October 2, 2025 • 36 mins
In this episode of On Boards, hosts Joe Ayoub and Raza Shaikh welcome Jamie Green, co-founder and CEO of Tutaki, an AI-powered workbench designed to make board directors significantly more effective.
Drawing from his background in consulting at McKinsey, Green created a technological solution to solve a widespread challenge in board rooms across the U.S. and abroad: information overload and limited time to prepare. Tutaki is an AI system that acts as an expert co-pilot to help directors surface context, identify risk and deliver impact.
Jamie also shares how AI is reshaping governance, what it means for liability and independence, and where the future of boardrooms may be headed.
Key takeaways
- Problems with board overload
- During his time as a consultant at McKinsey, Jamie would deliver 300-page strategy decks to board members 48 hours before a meeting
- Directors would arrive unprepared, leading to conservations about updates and recaps rather than strategic discussions. The lack of preparation risked slow decision making and stifles innovation.
2. New Zealand boardrooms
-
Jamie, based in New Zealand, said board practices in his home country are similar to the U.S. but New Zealand directors tend to have a wider board portfolio — some are on as many as 8 boards.
-
New Zealand has shifted to increase liability for board directors to hold them accountable for reading all board materials. If the business is impacted by the lack of action by a director, they can be criminally liable.
3. Practical pain points that led to Tutaki
- Information overload: directors are expected to digest hundreds of pages, news articles, and compliance documents across multiple boards.
- Contextual awareness: recalling what was decided three months ago or tracking delays in projects across several meetings often requires wading through thousands of pages.
- Portfolio management: professional directors often sit on 5–8 boards, making it difficult to organize materials, track follow-ups, and maintain oversight across them all.
4. AI as a co-pilot, not a replacement
- Tutaki is used to aid directors with meeting preparation by analysing board materials, delivering internal information, relevant news reports, competitor updates and deep dive reports.
- While Tutaki might suggest functions by providing certain information, Green said that the AI serves as a tool and the person using it is still liable
- The AI tool is meant to help directors spend more time doing critical thinking and being effective in meetings by being prepared with the context of the material.
Quotes
-
“We help you do the thinking, but the thinking is still yours, and ultimately that’s the end game.”
-
“I think the power of all this stuff is getting you the context in the way that's meaningful as opposed to just dumping hundreds of pages of information that you have to sift through.”
-
“If you had an AI tool listening to every conversation you had, every email, every document for the last three years on this board, I guarantee it would be a pretty damn good director.”
Links
Guest Bio
Jamie Green is the co-founder and CEO of Tutaki, an AI-powered workbench designed to make board directors radically more effective. Drawing on his background in strategy consulting at McKinsey and hundreds of conversations with chairs and directors, Jamie built Tutaki to solve one of the boardroom’s biggest challenges: staying on top of exploding information, tightening comp
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:05):
Hello and welcome to On Boards, a deepdive at what drives business success.
I'm Joe Ayoub, and I'm herewith my co-host Raza Shaikh.
Twice a month, On Boards is the place tolearn about one of the most critically
important aspects of any company ororganization; its board of directors
or advisors, with a focus on theimportant issues that are facing boards,
(00:28):
company leadership, and stakeholders.
Joe and I speak with a wide range ofguests and talk about what makes a board
successful or unsuccessful, what it meansto be an effective board member, and
how to make your board one of the mostvaluable assets of your organization.
Before we introduce our guest, wewant to thank the law firm of Nutter
(00:51):
McClennen & Fish who are again sponsoringour On Boards Summit this year taking
place on October 20 in their beautifulconference center in the Boston Seaport.
Nutter has been an incrediblepartner with us in every way.
We appreciate all they'vedone to support our podcast.
(01:11):
Our guest today is Jamie Green.
Jamie is the co-founder and CEOof Tutaki, an AI-powered Workbench
designed to make board directorssignificantly more effective.
Drawing on his background in strategyconsulting at McKinsey, Jamie built
Tutaki to solve one of the boardroom'sbiggest challenges; staying on top
(01:35):
of the explosion of information,tightening compliance requirements,
and limited time to prepare.
Currently, directors from 100organizations across Asia Pacific,
North America, and UK use Tutaki toprepare, engage, and make decisions.
(01:56):
Jamie is pioneering a new model ofaugmented directorship where AI acts as an
expert co-pilot to help directors surfacecontext, identify risk and deliver impact.
Tutaki is at the forefront of a movementthat will reshape how boards operate
(02:17):
in a world defined by increasingcomplexity and technological change.
Welcome, Jamie.
Thank you for joiningus today on On Boards.
Oh, it's amazing to be here.
Thank you for having me.
Excellent.
So, let's start with your backgroundat McKinsey because it did have some
influence on where you are today.
(02:39):
Talk about a little bit what youdid as a consultant with McKinsey
in terms of your board work.
Yeah, great question.
McKinsey's amazing because itgives you quite a broad view of
organization strategy and engagementwith management and boards.
I mean, most of what I was doing wasstrategy, M&A, and what that looks like
(03:00):
is you spend a very crunched periodof time coming up with a very dense
documents, usually 300 pages, heaps ofgraphs, a lot of insights, and you hand
it over to a board and management team.
You try and give it a week's worth oftime, but often it comes to their table
48 hours in advance, and it's complex.
Some of the hardest questions acompany has to decide is, do I spend
(03:22):
hundreds of millions of dollarson this company, and you're asking
directors to give their view on thatin a very condensed amount of time.
So, you'd arrive at these meetings withdirectors and you'd see the looks on
faces of what's going on, and I'm sureyou've seen it, the blank looks of,
"Damn, did I read that thing or didI miss this?" And there'd be generic
questions, and as a director, you'rehoping that other people fill your gaps.
(03:44):
But that was my first kindof look into this period of
preparation for these directors.
I mean, a lot of them are professional,they're sitting on other boards.
This is one of many thatthey have to get through.
And for the board meeting, usually it'sthe Thursday of the third week of the
month, and you've got a couple in thatweek, so there's just so much to get
through, and looking at that and going,"How does this process work? How can you
(04:06):
actually help directors prepare betterand be effective in meetings?" Because
the dumb questions need to be asked.
