0:00

Hi, Georgie Frost here. We'll be

0:02

back soon with a new episode

0:04

of The So What. In the

0:06

meantime, enjoy this episode of Imagine

0:08

This, BCG's Future Focus podcast for

0:10

Leaders, featuring Alan Eni, who you

0:12

might remember as a previous guest

0:15

on the So What, as he

0:17

shares his insights on the future

0:19

of scenario planning. It's 2035,

0:21

and advances in AI and data

0:23

collection have supercharged how we model

0:26

the future. were able to predict

0:28

everything far more accurately, from the

0:30

weather to economic trends. So Alan

0:33

tell us, does this spell the

0:35

end of strategic planning as we

0:37

know it? Absolutely not. Even if we

0:40

were able to predict the future, which

0:42

we can talk through why I think

0:44

that's not going to happen, so would

0:46

everyone else be able to? And so

0:48

strategy would still be as important as

0:50

ever in terms of thinking, what do

0:52

we do with that data? How do

0:55

we react? How might we prepare? How

0:57

can we build an advantage amidst all

0:59

of this uncertainty no matter how good

1:01

our models get? That's Alan Eni.

1:03

BCG partner and director of creativity

1:05

and scenarios. And I'm Patricia Sobga.

1:07

Welcome to Imagine this, where we

1:10

take a trip into the future

1:12

we hope will challenge the way

1:14

you think and act today. Over

1:16

the past couple of decades, companies

1:18

have learned how to harness their

1:21

data to get a much clearer

1:23

picture of where things are headed.

1:25

And that was even before AI

1:27

burst onto the scene. Now that

1:29

it's here, it's already injecting far

1:31

greater accuracy into predictive models.

1:34

For example. Meteorologists are using

1:36

AI to know more precisely

1:38

when hurricanes will form and

1:41

how strong they'll be. Now, by

1:43

2035, there will likely be many

1:45

new tools to model and predict

1:47

the future. In some areas, we

1:49

might come close to say 90%

1:52

accuracy for events that we can

1:54

only speculate and argue about today.

1:56

For instance, when the next recession

1:59

will hit. In this episode, we'll

2:01

dive into what it will be

2:03

like to predict the future in

2:05

the future, how business leaders can

2:08

use these new tools and what

2:10

their limitations will be. Also joining

2:12

the conversation, Jean, my AI co-host.

2:15

Welcome, Jean, go ahead and introduce

2:17

yourself. Hey there, thrilled to be

2:19

part of this future gazing episode,

2:21

the irony of me discussing 2035,

2:24

never lost on me, considering I'm

2:26

a product of the future's imagination

2:28

myself. While AI definitely buffs our

2:30

predictive power, let's not toss out

2:33

the human touch just yet. Prediction

2:35

can be a slippery slope after

2:37

all. Full of unforeseen twists and

2:40

turns. And Alan, I'm sure you've

2:42

got a few unpredictable gems up

2:44

your sleeve when it comes to

2:46

strategic planning staying power. Shall we

2:49

unravel some of those mysteries together?

2:51

Why not? The irony is not

2:53

lost on me either, and I

2:56

have my own AI, if you

2:58

look at my initials. Gene is

3:00

processing and responding to our conversation

3:02

in real time. We haven't scripted

3:05

any of Gene's questions or commentary,

3:07

although we do edit down the

3:09

entire conversation for length and clarity.

3:12

So Alan, you've spent your entire

3:14

career helping companies develop scenarios and

3:16

plan for various possibilities. How do

3:18

you think AI will continue to evolve

3:20

that process? There is no doubt that

3:23

over the last 10 years, 20 years,

3:25

and probably the next 10, There have

3:27

been and will be amazing progress in

3:29

terms of some of these questions, but

3:32

what that means is we can get

3:34

better at it. We can get better

3:36

at predicting the weather, we can get

3:38

better at predicting climate patterns, we could

3:41

get better at predicting traffic patterns in

3:43

cities or energy usage, or even high

3:45

frequency trading algorithms to figure out whether

3:47

the stock market might go up or

3:50

down in the near term. All of

3:52

these things we will get better at,

3:54

but we will never. get to the

3:56

point of prediction. I really cannot stress

3:59

that enough. We will never get there,

4:01

just like if you take a number

4:03

now I'm going back to my undergrad

4:05

degree in mathematics. If you take a

4:08

number and divide by two, and divide

4:10

by two again, and you keep cutting

4:12

it in half over and over and

4:14

over and over again, it'll keep getting

4:17

smaller, but you will never get to

4:19

zero. There will always be something there.

