Okay, Actually is a show about what’s really going on at work — and what we can do about it. New episodes come out every other Thursday and are always under 30 minutes.

AI can help us research, organize, generate, compare, and execute faster than ever. But it can also produce something polished enough that we don’t immediately notice it has solved a completely different problem from the one we actually had.

In Part Two of my conversation with my dear friend and former colleague Maury Postal, we talk about what happens when AI makes execution faster and cheaper without necessarily making the direction any better. We get into why “move fast and break things” is a questionable operating philosophy, how AI makes ideas more abundant but not necessarily more valuable, and why augmented intelligence may be a more useful framing than artificial intelligence.

This is a conversation about AI, but it’s also about judgment, accountability, creative distinctiveness, and the human conversations we can’t delegate.

This episode stands on its own, but you can also listen to Part One: Episode 10: The Brief Is the Work.

00:00 AI Editing Gone Wrong

02:19 Part Two Setup

03:10 Rigor Versus Speed

05:41 AI Makes Bad Direction Cheaper

09:11 Accountability Swim Lanes

11:50 Ideas Are Cheap; Judgment Isn’t

13:27 Algorithms, Sameness, and Slop

17:38 Process And Stepping Away

20:01 Three Four Questions Before You Begin

22:56 What Humans Still Need to Own

Four questions to ask before a project begins

1. Who will determine whether this work succeeds?
2. What does the work actually need to prove to that person?
3. How are we going to separate signal from noise?
4. What does a human being still need to own here?

About Maury

Maury Postal is EVP of Creative Strategy and Innovation Services for Grayling North America. He has more than 18 years of experience across creative strategy, storytelling, and digital innovation, helping organizations turn complex ideas into work that people can understand — and that can actually accomplish something.

He has worked with global brands including American Express, Audi, Ford, Genesis, P&G, Synchrony, and Japan Airlines. His work also includes AUTONOMY, a feature documentary made with Malcolm Gladwell that premiered at SXSW.

Connect with Maury

Instagram

LinkedIn

Find me

OkayDoak.com

[email protected]

It probably goes without saying that I rely

2

pretty heavily on AI these days.

3

I do a lot of writing and publishing,

and I have zero audio/visual editing

4

skills of any kind, so being able

to use AI tools to help me edit

5

a podcast is genuinely useful.

6

And at one point, I thought I had

completely nailed the process with Claude.

7

We were working through these

increasingly elaborate editing guides.

8

They were color-coded, there were

indicators for what to cut, what to

9

add, what to record, what needed a light

patch, blah, blah, blah, blah, blah.

10

They were complicated, but they worked.

11

And I wanna note they weren't complicated

because I made them complicated.

12

Every time we did another round,

Claude would add some kind of new

13

graphic, new feature, new indicator,

and I was like, "Oh, this is helpful.

14

This is easy to read."

15

I save one, I tell Claude,

"This is the template.

16

We're using this going forward."

17

Easy breezy, except apparently

asking Claude to use the template

18

that Claude created was becoming

increasingly challenging for Claude.

19

Every time I uploaded a new

transcript, I got a different system.

20

I had new colors, new categories, new

terminology, and most frustratingly, words

21

would just disappear from the transcript,

22

Which was a problem, because preserving

every word and telling me what to do

23

with it was literally the entire point.

24

I needed an accurate transcript

that showed me what to cut and

25

fix so I could make those changes

in Descript, my editing software.

26

And what I got was a beautiful, organized,

increasingly colorful document that

27

did not do the job I needed it to do.

28

But don't worry, it used way more of my,

, session usage or memory, so, , it also

29

slowed Claude down, which is not helpful.

30

And the output looked better.

31

The process was questionably faster,

but the underlying material was wrong.

32

And I keep thinking about that

because it kind of captures one of

33

the stranger risks of working with AI.

34

It can produce something polished

enough that you don't immediately notice

35

it has solved a completely different

problem from the one you actually had.

36

So two weeks ago, my friend and former

colleague, Maury Postal, joined me

37

to talk about why the creative brief

isn't paperwork before the real work.

38

The brief itself is the work.

39

But our conversation kept going

because Maury and I have never

40

been especially gifted at ending

a conversation expediently.

41

And once you start talking about

42

you inevitably end up talking about AI.

43

So this is part two.

44

You don't need to have heard part

one first, although obviously

45

I would love it if you did.

46

This time we're talking about what

happens when AI makes execution faster

47

and cheaper, but doesn't necessarily

make the direction any better.

48

I'm Karen Doak, and this is OK Actually,

a show about what's really going on

49

at work and what we can do about it.