They do.
Someone needs to ask them.
But if everyone is just asking that, itfeels like an update session, a learning
session, as opposed to a strategicdecision session, and I think that's
the risk if you don't actually prepareand have that effective process, is
you just talk about the same stuffmonth in, month out and you don't
(04:26):
actually move the business forward.
So, after McKinsey, what drew you tostartups, and what about your consulting
experience shaped your thinking?
Yeah.
Great question.
I'm sure there's many listeningat the moment and probably
tried and true consultants.
I got a bit used to telling peoplewhat to do and I felt like I needed
(04:47):
to live and breathe what I said.
So, at the end of consulting, youhave those three routes; you stay in
consulting and become a partner, notto say that's easy by any means, or you
join a corporate and try and steer anoil tanker, or you jump on the speedboat
per se and try and launch a startup.
And I thought at that point, you'reonly going to be young once, you
aren't going have much energy, tryand do the startup world and really
(05:10):
live and breathe what you say.
I think one of the things that McKinseygives you is that breadth of experience.
You work across 10, 15 organizationsin a very short period of time across a
very diverse set of questions, so thatgives you that generalist skillset where
you can jump into a company or launch acompany and do everything effectively.
(05:30):
I mean, my background is engineering,so I could also work in that realm and
understood the technical componentsas well as the operational side, and
I really wanted to take that and aproblem that I'd seen firsthand in
this new wave of technology and reallylaunch something that solve that need
in an effective way, which is Tutaki.
So, talk about the founding of Tutakiand, as you put it, the practical
(05:54):
pain points that inspired it.
Yeah.
So, the background really is, throughoutmy journey in startups, I met a guy called
Adam Clark, and Adam had launched andexited a number of startups, one being
M-Com, which was sold to Fiserv, which I'msure most of your listeners have heard of.
It was one of the first onlinebanking or mobile banking platforms.
(06:16):
And at the time, he was sittingon five boards and had this
constant series of pain points,one being information overload.
The world is moving faster than youever expected before, and the compliance
requirements are greater than everbefore, so what it means is not just
board documents getting longer, it'sall of the news articles you need to
read, all of the stuff you need tostay up to date with, all of that vast
(06:39):
amount of information is needed to bedigested to be an effective director
because if you're just coming in withthe context of the business now, you're
no longer giving the value you need to.
So, that was one, it's this painpoint of information overload.
The second was contextual awareness.
So, as a director, you've got, dependingon the board, one pack might be a hundred
pages, but you go, "I think we discussedthat three months ago." So, you have to
(07:01):
go back three months' worth of documents,each of them being a hundred pages.
You are also potentially as an executiveon a company, you've got limited time,
how do you find this information?
So, what it means is when you turnup to the meeting, you feel like
you've discussed something before.
You write down an action, which is thesame thing as it was three months ago.
Or you're talking about a projectthat's, you feel like it's been delayed
(07:22):
three months, but you don't quite know.
So, how do you have thatthread of contextual awareness
across board meetings?
And then the third was justgeneral portfolio management.
What have we said on this previously?
Where does this sit in my emails?
What document does this sit in?
And not just for one board,but for a portfolio of boards
if I'm on a number of boards.
So, how do you be that personalworkbench for a director?
(07:42):
And that's our goal, really, is tobe the vertical system of action that
solves the needs across all three stepsof the board meeting; the pre-meeting,
the end meeting, and the post-meeting,and we're just starting with the
pre-meeting space for directors, sobuilt to be the outcome or the system
of action for the director specifically.
So, just a little bit about New Zealand,I know that Tutaki is not just for New
(08:04):
Zealand boards, but talk a little bitabout how board practice in New Zealand
compares with the US and other countries.
Yeah, that's a great question.
From a structural point,they're very similar.
I mean, it's a unitary board.
They've all got fiduciary duties.
There's subcommittee structures dependingon the size of the organization, but I'd
say in general, New Zealand directorstend to have a wider board portfolio.
(08:28):
It's not uncommon that some of ourcustomers will be on eight boards
and they'll be diverse, some of themlarger, some of them smaller, but
they can be on quite a few boards.
Then the US tend to bea bit more reserved.
A lot of companies in New Zealand arestarting to do AI minutes, smaller
companies, not necessarily largerones, but the US that's pretty slim.
You would be a potential risktaker or a progressive director
(08:52):
if you had AI minutes in the USboard, so that's a different one.
There's a lot of more regulatorypressures in the US, obviously.
It's a bit more intense in that space.
But generally speaking, the painpoints, the process is roughly the same.
I mean, like you mentioned at thestart, we've got customers pretty
diverse across the world at thisstage, and it's the same process.
I get the BoardPAC, we have conversations.
(09:13):
I take my notes, I go to themeeting, and it's just how you
most effectively run that process.
I think you mentioned when we talkedactually a couple months ago, that
compliance requirements for boardsin New Zealand are becoming stricter,
especially when it comes to liability.
What has happened?
Because that's obviously an aspect thatmakes all of this information, all of
(09:36):
the things you've said, even more vitalfor board members because if you're going
to be held liable, then there's an extralevel of concern and maybe stress for
board members, especially if, with alldue respect to whoever they are, if you're
in eight boards, that's a lot of boards.
In this country we would say you'reover boarded, but it is what it is and
(09:57):
we know that we've talked to people inEuropean countries that have the same
practice; lots of boards, and they needhelp just to keep up with the material.
But what's going on?
How has liability in New Zealand shiftedto make it become something that board
members need to be very, very aware of?
Yeah, it's a great question.
(10:18):
The case that's used a lot in New Zealandand the Institute of Directors and other
training organizations is this case of acompany called DexMat, and basically, it
was supposedly a turnaround case, but whathappened was a bunch of visible oversight
that was missing and there was someaccounting issues, inventory management
issues, and what happened was there washundreds of millions of dollars of lost
(10:41):
value over a very short period of time.
While the directors didn't faceany criminal issues for that, it
has kind of spurred on the shift ofincreasing liability for directors
whereby you can't claim that youjust didn't read something anymore.