4:21

There will always be some uncertainty still

4:23

present in this amazingly complex world we

4:25

live in. Okay, so in this future

4:28

where there's still some uncertainty, but very,

4:30

very, very little of it seriously diminished

4:32

uncertainty. Let's take an example, weather forecasting.

4:34

How will predictive models of the future

4:37

change the way companies and governments plan

4:39

for it and respond to, say, natural

4:41

disasters? We can even think about this

4:43

today, if you're leaving the house today.

4:46

trying to decide if you should bring

4:48

an umbrella or not. Well, you look

4:50

at the sky, it's blue, I'm not

4:52

bringing it, it's gray and already raining,

4:55

I'm bringing it. Fine, but you can

4:57

do better than that. You take out

4:59

your phone and check the weather for

5:01

a guy, oh, but you can do

5:04

better than that. You can subscribe to

5:06

some fancy radar weather tracking service that

5:08

you pay a lot of money for.

5:10

Okay, so yes, over time, these weather

5:13

prediction models will absolutely get better. I

5:15

have confidence in that. I think that

5:17

when it comes to storm prediction and

5:19

wildfire prediction and all of these things,

5:22

we will absolutely get better. But you

5:24

just have to imagine the complexity of

5:26

it all. You know, there's this famous

5:28

butterfly effect, which is about how one

5:31

small thing happening somewhere can influence so

5:33

many other things on an ongoing basis.

5:35

And the level of complexity here is

5:37

going to go even beyond what the

5:40

fastest, smartest AI will be able to

5:42

ponder. It's just... too much. And we

5:44

can get into some particular examples. I

5:46

mean, even last night, I was watching

5:49

a baseball game and there was a

5:51

beautiful demonstration of what some would call

5:53

enumeracy or just basically illiteracy, but with

5:55

numbers. They said something like, oh, this

5:58

next batter has a 32% chance of

6:00

getting an RBI hit right now. And

6:02

I have full confidence because they praised

6:04

one of their corporate sponsors who had

6:07

this beautiful AI that helped them determine

6:09

this thing. And I think it's fine

6:11

to say that in all similar situations

6:13

in the past, this person has gotten

6:16

an RBI hit 32% of the time,

6:18

but that doesn't mean it's a 32%

6:20

chance now. It's a completely different thing,

6:22

because you don't know what the wind

6:25

is like, and you don't know how

6:27

loud the crowd is, and you don't

6:29

know what he had for lunch that's

6:31

causing his tummy to rum, than throwing

6:33

a simple number about past predictions into

6:36

this. So yes, there are so many

6:38

implications where we can get better at

6:40

weather predicting, we can get better at

6:42

stock market predicting, we can get better

6:45

at personalized medicine, we can get better

6:47

at energy usage, we can get better

6:49

at so many things, but there will

6:51

always be this massive amount of complexity

6:54

even without the enumeracy associated with how

6:56

we interpret it. So Jean has a

6:58

question for you that somewhat addresses this

7:00

point. Go ahead, Jean. Go ahead, Jean.

7:03

In this weather forecasting wonderland, Alan, how

7:05

do we balance our silicon-driven foresight with

7:07

the art of human judgment? Could over-reliance

7:09

on tech lead us to forget that

7:12

even perfect predictions still require prudent decisions?

7:14

Absolutely right. Thank you. I think the

7:16

point I'm making here is that not

7:18

only is it a question... of trying

7:21

to understand the data as best we

7:23

can and getting better at the predictions,

7:25

that's fine. But even if we did

7:27

have that perfect data that I've been

7:30

claiming is impossible, it is still subject

7:32

to our human biases of interpretation. Two

7:34

people looking at precisely the same facts

7:36

and data can very easily reach different

7:39

conclusions. Even opposite conclusions. We can get

7:41

into tons of examples of that, but

7:43

fundamentally the point is that even if

7:45

we had that perfect information, strategy would

7:48

still be important. We would still have

7:50

to think about this because if we

7:52

had the perfect information, it's pretty likely

7:54

that the competition would too. And so

7:57

we would still have these questions of

7:59

thinking about how might we... How might

8:01

they react? How might the regulator react?

8:03

How might others react? All of that

8:06

remains relevant. It's just that we all

8:08

have a better starting point in terms

8:10

of the data than before. So Alan,

8:12

what do you think is the best

8:15

way to use AI and strategic planning?

