June 18, 2025 • 21 mins
How the Smartest Directors Are Using AI
In this episode, we speak with Jamie Green, Co-Founder and CEO of Tutaki, the Director Workbench designed to make board directors radically more effective. Drawing on his experience at McKinsey and extensive conversations with Chairs, directors, and governance leaders, Jamie built Tutaki to help transform how board members prepare, engage, and act. With a strong foundation in strategy, governance, and AI product design, Jamie is at the forefront of a new era of “augmented directorship.” As complexity grows and time shrinks, AI is proving to be a performance accelerator, not just a support tool.
In this episode of the Better Boards podcast, we speak with Jamie Green, Co-Founder and CEO of Tutaki, the Director Workbench designed to make board directors radically more effective. Drawing on his experience at McKinsey and hundreds of conversations with Chairs, directors, and governance leaders, Jamie built Tutaki to transform how board members prepare, engage, and act. With a background in strategy, governance, and AI product design, he’s leading a new era of “augmented directorship.” In this conversation, Jamie outlines how the best boards are using AI not just to cope—but to lead.
Why effectiveness matters
The most valuable strategic decisions happen in the boardroom. AI can dramatically enhance this by helping directors focus on what matters, enabling more trust and transparency with management.
“Directors are the amplifiers, and it’s really key that their whole process works effectively to actually get the best out of it.”
What’s holding boards back
The main blockers? Lack of clarity on use, data privacy fears, and over-reliance. Jamie argues that these risks can be mitigated—and that the real risk is not using AI at all.
“If you blanket rule, ‘you can’t use AI,’ and your competitors do, they’ll make faster, more accurate decisions—and leave you behind.”
Where AI is heading
Jamie outlines four phases of AI maturity in the boardroom: analysis, automation, AI-assisted directors, and eventually AI boards. While AI boards are still a few years off, the steps to get there are already unfolding.
“Imagine if you had Warren Buffett sitting there analysing the financials. That’s the power of AI.”
Start today
Directors are already using AI to synthesise documents, explore competitor strategy, and broaden their thinking. The key is starting with clarity—ask AI to help you ask better questions.
Three takeaways:
“Not using AI is the riskiest option… The best directors are going to be AI-literate.”
- Not using AI is the greatest risk.
- The best boards are already using it.
- Seek out vertical tools that solve real problems.
Discover how you can take the first confident steps toward AI-powered boardroom effectiveness—starting today.
Remember to subscribe and never miss an episode of the Better Boards Podcast Series. It’s available on Apple, Spotify, or Google.
To find out how you can participate in the Better Boards Podcast Series or for more information on Better Boards’ solutions, please email us at info@better-boards.com.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_01 (00:01):
how the smartest directors are using AI.
Welcome to the Better Boardspodcast series, the podcast for
chairs, CEOs, non-executivedirectors, company secretaries,
and their advisors.
Every episode is filled withpractical insights and learnings
from those inside boardrooms.
We discuss what really mattersand highlight actionable steps
(00:25):
you can take to enhance theperformance of your board.
Today's episode It's for everyboard director wondering how to
stay ahead as complexity growsand time shrinks.
We are joined by Jamie Green.
Jamie is co-founder and CEO ofTutaki, an AI platform built
(00:45):
specifically for boarddirectors.
Jamie brings a bold perspectivethat AI isn't just a support
tool, it's a performanceaccelerator.
In this episode, we explore howforward-looking directors are
using AI to enhance theireffectiveness, how governance
workflows are evolving and whatdirectors can do today to stay
(01:06):
at the forefront and at the topof their game.
Jamie, thank you so, so much forcontributing to the Better
Boards podcast series.
SPEAKER_00 (01:14):
No, no, thank you.
Thank you for having
SPEAKER_01 (01:16):
me.
All right.
We love to have people fromaround the world.
So I sit currently in London.
Maybe tell our listeners whereyou are sitting.
SPEAKER_00 (01:24):
Yeah, I'm currently, if anyone's been to the other
side of the planet in NewZealand, I'm in Little Old New
Zealand and even up north in NewZealand, so a small town called
Whangarei.
So yeah, pleasure to be here.
SPEAKER_01 (01:37):
Fantastic.
It's important for the seriesbecause we really want to
capture the perspective ofdifferent people from different
countries, different roles.
So a very warm welcome from NewZealand.
Let's jump straight in.
Why is effectiveness, and welove effectiveness, Yeah, I
mean, it's a great
SPEAKER_00 (02:09):
first question.
You think about the purpose ofcompanies overall is to create
value.
And arguably the place where alot of that value is created is
in the board.
I mean, the board's making themost strategic decisions that
exist operationally and justgenerally.
strategically for that business.
So effectiveness overall at theboard level is massively
(02:30):
important.
And companies that do it well,move faster, make more accurate
decisions, better decisions.
And the companies that don't,the board can feel like a drag.
Management feel like they'repreparing for weeks for a
learning education session.
And not a lot can come out ofit.
There's a lot of trust.
