In this episode of the Future Proof HR podcast, Kellie Shotwell, VP of HR and Operations at SMS Group of Companies, talks about applying AI and automation across the candidate and employee lifecycle. Kellie shares how her team improved staffing operations by building smarter workflows into systems they already used rather than starting over with an entirely new technology stack.

The conversation focuses on a central tension for staffing and HR leaders: how to automate administrative work without making the candidate experience feel transactional. Kellie explains how SMS uses mobile-first onboarding, personalized workflows, automated scheduling, skill-based candidate routing, and ongoing employee check-ins while keeping one recruiter connected to each candidate throughout the hiring process.

Kellie also shares the operational results behind the approach. SMS reduced time to fill by roughly 75 percent, increased the rate at which candidates were hired by clients by 35 percent, extended assignment lengths by about 40 percent, and brought hospitality no-show rates below 2 percent. These gains came from faster communication, automated reminders, better use of existing candidate data, and fewer manual handoffs.

The episode closes with a practical approach to AI adoption. Start with repetitive administrative work and compliance-heavy processes, involve frontline employees in workflow design, introduce changes gradually, and preserve human judgment when evaluating candidates. AI should move the process forward, but it should not replace the relationship-building that helps candidates feel seen and supported.

Topics Discussed:

• How to apply AI across both the candidate and employee lifecycle
• Balancing high-volume hiring speed with a personalized candidate experience
• Building mobile-first application, onboarding, and document workflows
• Using AI to route applicants to better-fitting jobs and recruiters
• Re-engaging passive candidates within an existing talent database
• Automating 30-, 60-, and 90-day employee and client check-ins
• Reducing manual work in payroll, tax documentation, and compliance reporting
• Introducing AI workflows without eliminating recruiter ownership of relationships
• How employee involvement and small pilots improve AI adoption
• The impact of automation on time to fill, assignment length, hire-in rates, and no-shows

If you lead staffing, recruiting, HR operations, or talent acquisition and are trying to identify practical AI use cases, this episode offers a grounded model for automating repetitive work while preserving the human connection that candidates and employees value.

Additional Resources:

• Cleary's AI-powered HR Chatbot
• Connect with Kellie Shotwell on LinkedIn

I really feel strongly that the key to using AI is not letting

2

it take away too much of that human touch.

3

you could really miss an awesome

candidate for something, and you

4

missed it because you didn't take a

few minutes extra to really look at it.

5

Our time to fill has decreased by about

seventy-five percent across the board.

6

Speaker 2: What do you

think of when you hear AI?

7

What do you think of when you

hear from your leadership that

8

we're gonna be implementing a

lot of AI in our environment?

9

If you're in the staffing and

recruiting space, as soon as you

10

hear that, you might be thinking

that this is gonna cost you your job.

11

You might be thinking that this is

just gonna be another initiative that

12

makes me more stressed out, and it's

gonna have a steep learning curve.

13

You might be thinking that, and in this

particular case that we're gonna share

14

with you today, you would be wrong.

15

What we're gonna share with you today

is what one staffing firm did in their

16

organization, both in a high volume and

general professional perspective, that

17

paid massive dividends in terms of their

time to fill, in terms of their show

18

rates, in terms of the project durations,

all the things that matter to staffing

19

organizations and all the things that

matter to frontline staffing employees

20

that actually impact their pay- paychecks.

21

Think about it.

22

If you could shorten your time to fill,

extend your project durations, and have

23

your candidates convert to full-time

employees more often, that would be

24

a win across the board for you, and

it would mean more opportunities for

25

you within the clients that you serve.

26

And joining us is Kellie Shotwell,

who's the VP of HR and Operations

27

for the SMS group of companies.

28

She's skilled in HR law, permanent

placement, contract staffing, and employee

29

leasing across diverse industries.

30

She leads a team of 25 management

and administrative staff and oversees

31

1,500-member workforce delivering

innovative staffing solutions that

32

reduces costs and optimizes performance.

33

What

34

Kellie

35

Speaker 2: is going to share with us today

is how she leveraged existing systems and

36

looked for key workflows and automation

opportunities that resulted in massive

37

wins for her organization and her team.

38

Jim Kanichirayil: Kellie,

welcome to the show

39

Kellie Shotwell: Thank you, Jim

40

Jim Kanichirayil: So I'm looking

forward to this conversation mainly

41

because you operate in a space that I

spent quite a few years in, and that's

42

in the staffing space, so this is

gonna be an interesting conversation.

43

Before we get into the weeds on the

topic, which will focus on applying

44

AI and automation across the candidate

and employee lifecycle, I think it's

45

important for you to set the stage a

little bit and tell us a little bit more

46

about the the company and your role.

47

So why don't you take the mic and

fill us in on that a little bit?

48

Kellie: The company is SMS Group

of Companies, and it comprises of

49

actually two staffing companies.

50

One specializes in things like

security, janitorial, facility

51

services things like that.

52

A lot of municipalities, a lot of

government work, that kind of thing.

53

And then the other company, SMS Staffing

Solutions, is more of a complete full

54

boat staffing HR solution company.

55

We started as a very traditional

staffing company over 25

56

years ago, and we've grown.

57

As the times have changed, we've changed.

58

We have actually developed

several verticals.

59

We work strongly, as I mentioned, with

a lot of the municipalities in the area.

60

We do a lot of bid work, but we

also have a vertical in hospitality

61

staffing, which is very volatile,

very quick requires a lot.

62

So AI's come in well there.

63

But that's how we've developed

and where we're at now.

64

And like I said, we've been in

business for over 25 years, and we

65

are in about six states right now.

66

That's where we're at.