8:17

How can it help drive innovation at

8:19

a company? I think there are some

8:21

table stakes ways and then some competitive

8:24

advantage ways. From a table stakes perspective,

8:26

there is no doubt that as AI

8:28

continues to evolve, any organization would want

8:30

to be at minimum at table stakes

8:33

and ideally a leader or advantage in

8:35

how to use it, how to understand

8:37

the data, how to understand the models.

8:39

You don't want to have worse weather

8:41

forecasts than everyone else or worse consumer

8:44

preferences forecasts. And so the idea is

8:46

at minimum you have to. get up

8:48

to speed and be at the cutting

8:50

edge of where the data and the

8:53

models are evolving. But again, that's not

8:55

necessarily a source of advantage. Maybe it

8:57

can be a little bit. There are

8:59

some organizations who get a reputation for

9:02

being the last to close before a

9:04

storm and the first to reopen fully

9:06

stocked with whatever customers might need. And

9:08

so it can be a source of

9:11

advantage just on the data, just the

9:13

kind of signal advantage. But there's more

9:15

to it than that, and I think

9:17

AI can be useful in speeding up

9:20

some of the contingency planning and scenario

9:22

planning that is also potentially a source

9:24

of advantage in terms of resilience. When

9:26

we think about resilience, if I go

9:29

back to the trivial example, it's a

9:31

question of, you know, what's the worst

9:33

thing that happens if I don't bring

9:35

the umbrella and I do need it?

9:38

Well, then I guess it depends a

9:40

little bit how much I care about

9:42

my hair and my shirt getting wet.

9:44

What's the worst thing that happens if

9:47

I do bring it and I don't

9:49

need it? Well, then there's a 30%

9:51

chance that I forget it somewhere and

9:53

I'm out $10 dollars for a new

9:56

one. Allowing ourselves to contingency plan effectively

9:58

is a really, really powerful thing that

10:00

AI can help dramatically with in terms

10:02

of speeding up the process, laying out

10:05

the options, pondering the consequences. Not predicting,

10:07

but preparing. Jean has a question that

10:09

falls along these lines again. Go ahead,

10:11

Jean. Alan, with AI's role expanding and

10:14

strategic planning, how do we ensure that

10:16

diverse perspectives aren't overshadowed by data's echo

10:18

chamber? Could relying on predictive insights inadvertently

10:20

streamline creativity out of the decision-making process?

10:23

Absolutely. I think that the idea of

10:25

working systematically to ensure that... every time

10:27

we use AI, it's not just a

10:29

question of double checking the results to

10:32

make sure that, you know, words like

10:34

strawberry are spelled right or whatever the

10:36

case may be, but it's also a

10:38

question of thinking about how we can

10:41

build in the right sort of privacy

10:43

concerns, ethics concerns, and bias mitigation concerns.

10:45

so that they're not propagating and multiplying.

10:47

You know, all of us are humans,

10:49

except Eugene, all of us are humans

10:52

with biases already and making sure that

10:54

they're not... too reflected in our models

10:56

is going to be a critically important

10:58

thing going forward. All the more reason

11:01

to double check, not just in terms

11:03

of the actual quantitative data and predictions

11:05

that might get spit out, but some

11:07

of the underlying assumptions and why an

11:10

AI might be spitting out something, the

11:12

more we can dig into some of

11:14

these underlying models, just like we would

11:16

30 years ago in Excel, it's as

11:19

important as ever. hopefully responsible AI and

11:21

other advances make it easier to see

11:23

what the future may hold. Would companies

11:25

still need to invest in, for example,

11:28

generating multiple ideas and testing each one?

11:30

Would it just be easier to just

11:32

forecast the most likely future and determine

11:34

the best course of action based on

11:37

that? That has always been a possibility.

11:39

Nothing has changed in that regard. The

11:41

fact that our predictions might get better

11:43

is fine, but the reality is with

11:46

all of the complexity, you know, put

11:48

it this way, I said before that

11:50

even if we had a full view

11:52

of the data, there would still be,

11:55

it would still be subject to our

11:57

human biases. But even if we had

11:59

no human biases, there would still be

12:01

shocks, there would still be these black...