50

Before we got into AI directly, Maury and

I were talking about one of the reasons

51

people skip the brief in the first place,

which is that rigor sounds slow, and

52

process sounds like the thing preventing

you from moving quickly, especially in a

53

work culture that loves telling people,

"You need to move fast and break things."

54

So I asked Maury how he thinks about

the tension between rigor and speed

55

Maury: I also think that the idea of

rigor sounds scary and this is kind of

56

anecdotal, but it sounds like it's one

person trying to claim ownership idea

57

or a department or a function instead

of saying, "No, no, no, I want to make

58

sure everyone is accountable, has a clear

swim lane so that we can move fast."

59

, The move fast and break things mentality,

I think it's hysterical because it's like

60

you and I are, know, we don't work slowly.

61

um, the breaking things part is

bizarre because it's like actually,

62

you know, if you said that to like

a neurosurgeon, would you want

63

them to move fast and break things?

64

Probably not.

65

So I think , there's this idea

of, yeah, move quickly, but why?

66

at some point, , that compromises

any value you might get out of it.

67

It's like, this sounds so stupid, but like

this dehumidifier that we have, if you

68

turn the fan speed up, that doesn't make

it dehumidify faster then the dehumidifier

69

can't like suck in as much moisture.

70

It just sounds like it's

doing more, but it's not.

71

that's exactly what this is.

72

It's like of busyness, the activity

of busyness, and the activity of

73

like writing a brief because, oh wow,

we're, we're so physically in the weeds

74

tackling, instead of, , moving on.

75

Because oftentimes, again, it

might not be a final thing.

76

It might be a starting point.

77

be to prove that something else didn't

work or something else needs to be done.

78

So I, I'm kind of in that headspace,

especially when everything's

79

up in the air right now.

80

You know, like is AI going to replace

every human decision being made?

81

They're like big questions that are

being asked, so people are panicking

82

instead of relying on, you know,

everything that we've built to date

83

to get here and build on top of it.

84

Karen Doak: I liked the range of

analogies you were able to cover in that.

85

, Talking about AI a little more,

part of your new mandate and your

86

role is about scaling creative

work , in an AI environment.

87

So do you think that speed pressure makes

the brief discipline better or worse?

88

And do you think that AI makes bad

direction cheaper to execute and

89

therefore more dangerous as well?

90

Maury (2): Yes, yes,

yes, and all the above.

91

I think of it is, like last year

or two years ago, you know, the

92

idea of what an AI could do is

exciting it felt like it was useful.

93

, And often, it didn't feel

like a shortcut, right?

94

Like, it felt like it was

someone you could work with.

95

I think there was this mentality of like,

"Oh, this replaces people, this, this

96

speeds up processes," instead of more

research faster pressure testing an idea

97

before you have a discussion about it.

98

But I also think that that messy

process we just talked about is

99

actually really critical to figuring

out why things work or don't work.

100

And so, yes, AI makes things cheaper and

faster to kind of get to fail at, make

101

decisions about, but you're not seeing how

it impacts The other factors that aren't

102

being addressed because of these tools

are sophisticated enough to be a human.

103

Like, they always oversimplify.

104

And the crazy thing is when you have

to think and work with an AI, you do

105

think differently than when you're

working with a, a person that is--

106

that's a certain decision-maker.

107

I know, you know, in every role

I've ever been in, you get on client

108

calls, you work with decision-makers

in a way that fits them, fits their

109

needs, how they process information.

110

I mean, all the trainings we

did at Ogilvy about, you know,

111

what personality type are you?

112

How do you work with people who

are different, who are similar?

113

How do you make sure they have

the right amount of information?

114

It might not be too much or too little.

115

And all this stuff has to be

informed , and can be informed by AI.

116

But I like to say that that's just

like saying, you know, internet

117

or using another term to describe

something that's very nuanced,

118

is that you have to understand

the right tool for the right job.

119

And you know, I always thought it

should be augmented intelligence,

120

not artificial intelligence.

121

'Cause the whole point like, I know

for me, it helps me deal with anxious

122

decisions that I might have about, will

these two cars fit in our parking garage?

123

know, is this, , data

point actually significant?

124

Like, where did it come from?

125

at this, , collection of things

and pull out four points I

126

might wanna look at deeper.

127

Now, that last point sure

is oversimplifying it.

128

I might miss something, but given the

way that work has to happen, you need,

129

you need something to act as that filter.

130

I just think it's really important to not

lose the humanity on the other side of it.