67

Jim Kanichirayil: One of the interesting

things about your organizational

68

structure is that you have an arm

that's high-volume staffing and you

69

have another arm that's more general

professional i-in terms of the setup.

70

I guess one of the things that I'd

be curious to hear about from your

71

perspective is when you're looking at

two different flavors of candidates that

72

you're bringing into the pipeline, what

are some of the specific challenges that

73

you have to account for when you have

both the high-volume environment as well

74

as a general professional environment,

which tends to have a slower hiring cycle?

75

What are the things that you have

to account for from a recruitment

76

perspective, an onboarding perspective

that people that are outside of

77

staffing might not be aware of?

78

Kellie: When you're operating on

both sides of the fence like that,

79

one side requires a very quick

turnaround, very quick response

80

time with every applicant directing

them where they need to go quickly.

81

So that's a very fast-paced side, and

that AI has really helped with that a lot.

82

The other side, you've really gotta

make sure that you're using the AI to

83

help move things forward and avoid a

transactional experience, because on that

84

side, there-- the candidates are looking

for a much more hand-holding situation.

85

There's gotta be a lot more touchpoints,

there's a lot more discussion,

86

a lot more explanation, review.

87

And so it-- you have to be able to

set your system up so that it has

88

the proper cadence based on the

type of hiring that you're doing.

89

So the system needs to allow for that.

90

So that's something we've been

learning with AI and that we've been

91

developing and cr- you know, de-

pushing into different directions.

92

Jim Kanichirayil: One of the things

that you mentioned that I thought

93

was interesting was that in your

general professional staffing,

94

you're leveraging AI so that the

process isn't very transactional.

95

And why that's interesting to me is that

when I look at high-volume staffing,

96

that's extremely transactional.

97

It's speed to candidate that usually

ends up getting the win because those

98

candidate, the best candidates in

that space aren't on market very long.

99

So it's critical that

you operate very quickly.

100

So how did you bridge the gap between

both of those candidate pools when

101

you have two very different ways that

those candidates need to be treated

102

in order for you to win the placement?

103

Kellie: I think the key here is and I do

agree with you, Jim that they're gonna

104

go with the best offer that comes first.

105

But there's also a little bit of a

caveat there because something that I've

106

learned and that I feel is very unique

to our company and how we handle things

107

is, yes, that's true, money's important.

108

But you know what?

109

People wanna know they matter.

110

They wanna know they're important.

111

They wanna know they're

a part of something.

112

They wanna know that they're appreciated.

113

They don't wanna be a

number on a piece of paper.

114

They wanna be Bob or Sally

or what- whoever they are.

115

And so one of the things that we've

done is with the AI and the workflows,

116

we can develop the workflows within our

system so that they they reach out more

117

often to those people very quickly.

118

So when those people apply, there's a very

quick response that comes back to them.

119

It includes their first name.

120

It includes some detail.

121

It's very personalized, even though

it's an AI workflow response.

122

And it quickly, the trajectory is

to quickly put them right into an

123

orientation or a training session,

and they confirm right then and

124

there that they're gonna attend this.

125

So we've already pulled them into

where they're gonna come and see us.

126

It's not uncommon that this process could

start and three days later they could be

127

at work with orientation and everything.

128

But that was workflows.

129

It's developing the workflows.

130

And the workflows that we use on the

other side that's a little slower will

131

be a little bit more spaced out, a

little more developmental, a little

132

more human interaction in between

the steps because we're pulling

133

those along a little bit differently.

134

But that's how we,

that's how we've done it.

135

And it's been very important.

136

I personally am involved in writing all

of our workflows and our AI chatbots

137

because I don't want them to sound

like every other company out there

138

that, oh, you can tell it's a chatbot.

139

I'm sure at the end of the day they

probably know most of them are, but

140

I really try to put a personal touch

and a personal sp-spin on them 'cause

141

I think that's really important

142

Jim Kanichirayil: So let's

dig in there a little bit.

143

You're taking a hands-on approach in

establishing the AI workflows and the

144

chatbots, and you put an emphasis on

making sure that as much as you can make

145

it make it sound like it's not a bot.

146

So what was involved in setting up

those workflows and also getting

147

the language down so that it doesn't

sound like you're talking to a robot?

148

What did you do there?

149

Kellie: Honestly, the first step was

getting really involved with our IT

150

and our developmental teams that are

doing all this behind the scenes.

151

And it was funny because

that's not my area.

152

I'm not IT.

153

So as that can be a bit complex.

154

So I had to first understand the process.

155

And what I love is developing

the trigger points.

156

If this happens, then this goes out.

157

So the trigger points were the first

step that was really important.

158

Then from there, I just,

honestly, I asked my staff.

159

They do it every day.

160

They talk to people every day.

161

W-we would start to write the dialogue,

write the, the copy we wanted to

162

use, and then we would test it.

163

We would test it within our company.

164

I would take my core team.

165

I had, a core team that I

used for testing a lot of this

166

stuff, and I'd run it by them.

167

"What would you rather hear if you

were apply-- What sounds better?"

168

And sometimes we'd even pull in

some of our family members and say,

169

'cause they don't understand the

industry, and that's a good feedback.

170

And we'd say, "Okay, if you were

applying for a position and you

171

got this one, which of these

would make you feel more involved?

172

Which would make you feel like

you are a part of something

173

and not feel as transactional?"

174

And through that feedback, we

developed those and being willing

175

to change when we needed to change

it too, as things developed.

176

But that was really the

biggest thing we did was…

177

I think it's, I think it's really

important, and I think a lot of people

178

in our industry, and especially at my

level I'm the executive vice president.