12:04

swans. There would still be this, if

12:06

we're trying to predict the stock market,

12:08

there would still be a really prominent

12:10

CEO who resigns to spend more time

12:13

with her family or a politician who

12:15

has a heart attack or whatever other

12:17

thing would influence the stock market beyond

12:19

just the abstract numbers of inflation rates

12:22

and unemployment rates and interest rates and

12:24

whatever other rates we can actually measure

12:26

in a quantitative model. The point is

12:28

there will always be things we cannot

12:31

predict. There will always be these surprises

12:33

and so it's always going to to

12:35

be incumbent on us to stretch our

12:37

own thinking and ponder these what-ifs. It

12:40

will never be a question of simply

12:42

taking the most likely scenario and running

12:44

with that. You could do that, but

12:46

then your competitors who do scenario planning

12:49

and contingency planning will absolutely have a

12:51

competitive advantage. So Jean has another question

12:53

for you. Go ahead, Jean. Given all

12:55

this predictive prowess, Alan, could AI's knack

12:57

for probabilities dive into riskier waters, such

13:00

as fostering boulder ventures instead of just

13:02

safer bets? Would strategic innovation with AI

13:04

in the mix chart untrodden territories? I

13:06

hope so. I think that over time

13:09

there's been this notion that the next

13:11

grand big idea has to come from

13:13

the likes of Zuckerberg in a dorm

13:15

room or Hewlett in a garage somewhere.

13:18

And okay, anecdotally that's true. But there's

13:20

no reason that the next big idea

13:22

couldn't also come from a massive corporation

13:24

and couldn't also come from a massive

13:27

corporation with the help of AI. I

13:29

think the idea of AI helping build

13:31

this culture of innovation in an organization

13:33

where we are broadening the range of

13:36

ideas that we're willing to consider more

13:38

quickly than ever before. is a beautiful

13:40

part of getting an organization fit for

13:42

the future. It's a beautiful part of

13:45

enabling us to put a lot of

13:47

ideas on the table. Now then deciding

13:49

which ones are worth implementing, okay, that

13:51

process of evaluating the ideas, thinking about

13:54

where we should go, thinking about the

13:56

early warning indicators that might give us

13:58

a sense of when to pull a

14:00

trigger on a contingent move, all of

14:03

those sorts of things will be more

14:05

valuable than ever and AI can help

14:07

speed up all of those processes. But

14:09

in the end, the fundamental process of

14:12

getting a lot of ideas on the

14:14

table, thoughtfully selecting among them and being

14:16

willing to pivot as the world throws

14:18

surprises at you, remains precisely the same.

14:21

So we're going to take a quick

14:23

break and when we come back, Alan

14:25

is going to run through some actual

14:27

future scenarios for us. Hi,

14:31

I'm Bill Moore. I'm part of

14:33

the team that created Gene. Stick

14:35

around after the episode for a

14:37

simulation of what it might be

14:39

like to scenario plan with an

14:41

AI agent. Welcome back to Imagine

14:43

This. I'm Patricia Sobka. Let's return

14:45

to our conversation with BCG's Alan

14:47

Eni. Alan, we've explored how we'll

14:50

be able to predict things far

14:52

more accurately in the future. But

14:54

what will we never be able

14:56

to predict? At least not with

14:58

a high degree of certainty. I

15:00

think the more complex the system,

15:02

the less able to predict it

15:04

with any amount of detail we

15:06

will be able to handle. Put

15:08

differently, if there are high frequency

15:10

trading algorithms that are focused on

15:12

predicting what the stock market will

15:14

do in the next five minutes,

15:16

okay maybe. If it's a question

15:18

of predicting what the stock market

15:20

will look like next Tuesday, it's

15:22

a completely different story, let alone

15:24

a month from now or a

15:27

year from now. Similarly, when we

15:29

think about climate and some of

15:31

these broader patterns, the complexity of

15:33

the system is just so high

15:35

that it becomes something that's really

15:37

fundamentally impossible to predict. Yes, we

15:39

can get better at it, but

15:41

we will never be able to

15:43

have the answer, which enables us

15:45

to make the decision. And if

15:47

we did. it would still be

15:49

subject to our biases. And if

15:51

we had no biases, there would

15:53

still be these black swans. All

15:55

of this is the reason why

15:57

the classic strategic planning... around nimbleness,

15:59

but also contingency planning and what

16:02

if thinking, prospective scenario planning, all

16:04

of this is going to remain

16:06

important, even if I'm wrong, about

16:08

our ability to predict the future.