It is your full liability to read them,and if you don't and it's an impact on
the business and something happens, ifyou're trading in solvent, for example,
(11:03):
then you are criminally liable for that.
This is kind of one of those casesthat gets brought up a lot, and because
of that, it's actually becoming moreand more of a risk to be a director.
Like you said, over boarded is a bigone and now that the value case for a
professional director has to be prettymeaningful for you to take on a new board,
like there is just more and more riskfor you and more and more things like we
(11:26):
talked about before to be across, to makesure that your risk is being mitigated.
Jamie, I think we set up the problemreally well, and it's very, very obvious
to anybody who's in the boardroom.
Let's talk a little bit more abouthow Tutaki actually works, how it
actually would fit with existingtools that board members use for
board package, including tools allthe way from BoardPAC to BoardVantage,
(11:52):
how does Tutaki work in conjunctionwith what boards currently use?
Yeah, the short answer is alongside them.
So, if I give a couple use cases ofdirectors that use different tools in
how they use Tutaki, so there's all theway from the board has nothing or the
director has nothing and they use Tutakias a full portfolio management tool.
(12:14):
So, how that would look like is thesecretary for the organization or
the CEO, whoever's sending out theboard papers, would just cc Janet,
which is our LLM for you, then theywould be uploaded automatically.
You'd get an email sayingyour BoardPACs are ready.
You'd jump in, there would be a dashboardof what's happened with that organization
over the last month, including internalinformation, external news reports that
(12:36):
might be relevant, what your competitorsare doing, and then from there you can
jump into a series of deep dive reports.
So, I'll get into the structureof how that engineering works.
So, that's one case is I have nothing.
The second case is we use boardtools already, so BoardVantage,
Diligent for some of these largerorganizations or if you're smaller,
(12:58):
you might use Boardworks if you're anot-for-profit, that sort of thing.
We work completely alongside them.
So, for a director, all youneed to do is get the BoardPAC.
Now, for some of these organizations,that can be difficult because
they're locked in the black box.
How that works for us is we just have aconversation with the CTO and whoever's
the secretary managing that process willdo a secondary upload, takes 10 seconds,
(13:21):
and then you have your own portal there.
At some stage we'd love to integrate.
Right now, they're out in a bit of a blackbox, and then longer term we'll have a
lot of those processes as well in termsof agendas and minutes and the like.
Now, you are positioning thisas a co-pilot, as an augment for
directors, so it needs to kind ofbehave or think like a director.
(13:43):
How did you train the AI for that,like for basically becoming very
adept at all the questions and usecases that will be asked of it.
Yeah, this is the crux of the IPreally, so without giving away any trade
secrets, we'll get into the detail.
Directors are very, verybroad by definition.
(14:06):
I mean, there's the professionaldirector that has this set of questions
or very kind of compliance-relateddirector that has their checklist
of questions that they'll get down.
That's an easier thing to solve becauseyou just go, "Hey, when I look at
a financial document, what are the50 questions that I need to ask?"
But the general approach orstructure is, there's three elements.
(14:26):
One is how does a director actuallythink, and that's kind of the core
of Tutaki, across every document thatgets uploaded, how do you actually
want to interpret this information?
Because there's one thingis getting good information.
The second is how you use that,so that's one, it's how do you
actually think like a director?
The second is what is the relevantcontext on which to think?
(14:47):
So, that's internal information,external information, and we do a
lot of work there to chunk up therelevant information in the right
way so that it can be interpreted.
And then the third is what I liketo call the actual workflow prompt.
So, if our listeners, I'm sure a lot ofthem have heard of prompt engineering.
This is basically that architecture.
So, the director element is going, "IfI'm a director, what is the structure
(15:11):
in which I should think?" And then thesecond element is, "Okay, how should I
think about these documents?" And thenthe third element is, "Okay, what should
I do with that thinking?" And that generalflow is then how we get to the output for
you, and what it means is that you don'thave to be a prompt expert, you don't
have to get the right information in theright way, we can just run that workflow
for you in a really targeted format.
(15:33):
But just at the technical level, youwere able to leverage existing models
or foundational models to build all ofthis on top, or did you have to do any
specific model training yourself as well?
Yeah, we do.
So, we don't train on any of your data.
I'll just put that one out there.
All your data is private, secure.
(15:54):
It's a private LLM.
If you delete something,it's permanently deleted.
So, there's that security aspect.
The only training we do is on youruser interactions and thumbs up,
thumbs down, equivalent things.
So, the architecture we just talkedabout is the general process, and then
the next level is then, how do youtailor it to what you want to see?
(16:14):
What does Raza actuallywant to see over time?
And that's where the questions you askif you give us that feedback will then
input into how that director thinks.
So, if you are more of a financialdirector and you care about more about
forecast against actual and the likes,then you can tell us that, and then those
prompts will then evolve over time to giveyou more and more of that information.
(16:36):
Now, for us, we need to make sure thatthere's some bandwidth in there of
say like 10 to 20% of stuff that youhaven't said you're interested in, but
you should probably see so that youdon't get too finite in certain areas,
but that's the learning that happens.
It's less on your data and moreon your engagement with the tool,
and so that's basically the flow.
(16:56):
Yeah, well, you touched upon itand that's really important to
be upfront on that part, which isconfidentiality, security of information,
and even liability, if there's any.
How do you guys view that part?
How do you handle it?
How do you give comfort to yourclients and customers that all of their
(17:19):
information is secure and private?
Yeah, it's a great question.
There's three elements to this.
One is how do we store data?
The second is what do we do with it?
And then the third is probablywhat happens if I leave or what
happens if I delete something.
So, how do we store it?
It's all a private instance andthat goes to every organization.
(17:40):
So, say you're a directorand you've got three boards.
Let's just say for our purpose is one ofthem is Apple and one of them's Coca-Cola.
And in the Apple instance you askabout Coca-Cola, it's going to have
no idea what that company is unlessCoca-Cola exists within the Apple
Board documents or equivalent.