So an effective system is onewhere directors, they focus on
(02:50):
the right things, They're askingthe right questions and
generally not asking managementto make oodles and oodles of
models for them.
And secretaries are streamliningthat process, making sure the
right information is flowing tothe right people.
And the executive team has fulltrust in their board.
So they're not trying to hideinformation.
They're actually sharing theright information at the right
level to make not operational,but strategic decisions.
(03:13):
Often you see 200-page documentsand a board director looks at it
and goes, what am I supposed todecide here?
So directors are the amplifiersand it's really key that that
whole process works effectivelyto actually get the best out of
SPEAKER_01 (03:26):
it.
I mean, the whole community isat the moment really excited
about AI or you see so manypapers and documents and also
podcasts on AI.
Where in your view are mostboards really stuck right now
and what's holding them backfrom really harnessing AI more
fully?
SPEAKER_00 (03:45):
Yeah, it's a great question.
It's often the board themselvesholding them back.
I mean, if you think about AI,it's this thing that's moving
faster than the speed of lightand it changes every day.
So as a director, it's reallydifficult to go, how can I use
this effectively?
How can I be most effective withAI?
And typically that falls intothree categories.
It's what can I do with it?
(04:07):
How can I use it?
And how does this apply to myorganization?
Now, the first two can be kindof trained.
You can just use ChatGPT, ask acouple of questions, get a rough
idea.
But the third one of how thiscan be applied to my
organization, that's where thatunlock really is.
And if a board is sayingoutright, we can't use it, well,
they're the most strategic partof the business.
(04:27):
So if they don't understand howto use it and how it applies to
the organization, then you canbest believe that organization
is not going to get the best outof it.
So those are the threechallenges.
And I think every board...
and will talk to it, shouldembrace it at least at some
level, understand it at the veryleast.
Because if they're not using it,I can almost guarantee the
graduates that are coming out ofuniversity, out of college,
(04:49):
they've been brought up on thisstuff.
And if you tell them they can'tuse it, they're going to use it
without telling you.
And if you create the rightframeworks and guardrails for
them to use it effectively andyou use it yourself, then your
business is just going tothrive.
SPEAKER_01 (05:01):
Now, at least some months ago, I think things are
shifting now.
People were very concerned aboutthe risks.
So let's talk a little bit aboutpitfalls.
What risks do boards need towatch for when adopting AI?
Let's look at it from atechnical side and also from a
cultural side.
SPEAKER_00 (05:19):
Yeah, yeah.
It's one of those things that'smoving so quickly.
We're still actually exploringand finding out, you know, some
of these pitfalls.
But, you know, I've talked withmany a director, and the three
that come to mind are this feararound outputs.
Is it hallucinating?
Are they the wrong thing?
So that's one, and I'll talk tothat.
(05:39):
And the second is, where's mydata going?
I'm uploading, I'm askingquestions, I'm dragging and
dropping documents.
Where is it all going?
And then the third one is, Ithink the most prevalent, is
people are getting over-reliant.
So on outputs being wrong, Imean, if you ask a very broad
question on a broad topic with abroad data set.
AI can still hallucinate things.
(06:01):
It's not infallible.
There are things that can bewrong.
So you need to have a keen eyeon the question and what the
output is.
So that's one.
The data can be shared piece, Ithink, Easily solvable, but
still this fear.
I mean, AI is a black box.
Not even the researchers atAnthropic or OpenAI know
sometimes what it's doing.
So there is that fear aroundwhere it's going.
(06:22):
Again, it can be mitigated.
And then the third one is thisover-reliance.
I mean, we're seeing this atschools today.
Teachers are saying, you know,students coming through schools
are losing that logical part oftheir brain.
And that's no different fordirectors that become
over-reliant on this thing.
So the best directors really areusing it to supplement their own
insight.
They're using their own contextand just making faster, better
(06:42):
decisions.
I mean, they're not using it toreplace themselves.
So it's all about getting thatright data, asking the right
question, and having the rightguardrails in place to make sure
that your privacy is in place,the data is not going anywhere,
and that you're getting the bestout of it for your organization.
SPEAKER_01 (06:58):
You know, in our prep call, you talked about
security aspects.
And yes, it is also from what wecan see a concern for every
board.
But you've said it's riskier notto be using AI than actually
start using it.
Can you explain it a bit?
SPEAKER_00 (07:17):
Yeah, yeah, of course.
I mean, it's a bit of a cheekystatement, but I do think it's
true.
So the risk I mentioned earlier,I think the one that usually
sticks with people the most iswhat's happening to my data.
I mean, you sit on a board,you're making the most strategic
decisions that a company has,and these documents are usually
pretty secretive.
So what happens if it gets out?
(07:38):
But ultimately, if you put theright guardrails in place,
structure the LLM or the AI inthe right way, it's no different
than putting your documents onOneDrive or putting them on
Google Drive.
So I think that can bemitigated.
And it can be mitigated in threeways.
One is you can turn off trainingon the models.
So ChatGPT and Anthropic Cloud,all of them in the settings will
(08:02):
have a model that says stoptraining on this data.