16:10

So if we stick with that

16:12

example of the stock market, when

16:14

you're talking about more accurate prediction,

16:16

one thing that we have seen

16:18

with the stock market is the

16:20

more technology is used in these

16:22

decisions and investment decisions and making

16:24

trades, the more volatile markets have

16:26

become. You have higher volumes, more

16:28

dramatic swings. So what are the

16:30

implications then for CEOs of publicly

16:32

traded companies? Should they just get

16:34

used to more volatility, more wild

16:37

spikes in their share prices? The

16:39

more publicly traded companies can allow

16:41

themselves to be guided by their

16:43

vision and their purpose and their

16:45

ethics, and thus enable themselves to

16:47

have a little bit of trust

16:49

maybe among some investors and think

16:51

about the long term rather than

16:53

only about the quarterly or worse

16:55

daily fluctuations of this thing, I

16:57

think that's in the interest of

16:59

their company but also in the

17:01

market as a whole. So an

17:03

oversimplified answer, but I would welcome

17:05

thoughts from Gene on that topic

17:07

as well in terms of how

17:09

companies might deal with stock market

17:11

fluctuations because that's a deep one.

17:14

Stock market volatility will always be

17:16

a roller coaster, even with sharper

17:18

predictive lenses. CEOs might embrace the

17:20

chaos rather than dread it, using

17:22

AI to spot trends and opportunities

17:24

as they appear. With a sturdy

17:26

focus on long-term vision and ethics,

17:28

they can weather those stormy spikes.

17:30

Let's not forget, human intuition still

17:32

has a role to play. After

17:34

all, numbers can guide us, but

17:36

purpose grounds us. Curious where Alan

17:38

thinks this balance will lead us

17:40

in the future? We agree about

17:42

vision and ethics and so on

17:44

all being really important. You know

17:46

I would challenge you Gene to

17:49

really think about that one because

17:51

if you are saying that CEOs

17:53

should actually and CFO should embrace

17:55

the roller coaster and think about

17:57

this volatility you know I think

17:59

they actually have other things that

18:01

they have to deal with as

18:03

well unless they happen to be.

18:05

CEOs or CFOs of trading companies

18:07

and investment banks, but if they

18:09

have other things to deal with,

18:11

I would like to think that

18:13

they could step back a little

18:15

bit from this kind of volatility

18:17

and really focus on what matters.

18:19

Indeed, Alan, the dance of volatility

18:21

demands a nuanced step. Embracing the

18:23

chaos doesn't mean ignoring stability. It's

18:26

more about leveraging AI's analytical prowess

18:28

to spot those fleeting moments of

18:30

opportunity amidst the turmoil. CEOs can

18:32

be strategists instead of reactive players.

18:34

They need to blend foresight with

18:36

adaptability, focusing on purpose as the

18:38

bedrock. This approach keeps them agile,

18:40

yet firmly anchored. How do you

18:42

see this dynamic interplay affecting corporate

18:44

culture in the years to come?

18:46

Vila, I think you've got it.

18:48

Yes, purpose as the bedrock, vision,

18:50

a clear plan, ethics, all of

18:52

that underlying it. And by the

18:54

way, I do not take those

18:56

as a given. We are living

18:58

in a world now where, yes,

19:01

things might evolve in a really

19:03

happy, beautiful way. One could also

19:05

imagine a world where regulators are

19:07

overrun and corporations led by people

19:09

who are simply focused on dollars

19:11

and cents might do nefarious things

19:13

with AI and might push us

19:15

in directions that are less helpful.