So, we keep every company completelyseparate for our larger directors
(18:01):
such that the information doesn'tcross-contaminate, and then it also means
that the answers are a lot more refined.
So, you've all heard around AIhallucinations, that usually happens
when you've got massive data setsand they're very broad, and by
doing this we can narrow downthe context on which we're asking
questions and therefore reduce that.
So, that one is we store everythingprivately in individual instances for the
(18:25):
organizations on a private LLM so it'snot going back anywhere, so that's one.
The second is how do we use it?
And like I said before, we don't doany training on your data at all.
All of that architecture we've done inour side and then what we do with the
documents is effectively, some of yourlisteners might have heard about, it's
just RAG, retrieval augmented generation.
How do we chunk it up in the rightway, store that information such
(18:48):
that it can be retrieved for context.
And then the third one is,what do we do if you leave?
So, in that case, everything'spermanently deleted.
If you add in a board report,ask a question, delete it, ask
the same question, it won'tknow that board report exists.
And we've got bulk wipes.
So, if you want to wipe all of yournotes, you can do that with one click.
Jamie, what about liability?
(19:09):
If somebody comes to you and says,"Tutaki tool gave me certain information
that I relied on to do my function,and that turned out to be bad,
incorrect, or somehow erroneous, youshould be liable." How do you handle
the liability issue for Tutaki?
(19:30):
Yeah, the way we frame it is we are atool that you use, but ultimately, you
are the person that's still liable.
So, we help you do the thinking,but the thinking is still yours,
and ultimately, that's the end game.
I don't think we'd be in businessif we were liable for every single
decision that came out of our platform.
Look, we try and be as accurate aspossible on most things, but ultimately,
(19:51):
that liability sits with you.
So, I wonder about liability when itcomes to the concern that I've heard
many times, which is that as thisdevelops, and we'll talk about where
it might end, board members are lessand less likely to read the material.
I mean, if you're getting 400 pages andyour AI co-pilot can summarize it and
(20:13):
give you the context, I guess part ofthe question is, why would you read it?
Now, the AI may say, "Well, here arefour articles you can go read," and that
could make you a better board member.
But one concern that we've heard isthat board members will become more
and more reliant on their AI consultantand less involved in the materials
(20:38):
themselves, and then at some point,that liability will stick to them
for not doing their job properly.
They can say, "Well, I relied on this."
Certainly, from your company's point ofview, it's not your fault, it's the board
member, as you said, but now if boardmembers are not reading the materials,
and we have cited information in thepast that suggests that over 70% of board
(21:03):
members don't read all the materials.
So, that's a different issue, butit'll become more and more apparent.
And I just wonder about the roleof a human board member, who now
becomes very reliant for the threeor four boards he or she's on, to
just keep up with the information.
What kind of liability might thatgive rise to for the humans involved,
(21:30):
not the company and certainly notthe AI co-pilot who doesn't care.
What about that?
Yeah, look, I think you only need tolook at as far as schools today to
see probably a version of that impact.
You've got kids that are using AIto generate essays, and you've got
teachers that are using AI to gradethose essays, and what it creates is
(21:53):
this complete void of critical thinking,and I think what you want to avoid with
directors is helping them be betterand think better and get to the meeting
with more context and more prepared.
It doesn't necessarilyequate to reading more words.
It's how do you actually prepare better.
As long as they're still doingthe critical thinking and being
(22:13):
effective in those meetings,then that's a great outcome.
Now, I'm not saying that we shouldn'tread anything, and I think how it will
turn out is BoardPACs will get better.
So, it's not that you are reading200-page BoardPACs, it's that the AI
is actually creating a draft BoardPAC.
It's linked into all of your companyinputs, and this is where we want
to get to is we've got your HubSpot,we've got your Xero, we've got all your
(22:35):
accounting systems, your SAP, whateverit is, here's your draft BoardPAC
based on your company objectives.
The CEO and management team thengo through and add their context,
and instead of a 200-page BoardPAC,it's a targeted 40 pages that
you can actually digest at scale.
I think the power of all this stuffis getting you the context in the way
that's meaningful as opposed to justdumping hundreds of pages of information
(22:56):
that you have to sift through.
I don't know about you but if Iread a book, I forget the first
page by the time I'm on the last.
So, even if you're reading all 200 pages,it's a challenge to get the right context.
Great point, because actually it isall about how do you better prepare,
and if the result of this is thatBoardPACs become 40 or 50 pages, I think
(23:21):
everyone would agree that that is apositive development in the boardroom.
There's just no question about thatbecause increasingly, I think we all know
that management puts more and more outthere so they can't be criticized for
holding something back from the board,and that's just not the best way to do it.
It's just not.
(23:42):
Well, I was going to say, it's soineffective for you because if they put
200 pages in a BoardPAC because they'vedone the work and they want you to see
it, that liability now sits on you.
So, they've put in all this workand shared with you information that
probably isn't relevant and now it'son you to read it all which just
is not effective for anyone really.
But Joe, the point that youbrought up actually make me think
(24:03):
that the implication of all thison the business judgment rule.
In the US, like the basic theme is there'sno law against a company dying or not
successful at the business level as long.
As the directors made a good judgmentbased on the information that was provided
at that time, they are actually not liablebased on the business judgment rules.
(24:26):
So, I'm like, "Wow, does this changethis where it's saying, 'Well, now you
become complacent and you don't readit because you think the AI is telling
you, 'Are you making good judgment?'"I think Jamie's tool need to add more
Socratic method questions to challenge thereader, not just be a passive assistant,
(24:48):
but be like, "Whoa, are you thinkingabout it the right way?" Or actually
ask them questions to say the least.
I don't know, but this islike this is fascinating.
We'll see where it goes.
So, I want to talk a little bitabout the actual user experience.
because when we talked the last time,you put on the screen what a user might
see, and I wrote down a bunch of it.
(25:08):
because I think it's reallytelling what you're going to see.
So, inconsistency analysis, getinsights, ask questions, company
retrospective, board meetingpre-read, so obviously, a summary
of the pre-read, make me look smart.
That was one of them.