179

I'm not recruiting every day.

180

I'm not talking to these

candidates every day.

181

I'm helping my people.

182

I'm leveraging their workload

so they can do their job at

183

their best of their ability.

184

So I think that people at my level

sometimes forget, go back to the

185

people that are doing the work,

ask them, get their feedback.

186

They're talking to people every

day, all day, so they know.

187

They have good feedback.

188

And sometimes I think that when big

change is made within companies,

189

they skip that step, and that's

really an important step because

190

you're gonna get great feedback there

191

Jim Kanichirayil: So what I find

interesting about what you're

192

describing is that I keep thinking

about the two sides of the organization

193

and the two candidate pools that

you're dealing with, as well as

194

the frontline employees internally

that, that have to interact with it.

195

So So when I think about the high-volume

candidate generally speaking, these

196

folks are going to do everything that

they need to do off of their phones.

197

Whereas if I look at a general

professional candidate, it might

198

be a mix of primarily using their

phone or using an in-home setup.

199

And when you're thinking about your

messaging, your workflows, your

200

chatbots, how did you account for the

two different ways that your candidate

201

pools are likely to interact with

whatever infrastructure you s- you

202

put into place from an AI perspective

so that it's seamless for them?

203

Tell me about the process

that went into that.

204

Kellie: So that's a good question.

205

We recognized very quickly two factors.

206

One, you can't rely on email.

207

A lot of people do not use email.

208

They may even have an email,

but they check it rarely.

209

They're just not email people.

210

It's surprising to me the level of

people that don't utilize email.

211

But everybody has a smartphone today.

212

Even people for my entry-level

positions that are just starting in

213

the workforce or my, just getting

started they all have a smartphone.

214

And so we made sure that all of the

workflows and the various chatbots

215

and the AI that we did develop,

we made sure that it transcended

216

between the email and smartphones.

217

And we've actually, at this point in time,

taken our entire onboarding operation and

218

it's completely automated w- utilizing the

AI, and it all can be done on the phone.

219

They never have to look at a computer.

220

Everything can be done on the phone.

221

They can sign documents on the phone.

222

They can upload documents on the phone.

223

They can do everything on a smartphone.

224

We've tried to make it

as simple as possible.

225

That wouldn't have been possible

without the AI being able to write those

226

workflows and help us bridge those gaps.

227

Jim Kanichirayil: So that's pretty

cool that you've been able to take your

228

entire onboarding process and be able to

execute and complete it over the phone.

229

Which brings me to another question that

I have that, that comes up, is that if I'm

230

dealing with a candidate population that

might not be regularly doing this, dealing

231

with all of the paperwork completion,

I-9s, and all of that stuff that happens

232

in onboarding, now you're having that

process be facilitated over the phone.

233

Doesn't that create a barrier where people

might get frustrated or not understand

234

what they're looking at, not be able to

access it to the degree that they wanna

235

access it, and they abandon the process?

236

Did you encounter that?

237

And if so, how did you account for it?

238

Kellie: I don't think we really

have had too much trouble with that.

239

We do track that.

240

Where we see abandonment is

in the application process.

241

A lot of people will come in, and

they'll start the application.

242

One of the things we learned real quick

is to make sure that if a job doesn't

243

need a resume, like an entry-level

job don't request a resume because

244

there's a l- large amount of people out

there that just don't have a resume.

245

So we're very careful about that.

246

So anything-- That's where we've

worked to pull in more engagement.

247

We've tried to use AI.

248

It's more at the application.

249

Once they've gotten through that

part and the job has actually

250

been discussed, and there's…

251

Because the onboarding documents

are happening once they

252

know there's a job for them.

253

We have somewhere for them to work.

254

We have an income we've discussed.

255

We've had a schedule.

256

And so now very rarely do

we have a problem with that.

257

If we have, let's say, a candidate

that is challenged with doing things

258

on their smartphone, we have a couple

different ways to deal with that.

259

One of them is we've developed

a lot of quick process guides

260

quick little, with visual anchors.

261

This is what it will look like.

262

This is what you need to do.

263

All you have to do is do this.

264

And that takes care of probably

the majority of the questions

265

we would get at that point.

266

And in the worst-case scenario,

sometimes especially with people that

267

maybe are re-entering the workforce

or are not very savvy with smartphones

268

and even computers we'll have them

come in, and we have computers

269

set up here, and we'll help them.

270

We'll put them at a computer.

271

We'll let them log in, and we'll

do it online, and we help them

272

with the onboarding that way.

273

They can still do it-- They're still doing

it all electronically, but we're assisting

274

with walking them through the process.

275

Luckily, we don't have to do

a lot of that 'cause that is

276

a little more time-consuming,

but it's certainly available.

277

We've made, accommodations

for those types of things.

278

Jim Kanichirayil: So that,

that fills in the blanks on

279

the candidate experience side.

280

I wanna flip this internally and

think about implementing all of these

281

processes and AI tools internally.

282

Generally speaking, when I've talked

to high-volume staffing organizations,

283

they have multiple people at different

stages of the candidate life cycle.

284

So you have one person that'll

look at, whatever application,

285

and then they might forward it

to a recruiter for an interview.

286

That recruiter will interview and then

put the person to work, and then you

287

have an onboarding person that takes

care of all of the onboarding stuff.

288

Theoretically, I'm not saying

that's how your process is,

289

but this is what I've seen.

290

Now theoretically, if you apply AI to

the process like you described, I'm

291

seeing a scenario where two out of three

of those people that are touching the

292

candidate could likely be shifted off

or eliminated from the process where AI

293

automates the application process, AI

automates the evaluation process then the

294

only touchpoint is the recruiter at the

interview, and then you can completely

295

automate the the onboarding process.