131

you have to, think,

132

Karen Doak: Yeah

133

Maury (2): think about a lot of what

is being made right now and where

134

it's going, and it's being out to

talk to other non-human things.

135

Like, you're are being made to

exist in a world without the

136

consideration for the needs of people.

137

I think that's weird

138

Karen Doak: Yeah, we could go down a

whole rabbit hole on what that means

139

for society, but I think that idea

of responsibility and accountability

140

in the process is super important.

141

Whose responsibility is it in all

of that to stop and say, "Wait, do

142

we actually know where we're going?"

143

Because, that's what would

inform any AI, that's what would

144

inform any member of the team.

145

But, you know, who actually

really owns that responsibility?

146

Is that sitting within the strategy

team, creative, the client, someone else?

147

Maury (2): question.

148

I think what is really critical is that

everyone has a specific role, and I know

149

we used to do this great brainstorming,

technique basically gave everyone a role

150

. Because there was always a person who

was the naysayer, that was their role.

151

Because there was always someone that

was a scribe, someone who acted as the

152

client, someone who acted as the customer.

153

But , it was really helpful because

you could play out all those things

154

in a moment, and you didn't have the

devil's advocate, which is a problem.

155

You didn't have someone who felt

as though they didn't have a

156

voice because everyone had one.

157

And so, the accountability thing is

really important , you know, again,

158

if we think about how a lot of

these places were structured, that

159

was the role of an account person.

160

But , we had to be account people too.

161

Like, all of us are now everything.

162

So I think assigning that role when you're

creating a brief is really important.

163

Who are the players?

164

It, it matters internally too.

165

and even in a small team, it's

really critical to have clear

166

swim lanes because often you'll be

duplicating efforts you don't need to.

167

I, I think what, , is really tough

with the timelines and the tools

168

that we have, like AI, is that

it's the idea of failing faster.

169

So like just get in there,

make it messy, fail faster, and

170

then come out the other side.

171

you can't.

172

success and, being successful is

all about a clear articulation of a

173

problem statement, a, clear reading

of the, current state of things, and

174

then a clear endpoint, which I think

is often missing because, , if no

175

one can agree and consensus can't

be achieved, it's often like we'll

176

just kick the can down the road.

177

And, you know, I know I'm guilty of that

now, just certainly 'cause there's not

178

enough time in the week, day, month.

179

I mean, it-- we're all kind of

doing a million things at once.

180

So I think that a brief in its best form

actually helps to make better use of

181

people's time because it limits the swirl

182

Karen Doak: It does make me think that

especially with AI , that ideas are

183

almost cheaper and more disposable to us.

184

This idea that, like, I can just type

in , "Hey Claude, here's my brief.

185

Generate 100 creative ideas for me."

186

There was a point in time when, , and this

was not when we were working together,

187

but I remember to launch one particular,

, not very exciting mid-size sedan, we

188

submitted over 200 ideas to the client

and just kept cycling them through, and

189

we had a spreadsheet of every idea that

had been submitted and rejected and why.

190

And then we're at this point where

we're like, "Eh, can we revisit?"

191

'Cause there's only so many ideas.

192

And I do think now it's just a lot

easier to say, "Give me an idea."

193

And, and that whole idea of making sure

that everything that you're saying,

194

doing, building, looking at ties

back to an actual strategic purpose

195

sometimes is missing because it's just

sort of like a lot of vibe checking,

196

Maury (2): checking, but

197

Karen Doak: is, is how things move forward

198

Maury (2): you know, like every great

creative person I would always come

199

into at Pick a Place would always have

their way of doing things, their filter.

200

And I would often be shocked by how

much they would articulate that for you.

201

And in California, I remember

getting there with this creative

202

team that had just started and

hearing it like just very clearly.

203

Like they were very declarative about

the, "Huh, well, we flip it, we do

204

this, this is how we think of it."

205

And I was kind of amazed

that that's what they did.

206

I mean, it actually forced me to figure

out the what do I wanna be when I grow up.

207

But, , it worked because at that time

you couldn't duplicate something that

208

quickly, so you could come up with a way.

209

Karen Doak: Yeah

210

Maury (2): what's so tough is

that, With every pivot to video,

211

influencer, pick a thing, pick a

medium, it's all become so same-same.

212

And the cost of standing out now

means the algorithm won't pick you up.

213

So we are in a very weird moment of

creating a brief for some might feel

214

useless because you're like, "I'm gonna

do whatever the algorithm tells me to do.

215

I'm gonna see what performs and do the

same thing," instead of saying, "Well,

216

I know part of this campaign or part

of this thing or whatever it is needs

217

to fit and perform that way, but I

can buy my way into people's eyes and

218

minds and put whatever I want there."