And I thought, "Wow, you know what,that's the one I'm going to first,
(25:31):
right?" So, talk about how you tailorit for each individual or how that
comes about and how you make itinteractive, easy to interact with.
Yeah.
And I'll have to show you guys a new demo.
There's been a whole lot of additionssince, since the last time we caught up.
(25:52):
But if you think about it inthree ways, one is, what do you
want to see as an individual?
The second is, how do you wantto engage with that information?
And then the third is, at what level?
So, what do you want to see?
We ask all our directors, whatare their key interest points?
What are the topics thatthey want to follow?
And that then feeds into, like I saidat the start, that architecture of how
(26:14):
we gather and chunk up information.
So, if you're a financial directorthat only wants to see key metrics,
how they're performing, what are thekey changes in the train lines, then
we'll index more heavily in that.
If you're a director and moreoperational and you are looking at like
manufacturing plan and that's your keyfocus, we'll index more heavily on that.
So, that that first piecetailors the user experience.
(26:36):
And then the second one,how do you engage with it?
So, the general flow, and we're tryingnot to upset the cart too much in terms
of how directors engage, but allowdirectors to choose their own journey.
So, like I said before, mostdirectors get the BoardPAC.
They read it to some level, some allthe way down to every bullet point, and
then some will skim it or quote unquoteskim, which is basically, relying on gut.
(26:59):
And then we use that process andallow you to tailor it slightly.
So, the first step is you get in,here's your prepared company dashboard
based on your topics of interestof what's changed since the last
meeting, what are key red flags in thefinancial area, risk area, compliance,
what are your competitors doing?
Give me a 10-minute scan ofwhat I should be interested in.
(27:20):
Okay, cool.
Now, when I get to reading the BoardPAC,I have a lot of context on where
I want to focus my time, and alsocontext on the meaning behind some
of the things that are being said.
So, the next step is then obviouslyreading your BoardPACs, we align them
all to agenda items, and then we havethat 24/7 context as you read it.
So, you're reading through adocument, here's a project.
(27:42):
We actually put a red flag next to thatfor you, "Hey, this project's moved three
times in the last three board meetings,which you might not be aware of. The due
dates changed by six months and now, it'sat risk of completely missed delivery."
So, as you are reading the packet, youdon't have to go through all documents.
It's all there for you.
So, that's the second interms of how you engage.
Then the other one is form factor,and we're seeing a lot now is the
(28:05):
typical ways you read documents,but a lot more is, do you want to
actually speak to your documents?
I'm going for a walk, I'm sittingin a cab, I'm going to have a
conversation about my BoardPACs.
And that's kind of this next valuepoint of AI is having to be able to
have a live conversation at depthon topics and across your BoardPACs.
So, like we're having right now, I couldjust discuss with Tutaki the BoardPAC, ask
(28:30):
questions about the last three months andgo into a deep conversation about it in
a cab, and that's a new kind of avenue.
And then the third is, at whatlevel do you want to engage on it?
And like I said, you can read the wholeBoardPAC, you can jump from the dashboard
to your run sheet, which goes these arethe key questions that you've highlighted
as important for the meeting and thenall of that is tailorable to you, so
(28:51):
you can run all of it or some of it.
We know that the process foreach director is very different.
So, that's basically the userexperience and the button of
make me smarter hopefully isembedded throughout the process.
We can't tap into your brain, butbasically, if you think about the core of
it, it's help me ask the right questionson the right stuff with the right context,
(29:11):
and that's what we're trying to do really.
Yeah, that's a good way to think about it.
So, one thing we did when wetalked before, and this was a great
conversation that we had, which is,what does the future look like for AI?
And Raza mentioned we have the OnBoards AI Board Advisor Maturity Model.
(29:33):
So, on this model, here'swhat we talked about.
Level one is AI isassisting one board member.
Two is it's assisting the board.
Third is its AI is in the boardroom.
Next, AI is a non-voting board member.
(29:55):
Next level is AI is a voting board member.
Next is AI is the dominant voiceon the board, meaning there'd be
more AI than human in the room.
And the highest level, level seven is.
AI is the only voice on the board.
You get rid of the human elementand all the things that people
(30:17):
forget and mistake, and you reallyjust clean up the whole thing.
Now, I'm sort of kidding aboutthat, but the reality is, that
seems like a logical progression.
It seems like there's so much tosay about the advantages as AI
develops and as you develop thismodel, it's going to get better.
(30:39):
It's going to make fewer mistakes,it's going to understand better
what a board member needs to do.
All of the things that you trainboard members to do, AI will
remember and it won't forget them.
So, on this AI Board AdvisorMaturity Model, we're someplace at
(30:59):
number two, AI assisting the board.
I think that's where Tutaki is.
Do you think we'll get to levelseven anytime in our lifetime?
It's funny you asked this.
We did a talk and there wereabout 150 directors there, and
a similar question was asked.
It was like, "Am I going to be replacedat some point?" The scary thing is
this is every nature of our job, right?
(31:22):
And I think where it lacks nowis there's a lot of context that
sits outside of documentation.
You have a conversation betweentwo people over a coffee, and that
becomes context for then how youthink in a meeting and what questions
you ask and how you answer them.
If you had an AI tool listening to everyconversation you had, every email, every
document for the last three years onthis board, I guarantee it would be a
(31:45):
pretty damn good director and it would beborderline being able to sit on the board.
So, until we solve that informationgap, I think it'll be difficult, but
I think it will get there and I thinkthe adoption curve will look a lot
steeper in these smaller organizations.
Myself, I would love to have afull board there that I can share
my BoardPAC with and go grill me.
(32:06):
Give me all the questions thatyou've got so that I can answer
them and tailor my material beforeI get to the actual board meeting.
Like the value of that is crazily high.
And then for a board meeting, youmight see it similar to Wimbledon
that's just had AI line judges.
There are a couple of kinks they needto work out, but it might sit in the
board meeting and be kind of that devil'sadvocate or give an opposition view on
(32:27):
certain topics so that you don't havethis groupthink element and can actually
discuss things in a different way.