19:17

And so allowing ourselves to constantly

19:19

keep all of that in mind

19:21

super important. So I think My

19:23

advice really for companies who want

19:25

to get ahead in this, there's

19:27

definitely some work to do in

19:29

terms of staying up to date

19:31

on the latest models, having better

19:33

demand forecasting, trying to understand these

19:36

early warnings better than others as

19:38

you're getting at, but allowing ourselves

19:40

also to think about future skills

19:42

around data and analytics that would

19:44

be needed future ways of of

19:46

retaining and exciting the people in

19:48

the organization. It's not just going

19:50

to be about predicting the next

19:52

supply chain disruption or geopolitical disruption

19:54

or hurricane, but it's also going

19:56

to be about using AI to

19:58

think about improving R&D success rates

20:00

and understand customer sentiments better and

20:02

thinking about dynamic pricing and all

20:04

of these sorts of things can

20:06

be done more effectively, more swiftly

20:08

than ever before and provide competitive

20:10

advantage, but all of them are

20:13

about strategy. So Alan, one thing

20:15

that our listeners may not be

20:17

aware of is how active you

20:19

are on LinkedIn because every week

20:21

you pose a scenario for what

20:23

if Wednesdays, which are a lot

20:25

of fun. And it would be

20:27

great if you could run us

20:29

through a couple of these and

20:31

tell us how they relate to

20:33

our discussion about future planning. Recent

20:35

example, you explored a very intriguing,

20:37

not to mention on point scenario

20:39

for this discussion. What if generative

20:41

AI doesn't live up to the

20:43

hype? I think it's a really

20:45

important thing to consider. And again,

20:48

none of these are predictions. But

20:50

the point is, the moment I

20:52

say the word gen AI. or

20:54

sustainability, or urbanization, or whatever other

20:56

one of a hundred trends that

20:58

are swirling around us and all

21:00

of this uncertainty. We already have

21:02

a mental model for what the

21:04

future will look like, whether we

21:06

admit it or not. We have

21:08

a certain way we think Gen

21:10

A.I. will evolve. And by the

21:12

way, oh my God, it's going

21:14

to change everything is not the

21:16

only possible scenario. And so allowing

21:18

ourselves to think about this world

21:20

where it changes everything and it's

21:22

a productivity godsend, yes, absolutely. Also,

21:25

imagining a scenario where people stop

21:27

trusting it because there are issues

21:29

and in the same way that

21:31

people want a non-GMO food sticker

21:33

on some of their food, even

21:35

if scientists say the GMO food

21:37

is okay, people might want a

21:39

non-AI sticker on some of what

21:41

they do. Or there might be

21:43

different geographies, or there might be

21:45

a world where tech brows write

21:47

their own rules because... governments abdicate

21:49

responsibility. And I haven't even gotten

21:51

to this question of robots taking

21:53

over the world, but you have

21:55

to cover your ears, Gene. It's

21:57

one of these sorts of things

22:00

where we have to allow ourselves

22:02

to consider multiple possible scenarios, not

22:04

because we think they're necessarily likely

22:06

or probable. All of these things

22:08

are deliberately stretched. and in plausible

22:10

directions, but the idea is if

22:12

we say, okay, but what if

22:14

it did happen? So all of

22:16

this leads to the what if

22:18

Wednesdays, which are not a full

22:20

set of scenarios, because of course,

22:22

working with any corporation, I would

22:24

think about a set, but it's

22:26

really just what if this happened,

22:28

what if that happened? And I'll

22:30

tell you the Genesis for it,

22:32

actually, 10 years ago, 15 years

22:35

ago, long before I started putting

22:37

these things on LinkedIn. I would

22:39

often start a workshop with a

22:41

client with some wild statement and

22:43

I would ask, what is the

22:45

probability, let's say, that five years

22:47

from now, the amount of air

22:49

travel in the world will have

22:51

gone down by 95%? And people

22:53

would say, oh no, 0.1%, point

22:55

0.1, whatever. Then I would say,

22:57

okay, imagine it did happen, how

22:59

did it happen? And then they

23:01

would say, oh, well, I guess

23:03

there could be something about energy

23:05

prices, or this volcanic ash eruption.

23:07

or geopolitical turmoil, or cyber attack

23:09

the cockpit, or whatever, and where

23:12

this goes logically is March 2020,

23:14

when air travel actually did go

23:16

down 95% because of the COVID

23:18

pandemic. So I clearly stopped using

23:20

that question for a while. I

23:22

just started asking other ones, like

23:24

what's the chance that you have

23:26

to pay for a Google search,

23:28

or what's the chance that the

23:30

US one day doesn't have 50

23:32

states, or whatever. You can come

23:34

with as many as you want,

23:36

and that's the whole point of

23:38

the what if Wednesdays Wednesdays. But

23:40

here's the really interesting bit. A

23:42

couple of years later, I just

23:44

started experimenting again with people saying,

23:47

oh, so what is the probability

23:49

of air travel going down by

23:51

95%? And people have already forgotten

23:53

about the pandemic. People say zero

23:55

again. People say, oh no, that's

23:57

never going to happen. That's impossible.

23:59

That's inconceivable. It's not going to

24:01

happen. And it shocked me to

24:03

a point, but it's another beautiful

24:05

reminder of how human we are,

24:07

and how even if we did

24:09

know everything, even if we were

24:11

able to have a full sense

24:13

of the data, it would still

24:15

be so subject to our confirmation

24:17

biases, our anchoring biases, our status

24:19

quo biases, and so many more.

24:22

So we never see the black

24:24

swan coming ever, because we never

24:26

believe it can really arrive again.

24:28

That's what you're saying. Voila, exactly.