296

When you were talking about rolling

AI across the candidate life cycle

297

internally how did the team react,

and what were the implications as this

298

became a more mature process to how

your staffing levels changed internally?

299

Kellie: I think some of the recruiters

when we first started getting

300

into it more heavily were a little

resistant, but it's more like anything

301

else when it's something new, and

you don't know how to do it, and

302

you have to learn a new process.

303

Of course that was-- that's been o-

over a year ago now, so I don't really

304

have anybody that's resistant now.

305

I think the key is that you mentioned

something that's really interesting, and

306

it's completely different than what we do.

307

We don't have them touched

by more than one person.

308

Unless, and this is how this works.

309

So if someone applies, and let's say

they're applying for a janitorial

310

or a facility maintenance position,

but as they're applying, it-- they

311

don't have the right credentials for

that, but they have some experience.

312

Through the AI screening tool, we

discover that they have forklift

313

experience or something like that.

314

It will immediately redirect

them to the proper recruiter

315

that has those types of jobs.

316

That's a huge savings right there.

317

When you think about the s- the

operation and how it works, and you

318

have to talk to one person, and they

discover, "Oh, I can't use this person.

319

They're, they don't have what I'm

looking for this job, so now I'm

320

gonna hand them off to this other

recruiter," that's very time-consuming.

321

But when you have an application come

in, and you have AI going in there and

322

going, "Oh no, they don't qualify for

this job even though they applied, but

323

hey, they qualify for this one over here."

324

And then it forwards that applicant

over to the other individual recruiter

325

that can actually utilize that person.

326

First of all, that's a huge

time saver right there.

327

Second of all, that person's

being touched from the beginning

328

to the end by the same person.

329

They're-- We don't onboarding isn't

handled by somebody different.

330

It's not like that.

331

And I think that's where you develop

a strong rapport with your people.

332

So much so that in one of my larger

clients where there is a pass-off after

333

they're hired to an HR manager they keep

wanting to go back to their recruiter

334

because they love the interaction they

had with that recruiter, and they, and

335

the recruiter has to gently push them

now you have to work with the HR manager

336

because you've been hired, and you're an

employee, and I'm over here recruiting."

337

So it's interesting.

338

But

339

Kellie Shotwell: I really feel strongly

that the key to using AI is not letting

340

it take away too much of that human touch.

341

And so that's why I think we've

tried to develop-- The way we've used

342

it, we've been-- I think we've been

pretty smart in how we've used it.

343

We've really thought it through, and it's

been sprinkled back and forth between

344

the human touch, and that's why we do

have just the one person that handles

345

a candidate from beginning to end

once we discover where they're headed.

346

Jim Kanichirayil: The thing that I

like about what you described is in the

347

application process, if the intake by the

AI identifies a different skill that this

348

person is good for, that person can be

rerouted to that role and that recruiter.

349

That's a really cool process, which

actually had me thinking about how can

350

that be applied to your existing database

of candidates, because I know that one of

351

the biggest areas where any organizations,

and particularly in the TA function, they

352

lose a lot of opportunity is by constantly

going outside the organization to find new

353

candidates instead of looking at what's

existing in your database as a potential

354

fit for an open role that you have.

355

Because all that process, how tight

that process is, has all sorts of

356

impact in terms of your time to fill

and the number of candidates that

357

you need to present before you get

a fill and all that sort of stuff.

358

So tell me a little bit more about

how that's optimized for existing

359

candidates now that you've actually

talked about how it's optimized for

360

new candidates coming into the funnel.

361

Kellie Shotwell: As you can imagine,

we have been with, and I th- hopefully

362

it's okay to mention the name,

363

I have over 50,000 people in

my database at this point.

364

When you've got a-- When you've

been in the same software for that

365

long, that's what's gonna happen.

366

Obviously, they're not

actively lurking, all of them.

367

But that's how large the database is.

368

So when we have a job come in,

and let's say it's something

369

we haven't done in a while.

370

Recently we had one that came in,

we haven't done plastic injection

371

mold operators in quite some time,

but we had one come in for that.

372

And the beauty of it is part of the

intake for every candidate is, and we

373

really encourage them, and even after

we place them, we encourage them, "Go

374

back in and check off every skill that

you have, every skill you qualify,

375

the types of things you'd like to do."

376

So now what happens is I've got AI tools

written into the system, so when this

377

plastic injection mold, for example,

came in, first thing we could do is

378

we can go in and we can put the AI

on it, go through the database, find

379

people, and we give it the criteria.

380

We're looking for people that have

listed this, or this, plastic injection,

381

manufacturing, whatever it might be, how

many years we're looking for, et cetera.

382

And the system will go through

and pull us out a list of people.

383

Once it does that, it'll also generate

a message to them saying, "Hey,

384

we've got this great new opportunity.

385

If you're looking to make a

change, are you interested?

386

And here's how we can get

on our calendar," et cetera.

387

And think about if I have a database of

50,000, and even if it only identified

388

500 people that had plastic injection

mold experience, I can contact

389

those 500 people in five minutes.

390

My people can't do that.

391

There's no way.

392

It would take me a week of four

people calling all day, every day.

393

So that is an amazing thing in itself, is

that the ability to search that database

394

and utilize the, what we already have in

front of us to find candidates, especially

395

when new positions come in with a client,

that's something that we haven't--

396

'Cause when you're filling something

all the time, there's a constant flow.

397

There's a natural flow of candidates.

398

But when a new job comes in that

maybe you haven't done it in a year

399

or six months, you, all of a sudden,

you kinda gotta pump that well again.