So, I think that'll be the starting pointand then when we get to this step seven
of AI, everything, I think we'll be at thepoint in time where we've got AI boards
talking to AI companies, talking to AIconsumers, so that'll be completely a
(32:47):
paradigm shift in how we approach that.
Would there be a non-human voting member?
That's actually a moreinteresting question.
Like even one vote from AI boardmember makes it very, very interesting,
and that can be foreseeable in maybeless timeframe than the level seven.
(33:09):
But it would be fascinating to seewhether the law and legal frameworks
and all of those catch up to itand say, "Oh, yeah, you are allowed
to have one voting AI member."
And we talk about independentdirectors and non-independent
directors, the argument would be thatthis AI is even more independent.
Really, it's not beholden tothe CEO or to the board chair.
(33:33):
It doesn't care about compensation.
It doesn't care about beingrenewed for another term.
It doesn't care about, did itembarrass itself at the last meeting?
It's truly independent, notperfect, but independent.
And I could see that first vote, asyou said, I agree, that's the most
likely step that we're going to go.
(33:54):
That's a really great point, Joe.
That would be the prediction thing.
The independent tie-breaking,odd number director.
Jamie didn't take me out on his yacht lastweek and try to influence me on my vote.
I didn't go out to dinner at someguy's castle on a Greek island.
It's just, I looked at the stuff,I read all the emails, I saw every
(34:17):
board minute for the last fiveyears, and this is what I think.
And I have the humanities expert knowledgeon this industry on my fingertips, and
I'm saying, I will vote this or that.
Yes.
So, I would not be surprised if ithappens even in my lifetime, but I'm
positive it'll happen in yours, Jamie.
(34:37):
And I think the other thing for thefuture of AI in the boardroom, I think
what you mentioned that management hashad now to do a lot more disclosure and
provide a lot of materials, and I thinkwhat you alluded to as, well, even for
management, that's actually quite a bitof work to come up with these materials,
(35:00):
and you could foresee a set of tools thatcan gather all that primary sources of
information, as detailed as it can be inthe company down to, as you mentioned, the
transactions in the accounting system, butbring it up into a manner for the board
that is much more readable and so on.
(35:21):
So, hopefully, that also happenssoon because we are indeed inundated.
I hope that happens right away.
That's something everyone'sgoing to welcome.
Yeah, I mean, I was talking with oneof our customers who wants to use it
for the exact purpose across theirmanagement team's sub-teams, they
spend up to two weeks a month and it'slike a million dollars of people time.
(35:43):
Every month they do that, and thenby the time the board meeting's
over, they're starting the next one.
So, from a cost point of view, it'smassive, and these documents are huge.
So, when a director gets them,it's hard to actually digest them.
So, you've got arguably the moststrategic group in the company
struggling to ingest this stuff.
And arguably the most useful operatorsin the company spending half the year
(36:06):
creating those documents, which arehard to read, so it's a massive pain
point for the organization to solve.
Absolutely.
Jamie, this has been afascinating conversation.
Thank you so much for joining us today,and thank you all for listening to On
Boards with our guest, Jamie Green.
Please visit our websiteat OnBoardsPodcast.com.
(36:28):
That's OnBoardsPodcast.com.
We'd love to hear your comments,suggestions, and feedback.
If you're not already a subscriber, pleasebe sure to subscribe at Apple Podcast,
Spotify, or wherever you get your podcast.
And remember to leaveus a five star review.
And please tune in for thenext episode of On Boards.
(36:48):
Thanks.
Hello and welcome to On Boards, a deepdive at what drives business success.
I'm Joe Ayoub, and I'm herewith my co-host Raza Shaikh.
Twice a month, On Boards is the place tolearn about one of the most critically
important aspects of any company ororganization; its board of directors
or advisors, with a focus on theimportant issues that are facing boards,
(00:28):
company leadership, and stakeholders.
Joe and I speak with a wide range ofguests and talk about what makes a board
successful or unsuccessful, what it meansto be an effective board member, and
how to make your board one of the mostvaluable assets of your organization.
Before we introduce our guest, wewant to thank the law firm of Nutter
(00:51):
McClennen & Fish who are again sponsoringour On Boards Summit this year taking
place on October 20 in their beautifulconference center in the Boston Seaport.
Nutter has been an incrediblepartner with us in every way.
We appreciate all they'vedone to support our podcast.
(01:11):
Our guest today is Jamie Green.
Jamie is the co-founder and CEOof Tutaki, an AI-powered Workbench
designed to make board directorssignificantly more effective.
Drawing on his background in strategyconsulting at McKinsey, Jamie built
Tutaki to solve one of the boardroom'sbiggest challenges; staying on top
(01:35):
of the explosion of information,tightening compliance requirements,
and limited time to prepare.
Currently, directors from 100organizations across Asia Pacific,
North America, and UK use Tutaki toprepare, engage, and make decisions.
(01:56):
Jamie is pioneering a new model ofaugmented directorship where AI acts as an
expert co-pilot to help directors surfacecontext, identify risk and deliver impact.
Tutaki is at the forefront of a movementthat will reshape how boards operate
(02:17):
in a world defined by increasingcomplexity and technological change.
Welcome, Jamie.
Thank you for joiningus today on On Boards.
Oh, it's amazing to be here.
Thank you for having me.
Excellent.
So, let's start with your backgroundat McKinsey because it did have some
influence on where you are today.
(02:39):
Talk about a little bit what youdid as a consultant with McKinsey
in terms of your board work.
Yeah, great question.
McKinsey's amazing because itgives you quite a broad view of
organization strategy and engagementwith management and boards.
I mean, most of what I was doing wasstrategy, M&A, and what that looks like
(03:00):
is you spend a very crunched periodof time coming up with a very dense
documents, usually 300 pages, heaps ofgraphs, a lot of insights, and you hand
it over to a board and management team.
You try and give it a week's worth oftime, but often it comes to their table
48 hours in advance, and it's complex.
Some of the hardest questions acompany has to decide is, do I spend
(03:22):
hundreds of millions of dollarson this company, and you're asking
directors to give their view on thatin a very condensed amount of time.