219

and also if everything is so

accessible, then nothing is.

220

So I think there's a weird moment

right now where all the pivots are

221

kind of inverting on themselves.

222

And because, you know, if you're

listening to this, you're probably

223

part of the same algorithm that

you and I are a part at this point.

224

So you're served like the same four

things, often any big movement now

225

or creative thought comes from not

looking at your phone from pulling

226

out a book finding a thing and

making a connection that isn't made.

227

And I think what's great about that

is that that's how you differentiate.

228

I mean, we're def- we, we've completely

exited the try it and see, and now

229

it needs to be the it has to work.

230

then when you think about what every

generation does as a reaction to

231

another, living a life that feels,

you know, algorithmically significant

232

versus just like existing, there's

gonna be a huge reaction to that.

233

And so I think what's

important to remember is like

234

Everything's always changing.

235

And so maybe we're exiting the

era where the brief wasn't needed

236

'cause too much changed, but

237

we're probably going back to an era where,

people , that do things that are craft

238

forward, are, you know, about something

only a specific region can do or see,

239

that's actually what's succeeding in these

algorithms now because slop feeds slop.

240

And I think we're totally at the point now

where, you know, a lot of that training

241

data that helped make all the AI tools

useful is now so itself apart that,

242

know, you have organizations that have

been feeding themselves the same slop.

243

They can't generate

things that like anyone.

244

They just kind of have a non-answer.

245

Karen Doak: Well, that, makes me think,

Maury, of like the feeding itself slop

246

is almost the idea of the copy and paste

creative brief we talked about before

247

of like, "Hey, that thing you did to

launch product A, let's just repeat

248

it again, but in a different color

for product B," , and continue to do

249

that, which we, I mean, we got those

directions again and again and again

250

Maury (2): 'cause the copy-paste actually

started with a really good idea It's

251

like if the copy-paste, I mean, there

are things called Xerox machines,

252

Karen, I don't know if you're aware.

253

But when you made a Xerox of a Xerox

of a Xerox of a Xerox, you got nothing.

254

I think that's still the idea of copying

and pasting now, where, I mean, I think

255

we're all guilty of like, if I'm busy,

I have to use an AI tool to organize

256

something, paste it into something else.

257

I'll read it, I'll edit

it, but it's happening.

258

It's very weird, like we all

are operating at , that speed.

259

there will be a moment, I don't know

when it's gonna happen, where, you

260

know, you will see the first successful

company break through in this like

261

post-everything era by being exceptionally

human probably being different on

262

different channels, and that will be

the thing that's copied and pasted.

263

Like, most social media is slop.

264

I mean, if you go on any platform, entire

accounts are just generated garbage.

265

So, know, it's, I think, very

naive to think that now is an

266

exceptionally different time.

267

I think we're still in this mess.

268

I think what always

helps is the "Yes, and."

269

What worked from before?

270

What didn't?

271

Where are we right now,

and what's applicable?

272

Process always works, no matter what.

273

It always does.

274

a brief in some format is critical

to organizing and gaining consensus,

275

, in the same way that, like, a

transcript is very helpful now

276

because it can speed up meetings.

277

But if you speed up everything

deliver very quickly, what

278

are you doing during the day?

279

Like, that doesn't feel

like you're achieving much.

280

So, I think there's, an important

grounding that probably has to

281

happen about it is, the value of

you as the person creating the brief

282

or stewarding the brief and then,

, thinking about where it goes next.

283

But also, it takes a lot of great thinkers

to posit that and to kind of get us there.

284

So, I think it's really critical to both

put the right amount of thought into

285

things and then also step away from it.

286

I think the stepping away part is massive,

because even if you do your best, you have

287

the most I'll say rigid process with the

right people involved, stepping away and

288

coming back is still the best way to just

determine Why or why not did this work?

289

If I'm a stakeholder, the client I'm

doing this for, they probably need a

290

break too, because there's the same

type of pressures that inform the

291

successes as it does the failures.

292

I think it's a little bit of like

find the balance and then find the

293

purpose of both the document and

you yourself that are creating this.

294

Maybe you want to evolve how you fit in

295

Karen Doak: I, I love

everything you just said.

296

, I do wanna call out that this

podcast is not sponsored by

297

Xerox, , and we are normally very

cautious about brand name mentions.

298

But, that point about stepping

away also, I would call out

299

that, , your audience is not living

and breathing this stuff either.