400

You gotta get it going, and

that's a wonderful way to do

401

it, and it's so efficient.

402

It's so efficient.

403

That's been a real deal breaker for us.

404

Thomas Kunjappu: This has been

a fantastic conversation so far.

405

If you haven't already done so,

make sure to join our community.

406

We are building a network of the

most forward-thinking, HR and

407

people, operational professionals

who are defining the future.

408

I will personally be sharing

news and ideas around how we

409

can all thrive in the age of AI.

410

You can find it at go cleary.com/cleary

411

community.

412

Now back to the show.

413

Jim Kanichirayil: So that has me thinking

about one other element of the candidate

414

life cycle, which is, when a candidate

isn't active, you have two buckets of

415

candidates, active and passive candidates.

416

And depending on the space that

you're in, passive candidates can

417

often be your best opportunities to

fill a role that you have come open.

418

So how have you leveraged AI within

the environment to make sure that your

419

passive candidates, the ones who aren't

looking, are still engaged at some

420

level with either the organization in

general or a recruiter in particular?

421

Kellie: That's a great question,

and actually that's something we

422

didn't do right away, and then we

kind of happenstance fell on it,

423

a- and we realized it was a gap.

424

And we recently, fixed that because

what we did now is we have, I think

425

we're doing it quarterly right now.

426

We have it's almost like a news

bulletin approach, if you will, goes

427

out quarterly, and it's just hey…

428

And it might have a piece of information

that might grab their attention, could

429

be something that's going on right

now in terms of workforce, whatever.

430

It'll say, "Just glad you're

still a part of our organization.

431

If you're looking to make a change,

don't, hesitate to reach out to us.

432

If you've got, a- acquired new skills,

you may wanna go into the system and

433

update it so that we know that you're

available for these types of positions."

434

and then it may list a couple

of positions that we have.

435

"We've recently added this and that."

436

And by engaging them the other

thing is we have a an AI that sends

437

out a happy birthday message on

everybody's birthday in the system.

438

Just touch points like that, I

think that keeps-- They alwa-

439

they're- we're always in their mind.

440

They always know we exist.

441

They're not gonna forget,

even if it was two years ago.

442

Because honestly, a lot of our people do

get hired into some of our client sites.

443

And they may work there for a few years,

and then there may be a change, and

444

then they're looking for something new.

445

So we do try to maintain that connection.

446

Obviously once or twice a year, we

also have to send out something, "Hey,

447

update us on your email or your address

if it's changed or your phone number."

448

but I think by doing those things, it's,

it keeps them engaged so that when they

449

do get to the point in their life where

they're looking to make a change, we're

450

gonna be one of the things they think of.

451

We'll be one of the avenues

they wanna check out and see

452

what's available out there.

453

Jim Kanichirayil: Got it.

454

So up to this point, we've just been

talking generally about stuff that

455

happens during the candidate life cycle.

456

So when you take inventory of the

candidate life cycle from the process

457

of them actually seeing the job all the

way through onboarding, and you look at

458

that entire, th- those entire stages,

what were the biggest areas that were

459

creating the most amount of administrative

load for your team, and then how

460

did you go about solving for those?

461

Kellie: I think that the

largest pain points, of course,

462

are the heavy administrative

functions that are required.

463

The candidate applying and making sure

they're directed to the right person

464

so that they're handled properly.

465

If you were doing that in

man-hours, that takes a long time.

466

AI has really facilitated that

and made that much quicker.

467

The next step is making sure

they're getting on somebody's

468

calendar for an actual interview.

469

That happens much quicker

now because of AI.

470

And then the other side of it

is when you do have a position,

471

you're ready to do the onboarding.

472

The AI interaction for onboarding has

been just the probably the largest time

473

saver, I should say, of everything.

474

Because when you have an AI function

that's getting the onboarding done

475

and moving that forward so that

you can put this person to work,

476

I think that those have been the

biggest impact in the process.

477

They're very heavy administrative.

478

They take a lot of time.

479

I know I've saved at least two or

three recruiter spots that I haven't

480

needed because the administrative

flow is so much quicker now that my

481

people can handle so much more workload

482

Jim Kanichirayil: I'm glad that you

mentioned that because that's one of the

483

things that people need to think about

is how can I apply this internally to

484

not necessarily get rid of people, but

become more efficient so that you can

485

reallocate those to higher level internal

candidates or higher level initiatives

486

and priorities that actually help your

organization go to the next level.

487

So let's switch gears a little bit.

488

We've been talking about

the candidate life cycle.

489

What are some of the ways that

you've utilized AI to impact

490

the employee life cycle?

491

Kellie: Good question.

492

So we actually have done

some pretty unique things.

493

One thing we do is we have automated

messages that go out at 30, 60, and 90

494

days of the start of a new position.

495

Just checking in, how are you doing?

496

If you have any questions, give us a call.

497

Remember, we're here for you.

498

We're your HR rep, whatever you

need, et cetera, to keep those

499

lines of communication open.

500

In several instances, we also have

we have like reviews that go out.

501

So something goes to the hiring

manager at the client site that

502

says, "Hey, how is Bob Smith doing?"

503

And there's questions, four or five key

questions that they check real quick.

504

It has to be fast 'cause a hiring

manager at a client site's not gonna

505

spend a lot of time, but letting us know

how that candidate's doing so that we

506

know that we've made a good placement.

507

At the same time that goes out to the

hiring manager, we have one go out to

508

the employee asking kind of the flip

side of those questions about how's it

509

going for you, how, how are you doing?

510

And again, it's usually just where

they can check off some answers and

511

maybe a place to write a few things in.