So that's one.
But if you don't trust that,what you can do is create a
private model.
I mean, that's what we do atTutaki.
Every organization has a privatemodel.
And then the third one is youcan actually just host it
on-premise.
So you basically host it on yourservers and you know where that
data is going because you hostit.
(08:23):
You know, that can be mitigated.
And I think if you blanket rule,hey, you can't use AI, well, if
one of your competitors does,they're going to be making
decisions way quicker andthey're going to be more
accurate decisions.
And I think the risk to youactually is just being left
behind entirely.
How you use AI is using it inthe right way that works for
your organization.
(08:43):
I don't think that right way isjust saying you can't use it.
SPEAKER_01 (08:47):
It's, of course, a journey for every board.
They start somewhere and theyhave to take steps to develop,
but there is no finaldestination, of course.
Help our listeners to walk usthrough the journey you have
seen successful boards take.
SPEAKER_00 (09:04):
Yeah, I'll talk through.
I mean, I'd say there's foursteps, and two of them are
probably where companiescurrently are, and two of them
are where I see the futurebeing.
I mean, we're only two yearsinto this technology.
Think about in the context ofthe internet.
It's moving a whole lot faster,but two years into the internet,
we still had some of the biggestcompanies were online pet
(09:26):
stores.
So it's still quite an nascentphase of where it's going to end
up being.
So some of this is obviouslyliable to change, but this is
where I think the world is.
There's four phases.
The first is help me analyzesomething.
And everyone's doing that whenthey use ChatGPT.
Ask a question.
It's doing analysis.
It's giving you an answer.
(09:47):
The second is automate somethingfor me.
And typically that's around achain of events or a chain of
tasks, what we call workflows inthat community.
And if you think about them inthe context of a board, it's
Schedule a meeting, send theagenda, take minutes, take
actions.
That is a workflow, and that canall be automated.
(10:07):
The third phase and fourthphase, I don't think we're at
yet, but I think we're going toget there.
And the third one is AIdirectors, and the fourth is AI
boards.
If you think about directors,sport, for example, has line
judges.
Some of the referees, I mean,boxing had introduced an AI
referee in one of theirheavyweight matches, I think,
(10:27):
recently.
So you're starting to see itplay out as this kind of third
or alternative source ofinformation.
And I don't think boards willreplace directors, but they
might use it as a way to testthinking, to be that kind of
tell me about something in analternative view.
Imagine if you had WarrenBuffett sitting there analysing
(10:48):
the financials.
That's kind of the power of AI,and you can actually create
those agents to give that pointof view.
And the final is AI boards, andI don't think we're going to get
there for a long time.
But similarly, for small boardsand small organizations, like as
a CEO myself, I would love tohave a faux board there that I
could ask questions to thatgives advice, that tests and
(11:08):
prods my board pack before Isend it to my chair.
That would be an amazing way toactually get feedback rapidly.
And I think we will get thereeventually.
So those are the kind of keyfour phases that I see us going
through.
And ultimately, it's the casethat you need to create those
right structures for yourorganization so everything's
being done in the right way.
But what an amazing world that'sgoing to be.
SPEAKER_01 (11:29):
What do you think?
How long will it take us to getto this phase when we really
have an AI board?
Any guess?
Nobody knows, but give it
SPEAKER_00 (11:37):
a shot.
I would say within five years.
If you did today, if you gaveall of your email context,
document context, all thecontext that you're getting on
calls, so transcriptions, etcetera, to an AI algorithm and
chunked it in the right way, itwould have all the relevant
context to make decisions.
And that is often the laggingpiece with AI is it doesn't have
(11:59):
the right context.
But if you give it the rightcontext, it's an incredibly
smart tool.
So I actually think it's notgoing to be a technical
capability that's going to stopus.
It's going to be...
compliance, adoption, and justgeneral human sentiment around
some of these things.
I can imagine startups and smallboards using this stuff a lot
faster than obviously the bigorganizations, but I'd say
(12:23):
within five years.
SPEAKER_01 (12:24):
Amazing.
Absolutely amazing.
Now let's focus more on the now,where we are now in the year
2025.
I think it's important to maybesay 2025 because some of the
people dive into Old podcast.
So if someone listens to this in2027, they may think, oh my
goodness, what are they talkingabout?
(12:45):
So old fashioned.
What are some of the ways youhave seen directors?
How are they using AI in theirboard work today?
Inspire some of our listenerswhat they can do actually today
and what you have seen is reallyworking.
SPEAKER_00 (13:00):
There's too many to count.
But yeah, I mean, one of thebenefits of my role is I get to
talk to hundreds of directorsand many a day.
I'd say there's three bigbuckets, and they're broad, and
I'll get into some exampleswithin each.
But one is saving time.
The second is expanding theirown thinking, their own
(13:21):
aperture.
And the third is managing theprocess.
And save time is the classic, doa research report on my
competitors.
You go to ChatGPT, click deepresearch, here are my
competitors, tell me everythingabout them, what they've done,
Over the last three months, giveme an update.