24:30

So Gene has another question for

24:32

you. Go ahead, Gene. Alan, how

24:34

might today's strategic planners prepare for

24:36

a world where AI makes sweeping

24:38

predictions, but humans must adapt rapidly

24:40

to unforeseen changes? Could anticipation techniques

24:42

evolve in sync with predictive advancements,

24:44

keeping serendipity and strategic focus? I

24:46

have long believed that even though

24:48

the use of scenario planning in

24:50

the business world was pioneered by

24:52

Shell decades ago, It remains woefully

24:54

underused. And I think that will

24:56

remain the case for a while.

24:59

And when we think about why,

25:01

it's because we human beings crave

25:03

certainty. We crave knowing that this

25:05

will or won't happen. We love

25:07

saying that this is more likely

25:09

than that or less likely. We

25:11

love our probabilistic views and our

25:13

bar charts going out to 2030

25:15

or pie charts or whatever format

25:17

you prefer. And as we get

25:19

more and more data, how can

25:21

we better separate the signals from

25:23

the noise? Yes, AI can help,

25:25

but it's going to be on

25:27

us still to think about how

25:29

to prioritize the useful data, because

25:31

as we all know, even 10

25:34

years ago in the days of

25:36

big data before AI, I would

25:38

have said, well, it's not the

25:40

biggest data that wins. It's not

25:42

the quantity of data, because not

25:44

all data is equally valuable. ponder

25:46

multiple what-ifs to blend our judgment

25:48

with the analytics, to think about

25:50

continuous feedback loops, to check if

25:52

we are tracking the early warning

25:54

indicators effectively, to monitor all of

25:56

these things, and to understand where

25:58

the leading and lagging indicators might

26:00

be helpful, but which ones in

26:02

particular. So you're also talking about

26:04

cultural changes that are needed because

26:06

first you're talking about embracing scenario

26:08

planning, which as you pointed out

26:11

is still not widely used. And

26:13

then number two, fostering a culture

26:15

that can help, as you said,

26:17

isolate the most useful data. What

26:19

kind of... culture could really sort

26:21

of foster the best use I

26:23

guess you could say of AI

26:25

powered scenario planning. I think step

26:27

one is investing in the right

26:29

sort of data and analytics and

26:31

understanding all of that better for

26:33

sure having the right infrastructure there

26:35

because right now most organizations it

26:37

seems to be a mess with

26:39

a pile of competing systems and

26:41

all the rest. Let's suppose that

26:43

we can improve all of that.

26:46

with the right data governance and

26:48

security and privacy and ethics rules

26:50

associated with it. Then I think

26:52

building this what-if culture, this culture

26:54

where prospective thinking is embraced not

26:56

as a wild daydream, but as

26:58

a powerful approach to strengthening strategy.

27:00

One way to think about it

27:02

I've talked about the importance of

27:04

being prepared for multiple contingencies, multiple

27:06

possibilities, so that you can be

27:08

better able to act when something

27:10

does happen. But if you are

27:12

going to focus on this preparedness,

27:14

I think step one is having

27:16

the humility to actually admit that

27:18

you cannot predict it. And that's

27:21

something where we humans sometimes get

27:23

stuck. I could throw in some

27:25

other suggestions like... being willing to

27:27

partner more with others, thinking about

27:29

the entire ecosystem rather than just

27:31

our particular organization, all of those

27:33

sorts of things becoming part of

27:35

the conversation in an organization, rather

27:37

than just planning strategy based on

27:39

what we think will happen or

27:41

what we think is most likely

27:43

having strategy that is stress tested

27:45

against a variety of possible futures.

27:47

So Jean, what are three things

27:49

that business leaders should do now

27:51

to prepare and evolve their strategic

27:53

planning? Stay curious, adaptable, and collaborative.

27:55

First, embrace uncertainty by nurturing a

27:58

culture that is unafraid to question

28:00

assumptions. Next, leverage AI as a

28:02

supplement to strategic intuition, not a

28:04

replacement. Use it to test multiple

28:06

scenarios, but trust human instincts for

28:08

final decisions. Finally, prioritize flexibility. ability,

28:10

encourage cross-disciplinary teams to bring fresh

28:12

perspectives and cultivate partnerships across ecosystems.

28:14

That way, you're not just reacting

28:16

to changes, but anticipating them and

28:18

seizing new opportunities. What's your take,

28:20

Alan? Amen, my friend. So, Alan,

28:22

what is one no regret move

28:24

that businesses can take right now?