512

That has been really helpful to

make sure that we always are on top

513

of how our people are doing, and

our clients have really liked that.

514

And the other thing is updating their

files their tax documents when they

515

need employment verifications trying

to think of all the different way…

516

All these things have now got an AI

component that has just made that

517

administrative task so much easier.

518

And I know that my compliance

coordinator has a lot of use for

519

it in the sense that she tracks,

say, workers' comp for injuries or

520

unemployment for unemployment claims.

521

She has IAI components that can go into

the system and extract the data that

522

she needs for those types of things.

523

We have to do EEOC filings every year,

and the system has got AIs written

524

in to pull all that data for us.

525

There's a lot of reporting that we

have to do for different agencies, and

526

it's so much easier now that we have AI

to pull that candidate information in

527

Jim Kanichirayil: How long

has that been in place?

528

Kellie: We are going on two years

that we've had the full boat

529

of everything I've just discu-

about two years that we've been.

530

And the nice thing about the system

we use and I'm-- this is not meant

531

to be an Aviente pitch, but they are

very focused on growing their platform

532

and increasing the capabilities.

533

So I feel like every time I turn

around, they're adding new capabilities.

534

One of the newest one is the ability to

do chat right within the system with a

535

candidate, which is something we could

do text messages or emails within the

536

system, but there wasn't that actual

chat function like a text message.

537

And so now we'll be able to do

that, which is gonna be really nice

538

Jim Kanichirayil: So one of the things

that you mentioned when we were talking

539

about the candidate life cycle is

that based on the automations that

540

you've built in through the system,

it's saved you one or two recruiting

541

headcount since you've been doing this.

542

On the employee side you mentioned

automating the check-ins from a

543

30, 60, 90 day period, periodic

updates to employee files.

544

You've had impacts on compliance.

545

When you look at those sorts of

automations that you've built in,

546

how is that process being handled

before, and what did it mean in terms

547

of time saved once you automated

that through your internal systems?

548

Kellie: I feel like it was the Stone Age.

549

I think back to how we did it just

two, three years ago, and I don't

550

know how we survived, quite frankly.

551

Everything was touched at every

single point and that's very

552

time-consuming, number one.

553

You're chasing a candidate around.

554

You have to get ahold of them, and

then you're leaving messages, and I

555

can't even describe the administrative

nightmare when I think back.

556

It drives-- It just boggles my mind, quite

frankly, that we ever did it that way.

557

I feel like it was, rubbing two sticks

together to start a fire back then.

558

And the number of people that I would've

had to have internally just to do the

559

job that's a big drag on your overhead.

560

And what I like about the way we're

doing it now is that with less overhead,

561

my people get bigger commissions for

the work they do, so they're happy, and

562

they're more than willing to do more work.

563

But I think that's, th-that just

the manual part of everything, and

564

when you touch something manually

that many times, you really open

565

yourself up for more errors, too.

566

That's a big thing.

567

Think of taxes and employees' taxes.

568

They fill a form out, and then we have

to put it in the system, and we have

569

to set it up, and then it's gotta go

to payroll, and th-there's a lot of

570

touching of somebody's tax information.

571

Now, thanks to AI, when

they apply, they just…

572

We click a button, and the system

says, "Oh, they live in Detroit,

573

and they're working in Detroit.

574

Okay, here's all the tax

forms they need to…"

575

The system does it all, sends the tax

forms to them that they're, that they need

576

for that particular job in their location.

577

They fill them out, and it

flows right back into the system

578

and sets it up for payroll.

579

Nobody has to key anything in.

580

So not only is there no chance

for error, but it's gonna be

581

whatever the employee w-put in.

582

So if something's wrong,

it's not our error.

583

It's gonna be their error.

584

So we've really managed to cut out that.

585

The errors alone that we've cut out

by being able to set up sy-symmetry in

586

systems like that has been amazing, and

before that would've all been manual.

587

Everything was manual.

588

Even with a

589

computer system, you still had to

walk the paper through the system

590

Jim Kanichirayil: So when I think about

what you're describing my, my head

591

goes to the change management process.

592

Whenever you implement or institute

or introduce any sort of new thing

593

it's supposed to have, some vision

of what the future's gonna look

594

like by using this new process.

595

And I can imagine myself sitting at the

desk looking at an executive who's telling

596

me, "Hey, this is gonna be, like, the next

big thing," and I'm sitting here thinking

597

what I do is plenty fast, and it's gonna

take me longer to learn this new way of

598

thinking or new way of doing than it is

to just continue doing what I'm doing."

599

Did you encounter that, and how

did you overcome that resistance

600

to this change in workflow?

601

Because people don't like change,

and there might be plenty of people

602

within a- any organization that are,

like, pretty damn fast even though

603

the process is manual as hell.

604

So did you encounter it?

605

How did you overcome that?

606

Kellie: I did encounter it.

607

And a couple of individuals

probably more than others.

608

Some were a little more eager.

609

But first and foremost, I think

the first step was engaging them.

610

There's a difference between

handing you this and saying, "You

611

have to do this," and saying,

"Hey, let's take a look at this.

612

Let me show you how this works.

613

What do you think?

614

How can we work it into the system?"

615

If you make them a part of the

solution, I think that you get

616

that buy-in a little bit easier,

so that was something that I did.

617

And I wanted to get them excited about

it, so I, we would start small with maybe,

618

"Hey, just take one of your clients, try

it this way, and i- see what you think.

619

If you don't see the

value, then let's talk."

620

But if you can get them to that

point, they always saw the value.