So, you'd arrive at these meetings withdirectors and you'd see the looks on
faces of what's going on, and I'm sureyou've seen it, the blank looks of,
"Damn, did I read that thing or didI miss this?" And there'd be generic
questions, and as a director, you'rehoping that other people fill your gaps.
(03:44):
But that was my first kindof look into this period of
preparation for these directors.
I mean, a lot of them are professional,they're sitting on other boards.
This is one of many thatthey have to get through.
And for the board meeting, usually it'sthe Thursday of the third week of the
month, and you've got a couple in thatweek, so there's just so much to get
through, and looking at that and going,"How does this process work? How can you
(04:06):
actually help directors prepare betterand be effective in meetings?" Because
the dumb questions need to be asked.
They do.
Someone needs to ask them.
But if everyone is just asking that, itfeels like an update session, a learning
session, as opposed to a strategicdecision session, and I think that's
the risk if you don't actually prepareand have that effective process, is
you just talk about the same stuffmonth in, month out and you don't
(04:26):
actually move the business forward.
So, after McKinsey, what drew you tostartups, and what about your consulting
experience shaped your thinking?
Yeah.
Great question.
I'm sure there's many listeningat the moment and probably
tried and true consultants.
I got a bit used to telling peoplewhat to do and I felt like I needed
(04:47):
to live and breathe what I said.
So, at the end of consulting, youhave those three routes; you stay in
consulting and become a partner, notto say that's easy by any means, or you
join a corporate and try and steer anoil tanker, or you jump on the speedboat
per se and try and launch a startup.
And I thought at that point, you'reonly going to be young once, you
aren't going have much energy, tryand do the startup world and really
(05:10):
live and breathe what you say.
I think one of the things that McKinseygives you is that breadth of experience.
You work across 10, 15 organizationsin a very short period of time across a
very diverse set of questions, so thatgives you that generalist skillset where
you can jump into a company or launch acompany and do everything effectively.
(05:30):
I mean, my background is engineering,so I could also work in that realm and
understood the technical componentsas well as the operational side, and
I really wanted to take that and aproblem that I'd seen firsthand in
this new wave of technology and reallylaunch something that solve that need
in an effective way, which is Tutaki.
So, talk about the founding of Tutakiand, as you put it, the practical
(05:54):
pain points that inspired it.
Yeah.
So, the background really is, throughoutmy journey in startups, I met a guy called
Adam Clark, and Adam had launched andexited a number of startups, one being
M-Com, which was sold to Fiserv, which I'msure most of your listeners have heard of.
It was one of the first onlinebanking or mobile banking platforms.
(06:16):
And at the time, he was sittingon five boards and had this
constant series of pain points,one being information overload.
The world is moving faster than youever expected before, and the compliance
requirements are greater than everbefore, so what it means is not just
board documents getting longer, it'sall of the news articles you need to
read, all of the stuff you need tostay up to date with, all of that vast
(06:39):
amount of information is needed to bedigested to be an effective director
because if you're just coming in withthe context of the business now, you're
no longer giving the value you need to.
So, that was one, it's this painpoint of information overload.
The second was contextual awareness.
So, as a director, you've got, dependingon the board, one pack might be a hundred
pages, but you go, "I think we discussedthat three months ago." So, you have to
(07:01):
go back three months' worth of documents,each of them being a hundred pages.
You are also potentially as an executiveon a company, you've got limited time,
how do you find this information?
So, what it means is when you turnup to the meeting, you feel like
you've discussed something before.
You write down an action, which is thesame thing as it was three months ago.
Or you're talking about a projectthat's, you feel like it's been delayed
(07:22):
three months, but you don't quite know.
So, how do you have thatthread of contextual awareness
across board meetings?
And then the third was justgeneral portfolio management.
What have we said on this previously?
Where does this sit in my emails?
What document does this sit in?
And not just for one board,but for a portfolio of boards
if I'm on a number of boards.
So, how do you be that personalworkbench for a director?
(07:42):
And that's our goal, really, is tobe the vertical system of action that
solves the needs across all three stepsof the board meeting; the pre-meeting,
the end meeting, and the post-meeting,and we're just starting with the
pre-meeting space for directors, sobuilt to be the outcome or the system
of action for the director specifically.
So, just a little bit about New Zealand,I know that Tutaki is not just for New
(08:04):
Zealand boards, but talk a little bitabout how board practice in New Zealand
compares with the US and other countries.
Yeah, that's a great question.
From a structural point,they're very similar.
I mean, it's a unitary board.
They've all got fiduciary duties.
There's subcommittee structures dependingon the size of the organization, but I'd
say in general, New Zealand directorstend to have a wider board portfolio.
(08:28):
It's not uncommon that some of ourcustomers will be on eight boards
and they'll be diverse, some of themlarger, some of them smaller, but
they can be on quite a few boards.
Then the US tend to bea bit more reserved.
A lot of companies in New Zealand arestarting to do AI minutes, smaller
companies, not necessarily largerones, but the US that's pretty slim.
You would be a potential risktaker or a progressive director
(08:52):
if you had AI minutes in the USboard, so that's a different one.
There's a lot of more regulatorypressures in the US, obviously.
It's a bit more intense in that space.
But generally speaking, the painpoints, the process is roughly the same.
I mean, like you mentioned at thestart, we've got customers pretty
diverse across the world at thisstage, and it's the same process.
I get the BoardPAC, we have conversations.
(09:13):
I take my notes, I go to themeeting, and it's just how you
most effectively run that process.
I think you mentioned when we talkedactually a couple months ago, that
compliance requirements for boardsin New Zealand are becoming stricter,
especially when it comes to liability.
What has happened?
Because that's obviously an aspect thatmakes all of this information, all of
(09:36):
the things you've said, even more vitalfor board members because if you're going
to be held liable, then there's an extralevel of concern and maybe stress for
board members, especially if, with alldue respect to whoever they are, if you're
in eight boards, that's a lot of boards.
In this country we would say you'reover boarded, but it is what it is and
(09:57):
we know that we've talked to people inEuropean countries that have the same
practice; lots of boards, and they needhelp just to keep up with the material.
But what's going on?