Save time in understanding that.
Synthesize reports, summarizethings, all of that.
(13:44):
I mean, if you're reading one200-page board packet, it can be
hard.
And if you go, hey, I think wediscussed something three months
ago, it's even harder.
You have to dig through threedifferent other reports.
But ChatGPT, Tutaki can do thatin five seconds.
So that's one.
I think the second is expandtheir thinking and research.
A director is usually someonewho's got experience
(14:06):
operationally in one domain.
You're the finance director, thelegal director, the operational
director, or you've been trainedas a professional director, but
you typically kind of have yourarea of expertise.
And that's why having a verydiverse board is important
because you cover each other'sgaps.
But I'm seeing a lot ofdirectors use AI in a way that
(14:27):
they can fill their own gaps.
So they've got different agentswith different expertise that
answer the question or askquestions in different ways.
So they've got their WarrenBuffett for financials.
They've got their person foroperations, their person for
marketing, the person for sales.
And what it does is it raisesthings or flags things that they
would have entirely missed.
(14:48):
So as a director, you actuallycome to the meeting.
more prepared because youunderstand these other topics in
ways that you want to understandthem in.
So that's the second isexpanding their thinking.
And the third is managing aprocess.
Directors are using tools thatgo through their calendar, help
them get through their emailsquicker, and just generally, as
(15:09):
a board process, manage theprocess.
Help me write agendas.
Some boards are comfortable withtranscriptions.
Take my minutes for me.
Create my documents for me.
Save two weeks of work on themanagement team side by just
aggregating information in areport for me.
So I think there's those threelevels of ways that companies
(15:30):
and boards specifically areusing AI today to save time and
be more productive.
SPEAKER_01 (15:35):
Amazing.
When I see how I started workinglife and how it's evolving now,
I mean, I'm absolutely stunned.
what's possible and where all ofthis is going.
And maybe, you know, ourlisteners are always interested
in this real practical side thatreally inspires them and they
can take and basically, yeah,apply right away.
(15:59):
Maybe leave at the end ourdirectors with some very
practical things a director cando literally this week to begin
using AI safely and alsostrategically.
SPEAKER_00 (16:13):
Yeah, and I hope that if one thing comes out of
this, I hope people take it up.
I think if your listeners startusing it, they'll just be such
more effective directorstomorrow.
The one thing I would try and dois any tool you want, if it's a
general tool, if it's aboard-specific tool, even
(16:35):
better, but the best way to knowwhat you can do with AI is to
ask AI.
If you ask it the question, Andusually the framing is you tell
AI what it is and you ask it howto think.
So you are a director of thistype of organization.
You're trying to do this.
What are five ways you can doit?
(16:57):
Give me a prompt on that.
And by asking it in that generalsense and asking it to create
the prompt, it'll then feed backa very concise, structured
answer for how you can then askit that question.
And I think that's the biggestchallenge with AI today is you
have to be this prompt expert.
I mean, it's called promptengineering in the community.
(17:18):
I mean, we've done thousands ofiterations to get it to the
right output.
And for your director thatdoesn't have the time to be
learning this stuff, why not askthe source what it wants to be
told?
So if you frame it like that, itwill usually give back a really
powerful prompt that you canthen simply ask it to run.
(17:41):
And the output is honestlymind-blowing.
SPEAKER_01 (17:44):
Amazing.
Maybe at the end, there's alwaysthese very mean questions.
What are the three things ourlisteners should take away from
this episode?
SPEAKER_00 (17:53):
Yeah.
When you sent this question tome, I was really thinking deeply
on trying not to make it tooobvious.
I mean, I think...
The three things will beobvious, but I think they're
quite powerful and meaningful.
The first is not using AI is theriskiest option.
Every board today is thinking,should we use it?
(18:15):
If we use it, this might happen,et cetera, et cetera.
And there are more publicationsdaily on how you can structure
your guardrails around how toget the best out of it with
mitigating a lot of the risk.
And I would say, like a lot ofthings, that startup mentality
of, try it, get it going, use itin the right ways.
(18:36):
But that is the biggest risk isthat you just blanket rule, you
can't use this.
So I think that's one.
The best directors are going tobe AI literate and you have to
be able to use it on your board.
So that's one.
The second is the best boards.
If you think about the bestperforming boards in the world,
I can always guarantee they'reusing AI already.
(18:56):
And they might be quiet aboutit, but they're using it
confidently in the boardsetting.
Directors are getting thedocuments.
The management team might beusing AI to coach them through
making the best documents forthe board.
Once it gets to the board,they're probably using different
agents to help think better,answer questions, run reports.
(19:18):
They're integrating externalreports at the very least into
their own workflows.
Give me a report on macrochanges.
Okay, cool.
That's really interesting.
How does that feed into theboardroom?