621

When they saw how much time they could

save and again, when they saw how

622

much work they could get done, they

eventually, appreciated that, and

623

it was easier to bring them along.

624

So I think it was, one was engaging

them and making them a part of

625

the decision and the solution.

626

Not force-feeding them the entire

meal at once, but in pieces.

627

Let's start slow.

628

Giving them the time to learn

acknowledging their feedback

629

and trying to work within that.

630

I think that was real important.

631

I think that brought a lot of people

together quicker than it would have had I

632

just doled it out and said, "Here, this is

a new system, and now you have to do it."

633

so I think and we did do some reward

systems, like the first person to complete

634

the automated client and get their

workflow set up and, things like that.

635

Make it fun, make it engaging make

it something that they wanna see.

636

And my one employee that was my

absolute worst, she's the epitome

637

of everything you just said.

638

She fought it and fought it, and then

I literally, I put her aside, because

639

she was in our hospitality department.

640

So I put her aside, and I worked with

the other team, the other recruiters,

641

got them all up and running and got

things going well there, and then I went

642

and just did real hands-on with her.

643

And I showed her, and I showed her some

reports and I don't think she understood

644

how much control she had over the

workflows, and I think that's key too,

645

is saying, "Hey, this is your client.

646

I'm not gonna tell you which

workflows you need to use.

647

I'm not gonna tell you which AI you want.

648

But here's what's available to you.

649

Here's how it works.

650

What do you think would

make your job easier?

651

What's gonna help you do

your job faster every day?"

652

So when I did that with her, and I

gave her that one-on-one experience

653

with myself, and I really worked

with her as well as a trainer, she

654

couldn't believe what she was missing.

655

By the time I got done, she was like,

"Oh my God, I should've been doing

656

this," and she was like, she was all in.

657

So she took a little more work,

but once she saw the fruits

658

of the labor, she was there

659

Jim Kanichirayil: So one of the things

that I think about when I hear what

660

you're describing, both from the candidate

lifecycle perspective and also the

661

employee lifecycle perspective, the lens

that I have, I've been in staffing for

662

a while as well, and when I think about

staffing, I don't necessarily look at

663

staffing, regardless of what type it is,

as a particularly innovative industry.

664

Oftentimes when I talk to people

who are running staffing operations

665

"This is how we've always done it.

666

We're gonna continue doing it this way."

667

why I find your story interesting

is that you've maximized the

668

systems that you've had, and you've

automated as much as possible.

669

Now, the thing that usually convinces

staffing leaders and practitioners to

670

actually do something different is when

you can actually point to reductions

671

in time to fill and extensions or

lengthening of time on assignment

672

or retention at the end client.

673

What can you share that points to

either one of those things as an

674

impact or an outcome from all of

the different automation efforts

675

that you've taken internally?

676

Kellie Shotwell: Our time to fill

has decreased by about seventy-five

677

percent across the board.

678

Kellie: One of the areas that's

highly competitive with hospitality

679

is you get an order on a Monday,

it's for an event on Wednesday, you

680

literally have, forty-eight hours.

681

I can have thirty people in ten minutes

scheduled because of AI and the ability.

682

So the time to fill is, like I

said, seventy-five percent across

683

the board on average decreased.

684

And because of the engagement, we

have found that our hire-in ratio

685

has gone up by thirty-five percent.

686

Hire-in of our candidates to the client

site, it's gone up by thirty-five percent.

687

And the length of

assignments have extended up.

688

It's been about forty percent.

689

And I think that has a lot

to do with the engagement.

690

Again, asking those managers,

"How are they doing?"

691

They show that we care, we're engaged.

692

It's like anything else.

693

When you're engaged,

you can move mountains.

694

So those are some of the

immediate increases that I can

695

definitely or benefits I should

say, that I can definitely state

696

Jim Kanichirayil: As you were talking

through that, the other thing that I was

697

thinking about is that in a high-volume

environment, one of the big problems

698

that you deal with is you might go

through all of that process and hire

699

30 people to show up tomorrow for a

role, and half of them don't show.

700

What's been the impact on that

side of the equation where you

701

have a lot of attrition before they

even show up for the first day?

702

How have you managed that

piece of the equation?

703

Kellie: That's a great question, Jim,

because actually in hospitality I

704

follow those metrics very closely on

purpose because of the type of work

705

that it is, and I can honestly tell

you our no-show ratio is less than 2%.

706

But the reason it is goes back to

AI, because now what happens, let's

707

take that example of a Monday.

708

So Monday I get a request

for, 30 people on Wednesday.

709

Immediately we go out, the

assignments go out through AI.

710

They go to everybody's phone click.

711

Predetermined people that my, my person,

my recruiter in hospitality has selected.

712

She knows who she wants.

713

She knows who her A people are.

714

So she has sent this out to them.

715

They have either accepted or rejected.

716

So let's say she sends it to the

first 30 people, she fills 20 spots.

717

Then she sends it to 10 more and so

forth until all 30 spots are filled.

718

Once they're filled, now that was Monday.

719

Tuesday afternoon, they're all

gonna get a reminder, "Hey, you're

720

booked tomorrow at this time, at

this location for this reason."

721

They get a reminder Tuesday afternoon.

722

"If for any reason something's changed,

please let us know right away."

723

And it works.

724

We'll get one or two people that'll say,

"Oh God, I forgot I took that assignment.

725

I can't work tomorrow."

726

They'll immediately interact with

the recruiter, with the chatbot

727

that goes to the recruiter,

let them know they can't work.

728

We backfill those positions.

729

Then on Wednesday, day of

event, morning, same thing.

730

Goes out about midday.

731

If it's a, say it's a 4:00

start, it'll go out by noon.

