we've got kind of like our own internal satisfaction meters.

But I do think, again, with the assistance of a

transcript, you might be able to gauge or glean similar

insights that we're talking about as they leave.

This is an idea that we are actively implementing.

And it's this ability to ask the AI,

ask this chat bot about this transcript.

And I think that's where there is

a significant amount of use cases in

that one spot, it's like active insights. Right?

Like, how did the exit view

or the person that's leaving, right.

How did they feel about it?

And you can ask those types of questions

that you wouldn't necessarily ask the interviewee.

Like, hey, how did that go?

How did they feel?

You can just start asking hundreds of questions where

if you were the one doing the exit interview,

you're not going to want to listen to every

question that I could possibly come up with.

No, we talked about this

in our leadership team meeting.

It came up and I was like, I know

you guys sent this over to me right now.

I'm not making time for it.

What are the Cliff Notes? Right?

And then someone on the team went back,

listened to calls, took notes, collated that into

some insights, that went into slack.

Yeah, that was probably like, I don't

know how much time that took. Hour.

2 hours of somebody's time. Yeah.

Well, you have this idea of insights.

Let me ask questions of the transcript and

then you have this predetermined automation AI sentiment.

Let's ask the same questions of the transcript

and then let's just do that every time. Right.

Those two different use cases are going

to be valuable in their own ways.

And when you pair them together, it's going to

be even more valuable because you do have folks

that don't have enough time to sit down because

your calendar like, if we wanted to sit down,

we can't meet for like three weeks.

Well, the exit interview is long gone from our minds,

so I think there's a lot of opportunity there.

But all of it starts with really sitting down and asking,

well, if we wanted to automate this or if I wanted

some here, I need to come up with the use cases.

So you have to be thinking through those things.

Speaking of use cases.

So let's transition into our headlines segment.

So found this recent headline.

IBM's HR team saved 12,000 hours

in 18 months after using AI. What?

18 months I've been doing this.

They're on waiting edge.

I mean, I guess it is IBM.

So 18 months after using AI to automate 280 tasks.

And we quote, we're spending time

on the things that matter.

It's something that we say over and over again.

I said this in the first episode, right?

The goal is to get people back to

what they're good at, which inherently means they're

doing things that they're not good at.

Nobody is like a natural expert at data

manual entry or replace that with something else.

They get good at it because they do it over and

over again, but that doesn't mean that they like it, right?

18,000 hours.

I mean, sorry, 12,000 hours, 18 months.

That's a significant amount of time that was saved.

Those people can now do other things now.

Hopefully they were repurposed, right,

for doing more important things.

And that's kind of the hope and the idea.

But I imagine a lot of these things are

automation tasks and less like, I need critical thinking

on this to deliver an answer like sentiment analysis.

So I wouldn't be shocked, right, if a lot

of these folks are kind of pushing paper, if

you will like clicking the buttons to keep the

process rolling, but they went through enough work to

determine that it was 12,000 hours. Right.

Let's just say that's $100 an hour.

That's 120 grand.

That's not small dollars.

That's a decent amount.

Like, that's a full time employee, right? Yeah.

So I'd be really interested in understanding the tasks

that were automated here, but it's certainly no small.

It's no small like it is a big deal.

All right, it's time for our hot take.

So I know we here at Venn, when

we bring in almost all candidates, I think,

have some sort of assessment, right?

So, like, you have your application and

then you meet with a few folks.

You got phone screening, in person interviews,

and then depending on your role.

So for me, in marketing, it was, go away

and write a blog for someone, an integration consultant

or a salesforce consultant that looks a little different.

I don't know exactly what those assignments are,

but with tools like Chat JPT, should hiring

managers be concerned about these kinds of things?

I think it depends on what you're wanting

that person to come in and do. Right.

If you want them to come in and

write content and you're going to openly say,

well, yeah, you can use Chat GBT. Right.

And you ask them to you're assessing their

skills by having them write their own blog.

Maybe you're now trying to assess the wrong thing.

That's the first thing that comes to mind.