The best directors and the bestboards are already using this.
then the third one is similar tothe prompt question it can be
really hard to get the best outof tools and you're seeing this
(19:39):
in the engineering space forexample the best tools are the
vertical agents using ai similarto the internet it's going to be
the application layer or thethings built on the internet
that provide the highest valueand if you look at cursor
lovable dev like these are thetools that are getting the
biggest use because they'reapplying ai in very vertical
(20:02):
targeted ways and i think as aboard depending on what you're
looking to do you should seekout those vertical tools that
leverage ai to solve a problemyou're not asking the board
you're not asking yourorganization to use ai in a
myriad of ways to You havediscrete problems that you want
to solve and find the solutionsthat solve those problems.
(20:24):
I think that's the third piecethat I'd get listeners to take
away from this.
SPEAKER_01 (20:30):
Jamie, amazing.
Lots of food for thought.
Thank you so, so much forcontributing to the Better
Boards podcast series.
SPEAKER_00 (20:38):
No, amazing.
Pleasure to be here.
SPEAKER_01 (20:40):
Once again, thank you for listening to the
BetterBots podcast series.
If you would like to see that wecover a specific topic you are
interested in or a reference toa challenge you really try to
solve, let us know.
We are real people behind thispodcast.
We talk about AI, but this isnot generated completely by AI.
(21:03):
So do get in touch.
We are happy to hear from you.
If you would like to hear moreabout our work, visit our
website, thewww.better-boats.com
how the smartest directors are using AI.
Welcome to the Better Boardspodcast series, the podcast for
chairs, CEOs, non-executivedirectors, company secretaries,
and their advisors.
Every episode is filled withpractical insights and learnings
from those inside boardrooms.
We discuss what really mattersand highlight actionable steps
(00:25):
you can take to enhance theperformance of your board.
Today's episode It's for everyboard director wondering how to
stay ahead as complexity growsand time shrinks.
We are joined by Jamie Green.
Jamie is co-founder and CEO ofTutaki, an AI platform built
(00:45):
specifically for boarddirectors.
Jamie brings a bold perspectivethat AI isn't just a support
tool, it's a performanceaccelerator.
In this episode, we explore howforward-looking directors are
using AI to enhance theireffectiveness, how governance
workflows are evolving and whatdirectors can do today to stay
(01:06):
at the forefront and at the topof their game.
Jamie, thank you so, so much forcontributing to the Better
Boards podcast series.
SPEAKER_00 (01:14):
No, no, thank you.
Thank you for having
SPEAKER_01 (01:16):
me.
All right.
We love to have people fromaround the world.
So I sit currently in London.
Maybe tell our listeners whereyou are sitting.
SPEAKER_00 (01:24):
Yeah, I'm currently, if anyone's been to the other
side of the planet in NewZealand, I'm in Little Old New
Zealand and even up north in NewZealand, so a small town called
Whangarei.
So yeah, pleasure to be here.
SPEAKER_01 (01:37):
Fantastic.
It's important for the seriesbecause we really want to
capture the perspective ofdifferent people from different
countries, different roles.
So a very warm welcome from NewZealand.
Let's jump straight in.
Why is effectiveness, and welove effectiveness, Yeah, I
mean, it's a great
SPEAKER_00 (02:09):
first question.
You think about the purpose ofcompanies overall is to create
value.
And arguably the place where alot of that value is created is
in the board.
I mean, the board's making themost strategic decisions that
exist operationally and justgenerally.
strategically for that business.
So effectiveness overall at theboard level is massively
(02:30):
important.
And companies that do it well,move faster, make more accurate
decisions, better decisions.
And the companies that don't,the board can feel like a drag.
Management feel like they'repreparing for weeks for a
learning education session.
And not a lot can come out ofit.
There's a lot of trust.
So an effective system is onewhere directors, they focus on
(02:50):
the right things, They're askingthe right questions and
generally not asking managementto make oodles and oodles of
models for them.
And secretaries are streamliningthat process, making sure the
right information is flowing tothe right people.
And the executive team has fulltrust in their board.
So they're not trying to hideinformation.
They're actually sharing theright information at the right
level to make not operational,but strategic decisions.
(03:13):
Often you see 200-page documentsand a board director looks at it
and goes, what am I supposed todecide here?
So directors are the amplifiersand it's really key that that
whole process works effectivelyto actually get the best out of
SPEAKER_01 (03:26):
it.
I mean, the whole community isat the moment really excited
about AI or you see so manypapers and documents and also
podcasts on AI.
Where in your view are mostboards really stuck right now
and what's holding them backfrom really harnessing AI more
fully?
SPEAKER_00 (03:45):
Yeah, it's a great question.
It's often the board themselvesholding them back.
I mean, if you think about AI,it's this thing that's moving
faster than the speed of lightand it changes every day.
So as a director, it's reallydifficult to go, how can I use
this effectively?
How can I be most effective withAI?
And typically that falls intothree categories.
It's what can I do with it?
(04:07):
How can I use it?
And how does this apply to myorganization?
Now, the first two can be kindof trained.
You can just use ChatGPT, ask acouple of questions, get a rough
idea.
But the third one of how thiscan be applied to my
organization, that's where thatunlock really is.