732

We still have time to backfill if

there's a last-minute call off.

733

So when you get to the point of

the event actually starting, we've

734

already backfilled the positions that

would've normally been a no-show,

735

and that has really been major.

736

When I look at our metrics

from two years ago and today in

737

hospitality for our no-show rates,

unbelievable the difference.

738

I have many clients that are tracking at

100% show ratio for a month at a time.

739

It's insane.

740

But that's because of the AI and the

reminders and, putting it in their

741

face because, you know how people

are when you're dealing with people.

742

Sometimes they're forgetful.

743

Jim Kanichirayil: Yep, absolutely do.

744

So great conversation.

745

I want you to tie everything

up in for our listeners.

746

So when we think about all the things

that you've done from an employee

747

lifecycle perspective as well as

a candidate lifecycle perspective.

748

Really what we've been talking

about is leveraging your existing

749

systems and automating as much as

you can within those systems to

750

optimize what your recruiters can do.

751

So when you think about that and you're

advising other staffing leaders on

752

how they can do the same thing, what

are the key things that they need

753

to keep in mind as they're looking

at not only the candidate lifecycle,

754

but also the employee lifecycle,

and what are those biggest areas of

755

opportunities for them to automate?

756

Kellie: I think first and foremost,

look at your heavy administrative.

757

Things that are just pushing a

paper forward, pushing a process

758

forward, getting something

from, point A to point B.

759

That's where you can use

your AI and your automation.

760

Make sure you're not giving

up the people component.

761

Make sure you're not giving up what

we call around here the warm and

762

fuzzies, because that's where you

develop the loyalty with your people.

763

So I think unfortunately, there's a lot

of companies that aren't doing that.

764

I think they're just throwing the AI out

there and trying to use it from A to Z.

765

That's not, in my opinion,

that is not the answer.

766

It has to be integrated within the human

component, within the human touch piece.

767

So I think looking at your

heavily administrative areas

768

is the first and foremost.

769

The next is your risk factors, where you

need to make sure that certain documents

770

are signed or certain things are in place.

771

Look at where in, within your compliance.

772

Those are our heavy cost areas, and

so anything you can do to button those

773

areas up is always helpful to the system.

774

And I think the other component is you

really have to look at your recruiting

775

process and how you're screening

these people and how you're using

776

the AI, because in this day and age,

there's not a lot of unemployment.

777

You said yourself passive

recruiting is really important.

778

We wouldn't survive without it.

779

Diamonds come in a lot of shapes and

sizes and places, and if you're not

780

somehow making sure that you're looking

at a resume, looking behind the scenes,

781

looking between the, reading between

the lines, talking to that candidate,

782

Kellie Shotwell: you could really miss

783

an awesome candidate for something, and

you missed it because you didn't take a

784

few minutes extra to really look at it.

785

Kellie: So I think it's really

important that AI accentuates

786

but doesn't replace that process.

787

I have so many success stories of someone

who came in thinking that they only

788

qualified for this or that, and because

we were able to really look deeper, we

789

found a great position for them where they

were really able to excel that that if

790

you had just let AI screen them out, you

would've never known where to put them.

791

It's a melding.

792

It's not a one or the other.

793

It's a melding

794

Jim Kanichirayil: Great stuff, Kellie.

795

Appreciate you sharing that.

796

If people wanna continue the

conversation, what's the best way

797

for them to get in touch with you?

798

Kellie: I can be reached

on LinkedIn, of course.

799

I have a LinkedIn profile under my

name there as well as my work email.

800

Speaker 2: Thanks for

hanging out with us, Kellie.

801

I appreciate you sharing your story with

us, and it's been a pretty impactful

802

discussion that we've had so far.

803

So I'm sure that our listeners, and

particularly those who are in staffing

804

environments, are gonna be interested

in, hearing this story and seeing how

805

they can apply it in their own world.

806

When I listen to the stuff that you

described, I think one of the things

807

that worries a lot of people when

we get into this discussion about

808

AI is that they often think about it

in terms of reinventing the wheel.

809

And what I found particularly interesting

about how you actually approached

810

this is that a lot of your story was

about optimizing existing systems and

811

identifying and automating, key processes

and workflows that were ripe for

812

automation, and using that as a lever to

become more efficient as an organization.

813

And what was interesting is that you

applied this not only to the candidate

814

life cycle but also the employee life

cycle, and the impact speaks for itself.

815

We talked about your massive

reductions in time to fill.

816

We talked about the benefits that you

got in terms of longer assignments.

817

We talked about your show rates and

how that improved over time, and this

818

wasn't, all driven out of completely

new systems that were put into place.

819

This was driven by optimizing existing

workflows, identifying areas for

820

automation, and then pushing those

out to not only the candidates but the

821

internal employees who were spending a

ton of time on the manual workflow side.

822

So what that means for everybody

that's listening is that when you

823

think about AI and automation and

optimizing your workflows internally,

824

you don't have to start from zero.

825

You can look internally at the

systems that you currently have.

826

You can look internally at how you're

currently doing things from a workflow

827

perspective, and those are the easiest

opportunities that might be in front

828

of you that you're ignoring because

you are looking at the wrong context

829

and not thinking about this, at a

more granular and practical level.

830

So that's what stood out

to me in this conversation.

831

I appreciate you sharing that with us.

832

For those of you who've been

listening to this conversation,

833

we appreciate you hanging out.

834

If you like the discussion, make

sure you leave us a five-star review

835

on your favorite podcast player.

836

And then make sure that you tune in

next time where we'll have another

837

leader hanging out with us and sharing

with us the things that they've been

838

doing with AI to future-proof HR.