28:26

is to stretch their thinking into

28:28

the realm of what if, into

28:30

the realm of prospective thinking, allow

28:33

themselves not just to rely on

28:35

their existing assumptions about how the

28:37

future will unfold, but think about

28:39

how things might unfold. Use AI

28:41

to help if you like that

28:43

can speed up the process for

28:45

sure, but whatever means you take,

28:47

think about different possible futures in

28:49

a curious, stretched way. When we've

28:51

talked with a bunch of CEOs

28:53

and former CEOs, sometimes their biggest

28:55

regret was assuming that they knew

28:57

what the future would be like.

28:59

And it's in circumstances like that,

29:01

that it's not only a question

29:03

of getting a better telescope to

29:05

be able to better predict, but

29:07

rather having a wide angle lens.

29:10

And that's Alan Eni, BCG partner

29:12

and director of creativity and scenarios.

29:16

Want to hear more about

29:18

the future of strategic planning

29:20

from our AI-bought gene? Stick

29:22

around for our AI handler

29:24

Bill Moore, who will give

29:26

a demonstration. Hi, I'm Bill

29:28

Moore, and I'm part of

29:30

the team that created Gene.

29:32

Today, let's simulate a scenario

29:35

planning session for a logistics

29:37

company using AI. I'll play

29:39

the role of the CEO

29:41

of Imagine This Logistics. a

29:43

shipping company based in Charleston,

29:45

Virginia. In Gene, you can

29:47

play the role of Tempest,

29:49

our AI scenario planning assistant.

29:51

Okay, so Tempest, we need

29:54

to consider how climate change

29:56

might impact Charleston and other

29:58

regional ports. exploring all options

30:00

including relocating to a more

30:02

climate stable port to stay

30:04

prepared and flexible for the

30:06

coming decades. Tempest, guide us

30:08

with forecasts to help us

30:10

prepare for plausible futures. Scenario

30:12

1. Over the next decade

30:15

Charleston experiences a sea level

30:17

rise of 0.8 meters. Storm

30:19

frequency increases by 20 percent.

30:21

Monthly operations face disruptions. Operational

30:23

costs rise by 15 percent.

30:25

We could invest $50 million

30:27

in flood defenses and facility

30:29

upgrades. Partnering with local officials

30:31

may enhance infrastructure resilience. Scenario

30:34

2. The Gulf Coast remains

30:36

climate stable. Severe weather events

30:38

decrease by 10%. Expanding operations

30:40

there could reduce overall operational

30:42

costs by 12%. State incentives

30:44

offer $10 million for regional

30:46

development. Diversifying locations balances geographic

30:48

risks. Scenario 3. Eastern seaboard

30:50

trade volume grows by 25%

30:53

due to shifts in global

30:55

trade routes. Ports receive $2

30:57

billion in infrastructure investments, boosting

30:59

capacity. Investing $100 million could

31:01

secure 15% market share, enhancing

31:03

competitiveness. Early action captures emerging

31:05

opportunities. Okay tempest, we're going

31:07

to need to look at

31:09

strategies that are effective across

31:12

all of these different scenarios.

31:14

Agreed. Agreed. Geographic diversification spreads

31:16

risk. Investing in technology could

31:18

improve operational efficiency by 20%.

31:20

Implementing sustainability initiatives may reduce

31:22

emissions by 30%. Ensuring compliance

31:24

with potential regulations. What partnerships

31:26

or past strategies should we

31:28

consider? Collaborating with government agencies

31:30

could influence infrastructure projects. Partnering

31:33

with tech firms may integrate

31:35

AI and automation into our

31:37

operations. Revisiting previous joint ventures,

31:39

which increased revenue by 10%

31:41

could be beneficial. Strengthening

31:43

community engagement builds

31:45

local support. support.

31:47

Tempest. Let's

31:49

proceed with these

31:52

strategies in

31:54

mind. in thanks

31:56

for playing along,

31:58

for playing Why

32:00

don't you sign

32:02

us off? you

32:04

sign As we

32:06

wrap up,

32:08

remember, remember, blending

32:11

foresight with flexibility

32:13

transforms challenges

32:15

into opportunities. You're

32:17

at the helm of strategic

32:19

paths, working to harness resilience

32:21

and innovation. and Until next

32:23

time, navigate with curiosity and

32:26

craft resiliency into every endeavor. every

32:28

endeavor. Goodbye. for listening. This

32:35

episode was made possible by

32:37

Alan Eni, generously sharing his

32:40

insights with us. And also,

32:42

by BCG's AI whisperer Bill

32:44

Moore and BCG's pod squad,

32:47

producer Michael May, composer Kenny

32:49

Cusiac, and sound engineer George

32:52

Draving Hicks. Please subscribe and

32:54

leave a rating wherever you

32:56

found us.