And if a board is sayingoutright, we can't use it, well,
they're the most strategic partof the business.
(04:27):
So if they don't understand howto use it and how it applies to
the organization, then you canbest believe that organization
is not going to get the best outof it.
So those are the threechallenges.
And I think every board...
and will talk to it, shouldembrace it at least at some
level, understand it at the veryleast.
Because if they're not using it,I can almost guarantee the
graduates that are coming out ofuniversity, out of college,
(04:49):
they've been brought up on thisstuff.
And if you tell them they can'tuse it, they're going to use it
without telling you.
And if you create the rightframeworks and guardrails for
them to use it effectively andyou use it yourself, then your
business is just going tothrive.
SPEAKER_01 (05:01):
Now, at least some months ago, I think things are
shifting now.
People were very concerned aboutthe risks.
So let's talk a little bit aboutpitfalls.
What risks do boards need towatch for when adopting AI?
Let's look at it from atechnical side and also from a
cultural side.
SPEAKER_00 (05:19):
Yeah, yeah.
It's one of those things that'smoving so quickly.
We're still actually exploringand finding out, you know, some
of these pitfalls.
But, you know, I've talked withmany a director, and the three
that come to mind are this feararound outputs.
Is it hallucinating?
Are they the wrong thing?
So that's one, and I'll talk tothat.
(05:39):
And the second is, where's mydata going?
I'm uploading, I'm askingquestions, I'm dragging and
dropping documents.
Where is it all going?
And then the third one is, Ithink the most prevalent, is
people are getting over-reliant.
So on outputs being wrong, Imean, if you ask a very broad
question on a broad topic with abroad data set.
AI can still hallucinate things.
(06:01):
It's not infallible.
There are things that can bewrong.
So you need to have a keen eyeon the question and what the
output is.
So that's one.
The data can be shared piece, Ithink, Easily solvable, but
still this fear.
I mean, AI is a black box.
Not even the researchers atAnthropic or OpenAI know
sometimes what it's doing.
So there is that fear aroundwhere it's going.
(06:22):
Again, it can be mitigated.
And then the third one is thisover-reliance.
I mean, we're seeing this atschools today.
Teachers are saying, you know,students coming through schools
are losing that logical part oftheir brain.
And that's no different fordirectors that become
over-reliant on this thing.
So the best directors really areusing it to supplement their own
insight.
They're using their own contextand just making faster, better
(06:42):
decisions.
I mean, they're not using it toreplace themselves.
So it's all about getting thatright data, asking the right
question, and having the rightguardrails in place to make sure
that your privacy is in place,the data is not going anywhere,
and that you're getting the bestout of it for your organization.
SPEAKER_01 (06:58):
You know, in our prep call, you talked about
security aspects.
And yes, it is also from what wecan see a concern for every
board.
But you've said it's riskier notto be using AI than actually
start using it.
Can you explain it a bit?
SPEAKER_00 (07:17):
Yeah, yeah, of course.
I mean, it's a bit of a cheekystatement, but I do think it's
true.
So the risk I mentioned earlier,I think the one that usually
sticks with people the most iswhat's happening to my data.
I mean, you sit on a board,you're making the most strategic
decisions that a company has,and these documents are usually
pretty secretive.
So what happens if it gets out?
(07:38):
But ultimately, if you put theright guardrails in place,
structure the LLM or the AI inthe right way, it's no different
than putting your documents onOneDrive or putting them on
Google Drive.
So I think that can bemitigated.
And it can be mitigated in threeways.
One is you can turn off trainingon the models.
So ChatGPT and Anthropic Cloud,all of them in the settings will
(08:02):
have a model that says stoptraining on this data.
So that's one.
But if you don't trust that,what you can do is create a
private model.
I mean, that's what we do atTutaki.
Every organization has a privatemodel.
And then the third one is youcan actually just host it
on-premise.
So you basically host it on yourservers and you know where that
data is going because you hostit.
(08:23):
You know, that can be mitigated.
And I think if you blanket rule,hey, you can't use AI, well, if
one of your competitors does,they're going to be making
decisions way quicker andthey're going to be more
accurate decisions.
And I think the risk to youactually is just being left
behind entirely.
How you use AI is using it inthe right way that works for
your organization.
(08:43):
I don't think that right way isjust saying you can't use it.
SPEAKER_01 (08:47):
It's, of course, a journey for every board.
They start somewhere and theyhave to take steps to develop,
but there is no finaldestination, of course.
Help our listeners to walk usthrough the journey you have
seen successful boards take.
SPEAKER_00 (09:04):
Yeah, I'll talk through.
I mean, I'd say there's foursteps, and two of them are
probably where companiescurrently are, and two of them
are where I see the futurebeing.
I mean, we're only two yearsinto this technology.
Think about in the context ofthe internet.
It's moving a whole lot faster,but two years into the internet,
we still had some of the biggestcompanies were online pet
(09:26):
stores.
So it's still quite an nascentphase of where it's going to end
up being.
So some of this is obviouslyliable to change, but this is
where I think the world is.
There's four phases.