300

So taking a step away is actually

your best chance of giving the

301

perspective that your audience has,

which is I'm living my life, and

302

then I'm confronted with this idea,

concept, thought in front of me anyway.

303

, And I also think that that point of

why did something work, you know, what

304

worked before, okay, but why did it work?

305

Because sometimes the reason why it

worked was no one had done it before,

306

and replicating it, it's already been

done, and so that's not gonna work again.

307

But if you can't agree on why, , you

can end up going in circles.

308

So , if you could get every sort of

client or key stakeholder to nail,

309

a defined shortlist of things before

a brief goes out or a project kicks

310

off, what would those things be?

311

Maury (2): I'm a client and I'm

putting it together, I would say

312

very clearly who is going to be

determining the success of this project?

313

What is the ultimate thing that

you have to prove to that person

314

who's, , considering it a success?

315

And how are you filtering

out noise from signal?

316

Because I would argue,

317

Karen Doak: I love number three

318

Maury (2): It's weird because

even those words are like AI

319

words now, but it, it's true.

320

, The understanding very clearly what

brand stands for is often the job of

321

the person putting it into practice,

whether that's the agency or that client.

322

It kind of means something a little

bit different, and I think it's really

323

clear to make a point of what one thing

is supposed to do to get to another

324

checkpoint, another, another moment.

325

Because too big of an ask, too

sprawling of a, of a project

326

actually won't help the thing.

327

And I would argue that having a

thought partner and a client is

328

probably the most critical step.

329

Like, they take the inputs, filter it

as well, and understand on their side

330

upstream in a decision-making way, you

know, who, who wants to hear certain

331

things, what words are good or bad to use?

332

And on the other side, it's, "Well,

what has worked or hasn't worked, why?"

333

To our point.

334

, But I would also argue that the hidden

factors at play should be surfaced, the

335

things that no one wants to talk about.

336

Like, you might have the best brief

in the world, but you're putting it

337

out into an ecosystem where, you know,

you have a big misinformation problem.

338

You do all the things that you want.

339

Did you solve the misinformation problem?

340

this thing supposed to, like, suppress it?

341

A lot of things I see are

trying to a Band-Aid on a thing

342

that should be fixed first.

343

And oftentimes it can't

344

Karen Doak: Totally agree with that.

345

Maury (2): I organizationally get it.

346

You're a division that

can only do so much here.

347

You can't affect that.

348

Totally fine.

349

You need to acknowledge it in the brief.

350

Like sure, there might

need to be two versions.

351

Oops, that's the other scary thing.

352

There might need to be two versions.

353

There's the one you're actually

working on, then the one that

354

maybe someone has to approve.

355

That's like the deep, dark secret

of you, you wanna be successful, you

356

the practitioner, but you also wanna

make sure your client is successful.

357

that's where you have to be smart about it

358

Karen Doak: The thing I keep coming back

to from this conversation is that making

359

the work faster does not eliminate the

need to decide what the work is for.

360

So AI can help us research, organize,

generate, compare, pressure test.

361

It can give us more options and

less time, but more options are not

362

the same thing as better judgment.

363

And a polished output is not

evidence that the people involved

364

ever reached an actual agreement.

365

The faster execution becomes,

the more important it is to

366

settle the human questions first.

367

Maury offered three that I think are

worth asking before any project begins.

368

Who will determine whether

this work succeeds?

369

What does the work actually

need to prove to that person?

370

And how are we going to

separate signal from noise?

371

But I think I'd add one more,

which is, what does a human

372

being still need to own here?

373

Because AI can participate in the process,

and it can extend our thinking, but

374

it cannot take responsibility for the

direction or protect the purpose of the

375

work or have , the difficult conversation

that everyone else is avoiding.

376

And in case you're wondering, AI did

eventually become useful in helping

377

me edit these episodes, but only after

I became much clearer about what I

378

needed it to preserve, what decisions

I wanted it to support, and what I

379

was not willing to delegate, which,

as I'm saying all of that, sounds

380

suspiciously like a solid brief.

381

So I think the larger lesson across

both episodes is the brief is the

382

work, because reaching clarity

together is the work, and everything

383

that happens afterward is execution.

384

I'm incredibly grateful to Maury for this

conversation and for every conversation

385

we ever had, , and for letting all of you

listen to the way we apparently talk when

386

someone remembers to turn on a microphone.

387

It's not quite the same as the old

days when we were just one office

388

phone extension apart, but those

calls, , did involve significantly

389

fewer audio settings and way less

reliance on me for technology.

390

I'm Karen Doak, and this is Okay,

Actually, a show about what's really going

391

on at work and what we can do about it.