You need to reassess your assessments.

The second thing that comes up in my mind

is that it is possible, for instance, in our

qualified IO assessments, I did this, I typed in

the question that was being asked, OpenAI generated code.

It was in the correct syntax.

It did function, but it didn't function

exactly the way it should have.

And that gets back to, well, I need to check the work.

I think what we're going to run into here

is people are going to use this to effectively

generate the answers and they're going to be correct,

and it's going to look like they're really good.

The problem that they're going to have is that

they're going to get into the gig and they're

not going to be able to do it.

So maybe there needs to be a level of remember back

in grade school when you kind of turned in your paper

and sometimes there was the how did you get there?

You kind of had to present to the class.

Oh, that's genius. Yeah.

So it's do the assignment right.

And then show me how you got there. Right.

And maybe you can use these tools,

but maybe you can use Chat GPD.

But tell me what the prompt was,

how you got to your research.

Why do you think that?

Why do we need to go after this keyword?

I think in person or recorded interviews, oral

examination is going to become even more. Important.

Like, sure, I had the bot do the work for

me, but you have to understand that there's this one

spot where the bot did it wrong, and if you

don't understand the syntax and the coding language itself and

all of the methods and how all this works, you're

not going to be able to spot that problem.

And those are the things that you need to

hone in on in your assessments of these individuals

to determine if they're really good or they're not.

Another one that really takes me back to college.

I took a class that I didn't like, and

they were oral exams, so you can't BS anything.

And I remember that the prophet, that's

how you become a good BSR.

Maybe here I am.

But he called me out and he's like,

you didn't study any of this, did you?

And I was like, not really, because I really

don't like, I didn't say this to him because

I didn't want to make him feel bad, but

I just didn't really like the class anyway.

I got an F, of course,

naturally, because I didn't study.

But that's going to be the overriding thing that pops up

now is that, hey, Mel, tell me of these concepts.

Tell me why you did this. Right?

And you can't plagiarize that when you're

sitting in a room without a computer.

There's no calculator to do the math problem for you.

I mean, I remember it was just three

years ago when you were interviewing me for

this role, and I can only imagine now

how that conversation might look a little different.

So you might ask me, well, what's your take

on tools like applying Chat JPT to marketing?

And then let's say I'm in favor of it.

Okay, so how much time is that going to save you, and

what else are you going to focus on in this role?

I mean, I do think that you could actually

have some really powerful conversations with recruit or candidates

if you ask those kinds of questions.

Again, it's not going away.

The AI functionality is, whether we like it or not,

just being rolled out left and right in app.

Like, even if you're not out there using Chat GBT, ODS,

are the multiple many apps you're using to do your job

right, have some level of AI that you didn't say you

didn't turn on or off the boat's already left.

You either need to swim to the boat and get on right,

or you can stay back there in the back of the bus.

Well, I think that's a wrap for today's episode, but hopefully

for anyone who's kind of either a hiring manager I mean,

even if you're not in an HR function, hopefully you kind

of got some ideas for how you might be able to

leverage AI to enhance that recruiting experience.

Enhance the employee experience once

they come on board.

I don't think that your job's going away.

If you're in HR.

The one thing that I've taken out of our conversation today,

and it just happened at the very end, kind of like

a little simple light bulb moment for me was, we will

now have more brain power because you said there still has

to be like, that oral examination, right?

If you're running, like I've been in that place where

you're running from meeting to meeting to meeting, and you

get to a point where your battery is low and

the level of brain power that you're able to apply

and level of critical thinking, that's probably a better way

to say it is less.

Hopefully, by automating some of these tedious things

that burn us out and frustrate us, we're

able to be more present with people and

to give them an honest, authentic chance.

The human element.

The human element. Yeah.

That's what I hope.

So we want to know what questions

or thoughts you have on the topic,

so email us your take at thejunction@zentechnology.com.

Thanks again for tuning in. Say this again.

Until next time.

Keep exploring, stay curious, and

embrace the power of AI.

It's not going anywhere. Keep it automated.