The first is help me analyzesomething.
And everyone's doing that whenthey use ChatGPT.
Ask a question.
It's doing analysis.
It's giving you an answer.
(09:47):
The second is automate somethingfor me.
And typically that's around achain of events or a chain of
tasks, what we call workflows inthat community.
And if you think about them inthe context of a board, it's
Schedule a meeting, send theagenda, take minutes, take
actions.
That is a workflow, and that canall be automated.
(10:07):
The third phase and fourthphase, I don't think we're at
yet, but I think we're going toget there.
And the third one is AIdirectors, and the fourth is AI
boards.
If you think about directors,sport, for example, has line
judges.
Some of the referees, I mean,boxing had introduced an AI
referee in one of theirheavyweight matches, I think,
(10:27):
recently.
So you're starting to see itplay out as this kind of third
or alternative source ofinformation.
And I don't think boards willreplace directors, but they
might use it as a way to testthinking, to be that kind of
tell me about something in analternative view.
Imagine if you had WarrenBuffett sitting there analysing
(10:48):
the financials.
That's kind of the power of AI,and you can actually create
those agents to give that pointof view.
And the final is AI boards, andI don't think we're going to get
there for a long time.
But similarly, for small boardsand small organizations, like as
a CEO myself, I would love tohave a faux board there that I
could ask questions to thatgives advice, that tests and
(11:08):
prods my board pack before Isend it to my chair.
That would be an amazing way toactually get feedback rapidly.
And I think we will get thereeventually.
So those are the kind of keyfour phases that I see us going
through.
And ultimately, it's the casethat you need to create those
right structures for yourorganization so everything's
being done in the right way.
But what an amazing world that'sgoing to be.
SPEAKER_01 (11:29):
What do you think?
How long will it take us to getto this phase when we really
have an AI board?
Any guess?
Nobody knows, but give it
SPEAKER_00 (11:37):
a shot.
I would say within five years.
If you did today, if you gaveall of your email context,
document context, all thecontext that you're getting on
calls, so transcriptions, etcetera, to an AI algorithm and
chunked it in the right way, itwould have all the relevant
context to make decisions.
And that is often the laggingpiece with AI is it doesn't have
(11:59):
the right context.
But if you give it the rightcontext, it's an incredibly
smart tool.
So I actually think it's notgoing to be a technical
capability that's going to stopus.
It's going to be...
compliance, adoption, and justgeneral human sentiment around
some of these things.
I can imagine startups and smallboards using this stuff a lot
faster than obviously the bigorganizations, but I'd say
(12:23):
within five years.
SPEAKER_01 (12:24):
Amazing.
Absolutely amazing.
Now let's focus more on the now,where we are now in the year
2025.
I think it's important to maybesay 2025 because some of the
people dive into Old podcast.
So if someone listens to this in2027, they may think, oh my
goodness, what are they talkingabout?
(12:45):
So old fashioned.
What are some of the ways youhave seen directors?
How are they using AI in theirboard work today?
Inspire some of our listenerswhat they can do actually today
and what you have seen is reallyworking.
SPEAKER_00 (13:00):
There's too many to count.
But yeah, I mean, one of thebenefits of my role is I get to
talk to hundreds of directorsand many a day.
I'd say there's three bigbuckets, and they're broad, and
I'll get into some exampleswithin each.
But one is saving time.
The second is expanding theirown thinking, their own
(13:21):
aperture.
And the third is managing theprocess.
And save time is the classic, doa research report on my
competitors.
You go to ChatGPT, click deepresearch, here are my
competitors, tell me everythingabout them, what they've done,
Over the last three months, giveme an update.
Save time in understanding that.
Synthesize reports, summarizethings, all of that.
(13:44):
I mean, if you're reading one200-page board packet, it can be
hard.
And if you go, hey, I think wediscussed something three months
ago, it's even harder.
You have to dig through threedifferent other reports.
But ChatGPT, Tutaki can do thatin five seconds.
So that's one.
I think the second is expandtheir thinking and research.
A director is usually someonewho's got experience
(14:06):
operationally in one domain.
You're the finance director, thelegal director, the operational
director, or you've been trainedas a professional director, but
you typically kind of have yourarea of expertise.
And that's why having a verydiverse board is important
because you cover each other'sgaps.
But I'm seeing a lot ofdirectors use AI in a way that
(14:27):
they can fill their own gaps.
So they've got different agentswith different expertise that
answer the question or askquestions in different ways.
So they've got their WarrenBuffett for financials.
They've got their person foroperations, their person for
marketing, the person for sales.
And what it does is it raisesthings or flags things that they
would have entirely missed.
(14:48):
So as a director, you actuallycome to the meeting.
more prepared because youunderstand these other topics in
ways that you want to understandthem in.
So that's the second isexpanding their thinking.
And the third is managing aprocess.
Directors are using tools thatgo through their calendar, help
them get through their emailsquicker, and just generally, as
(15:09):
a board process, manage theprocess.
Help me write agendas.
Some boards are comfortable withtranscriptions.
Take my minutes for me.
Create my documents for me.
Save two weeks of work on themanagement team side by just
aggregating information in areport for me.
So I think there's those threelevels of ways that companies
(15:30):
and boards specifically areusing AI today to save time and
be more productive.
SPEAKER_01 (15:35):
Amazing.
When I see how I started workinglife and how it's evolving now,
I mean, I'm absolutely stunned.
what's possible and where all ofthis is going.
And maybe, you know, ourlisteners are always interested
in this real practical side thatreally inspires them and they
can take and basically, yeah,apply right away.
(15:59):
Maybe leave at the end ourdirectors with some very
practical things a director cando literally this week to begin
using AI safely and alsostrategically.
SPEAKER_00 (16:13):
Yeah, and I hope that if one thing comes out of
this, I hope people take it up.
I think if your listeners startusing it, they'll just be such
more effective directorstomorrow.
The one thing I would try and dois any tool you want, if it's a
general tool, if it's aboard-specific tool, even
(16:35):
better, but the best way to knowwhat you can do with AI is to
ask AI.
If you ask it the question, Andusually the framing is you tell
AI what it is and you ask it howto think.
So you are a director of thistype of organization.
You're trying to do this.
What are five ways you can doit?
(16:57):
Give me a prompt on that.
And by asking it in that generalsense and asking it to create
the prompt, it'll then feed backa very concise, structured
answer for how you can then askit that question.
And I think that's the biggestchallenge with AI today is you
have to be this prompt expert.
I mean, it's called promptengineering in the community.
(17:18):
I mean, we've done thousands ofiterations to get it to the
right output.
And for your director thatdoesn't have the time to be
learning this stuff, why not askthe source what it wants to be
told?
So if you frame it like that, itwill usually give back a really
powerful prompt that you canthen simply ask it to run.
(17:41):
And the output is honestlymind-blowing.
SPEAKER_01 (17:44):
Amazing.
Maybe at the end, there's alwaysthese very mean questions.
What are the three things ourlisteners should take away from
this episode?
SPEAKER_00 (17:53):
Yeah.
When you sent this question tome, I was really thinking deeply
on trying not to make it tooobvious.
I mean, I think...
The three things will beobvious, but I think they're
quite powerful and meaningful.
The first is not using AI is theriskiest option.
Every board today is thinking,should we use it?
(18:15):
If we use it, this might happen,et cetera, et cetera.
And there are more publicationsdaily on how you can structure
your guardrails around how toget the best out of it with
mitigating a lot of the risk.
And I would say, like a lot ofthings, that startup mentality
of, try it, get it going, use itin the right ways.
(18:36):
But that is the biggest risk isthat you just blanket rule, you
can't use this.
So I think that's one.
The best directors are going tobe AI literate and you have to
be able to use it on your board.
So that's one.
The second is the best boards.
If you think about the bestperforming boards in the world,
I can always guarantee they'reusing AI already.
(18:56):
And they might be quiet aboutit, but they're using it
confidently in the boardsetting.
Directors are getting thedocuments.
The management team might beusing AI to coach them through
making the best documents forthe board.
Once it gets to the board,they're probably using different
agents to help think better,answer questions, run reports.
(19:18):
They're integrating externalreports at the very least into
their own workflows.
Give me a report on macrochanges.
Okay, cool.
That's really interesting.
How does that feed into theboardroom?
The best directors and the bestboards are already using this.
then the third one is similar tothe prompt question it can be
really hard to get the best outof tools and you're seeing this
(19:39):
in the engineering space forexample the best tools are the
vertical agents using ai similarto the internet it's going to be
the application layer or thethings built on the internet
that provide the highest valueand if you look at cursor
lovable dev like these are thetools that are getting the
biggest use because they'reapplying ai in very vertical
(20:02):
targeted ways and i think as aboard depending on what you're
looking to do you should seekout those vertical tools that
leverage ai to solve a problemyou're not asking the board
you're not asking yourorganization to use ai in a
myriad of ways to You havediscrete problems that you want
to solve and find the solutionsthat solve those problems.
(20:24):
I think that's the third piecethat I'd get listeners to take
away from this.
SPEAKER_01 (20:30):
Jamie, amazing.
Lots of food for thought.
Thank you so, so much forcontributing to the Better
Boards podcast series.
SPEAKER_00 (20:38):
No, amazing.
Pleasure to be here.
SPEAKER_01 (20:40):
Once again, thank you for listening to the
BetterBots podcast series.
If you would like to see that wecover a specific topic you are
interested in or a reference toa challenge you really try to
solve, let us know.
We are real people behind thispodcast.
We talk about AI, but this isnot generated completely by AI.
(21:03):
So do get in touch.
We are happy to hear from you.
If you would like to hear moreabout our work, visit our
website, thewww.better-boats.com
Popular Podcasts
United States of Kennedy
United States of Kennedy is a podcast about our cultural fascination with the Kennedy dynasty. Every week, hosts Lyra Smith and George Civeris go into one aspect of the Kennedy story.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com