How has liability in New Zealand shiftedto make it become something that board
members need to be very, very aware of?
Yeah, it's a great question.
(10:18):
The case that's used a lot in New Zealandand the Institute of Directors and other
training organizations is this case of acompany called DexMat, and basically, it
was supposedly a turnaround case, but whathappened was a bunch of visible oversight
that was missing and there was someaccounting issues, inventory management
issues, and what happened was there washundreds of millions of dollars of lost
(10:41):
value over a very short period of time.
While the directors didn't faceany criminal issues for that, it
has kind of spurred on the shift ofincreasing liability for directors
whereby you can't claim that youjust didn't read something anymore.
It is your full liability to read them,and if you don't and it's an impact on
the business and something happens, ifyou're trading in solvent, for example,
(11:03):
then you are criminally liable for that.
This is kind of one of those casesthat gets brought up a lot, and because
of that, it's actually becoming moreand more of a risk to be a director.
Like you said, over boarded is a bigone and now that the value case for a
professional director has to be prettymeaningful for you to take on a new board,
like there is just more and more riskfor you and more and more things like we
(11:26):
talked about before to be across, to makesure that your risk is being mitigated.
Jamie, I think we set up the problemreally well, and it's very, very obvious
to anybody who's in the boardroom.
Let's talk a little bit more abouthow Tutaki actually works, how it
actually would fit with existingtools that board members use for
board package, including tools allthe way from BoardPAC to BoardVantage,
(11:52):
how does Tutaki work in conjunctionwith what boards currently use?
Yeah, the short answer is alongside them.
So, if I give a couple use cases ofdirectors that use different tools in
how they use Tutaki, so there's all theway from the board has nothing or the
director has nothing and they use Tutakias a full portfolio management tool.
(12:14):
So, how that would look like is thesecretary for the organization or
the CEO, whoever's sending out theboard papers, would just cc Janet,
which is our LLM for you, then theywould be uploaded automatically.
You'd get an email sayingyour BoardPACs are ready.
You'd jump in, there would be a dashboardof what's happened with that organization
over the last month, including internalinformation, external news reports that
(12:36):
might be relevant, what your competitorsare doing, and then from there you can
jump into a series of deep dive reports.
So, I'll get into the structureof how that engineering works.
So, that's one case is I have nothing.
The second case is we use boardtools already, so BoardVantage,
Diligent for some of these largerorganizations or if you're smaller,
(12:58):
you might use Boardworks if you're anot-for-profit, that sort of thing.
We work completely alongside them.
So, for a director, all youneed to do is get the BoardPAC.
Now, for some of these organizations,that can be difficult because
they're locked in the black box.
How that works for us is we just have aconversation with the CTO and whoever's
the secretary managing that process willdo a secondary upload, takes 10 seconds,
(13:21):
and then you have your own portal there.
At some stage we'd love to integrate.
Right now, they're out in a bit of a blackbox, and then longer term we'll have a
lot of those processes as well in termsof agendas and minutes and the like.
Now, you are positioning thisas a co-pilot, as an augment for
directors, so it needs to kind ofbehave or think like a director.
(13:43):
How did you train the AI for that,like for basically becoming very
adept at all the questions and usecases that will be asked of it.
Yeah, this is the crux of the IPreally, so without giving away any trade
secrets, we'll get into the detail.
Directors are very, verybroad by definition.
(14:06):
I mean, there's the professionaldirector that has this set of questions
or very kind of compliance-relateddirector that has their checklist
of questions that they'll get down.
That's an easier thing to solve becauseyou just go, "Hey, when I look at
a financial document, what are the50 questions that I need to ask?"
But the general approach orstructure is, there's three elements.
(14:26):
One is how does a director actuallythink, and that's kind of the core
of Tutaki, across every document thatgets uploaded, how do you actually
want to interpret this information?
Because there's one thingis getting good information.
The second is how you use that,so that's one, it's how do you
actually think like a director?
The second is what is the relevantcontext on which to think?
(14:47):
So, that's internal information,external information, and we do a
lot of work there to chunk up therelevant information in the right
way so that it can be interpreted.
And then the third is what I liketo call the actual workflow prompt.
So, if our listeners, I'm sure a lot ofthem have heard of prompt engineering.
This is basically that architecture.
So, the director element is going, "IfI'm a director, what is the structure
(15:11):
in which I should think?" And then thesecond element is, "Okay, how should I
think about these documents?" And thenthe third element is, "Okay, what should
I do with that thinking?" And that generalflow is then how we get to the output for
you, and what it means is that you don'thave to be a prompt expert, you don't
have to get the right information in theright way, we can just run that workflow
for you in a really targeted format.
(15:33):
But just at the technical level, youwere able to leverage existing models
or foundational models to build all ofthis on top, or did you have to do any
specific model training yourself as well?
Yeah, we do.
So, we don't train on any of your data.
I'll just put that one out there.
All your data is private, secure.
(15:54):
It's a private LLM.
If you delete something,it's permanently deleted.
So, there's that security aspect.
The only training we do is on youruser interactions and thumbs up,
thumbs down, equivalent things.
So, the architecture we just talkedabout is the general process, and then
the next level is then, how do youtailor it to what you want to see?
(16:14):
What does Raza actuallywant to see over time?
And that's where the questions you askif you give us that feedback will then
input into how that director thinks.
So, if you are more of a financialdirector and you care about more about
forecast against actual and the likes,then you can tell us that, and then those
prompts will then evolve over time to giveyou more and more of that information.
(16:36):
Now, for us, we need to make sure thatthere's some bandwidth in there of
say like 10 to 20% of stuff that youhaven't said you're interested in, but
you should probably see so that youdon't get too finite in certain areas,
but that's the learning that happens.
It's less on your data and moreon your engagement with the tool,
and so that's basically the flow.
(16:56):
Yeah, well, you touched upon itand that's really important to
be upfront on that part, which isconfidentiality, security of information,
and even liability, if there's any.
How do you guys view that part?
How do you handle it?
How do you give comfort to yourclients and customers that all of their
(17:19):
information is secure and private?
Yeah, it's a great question.
There's three elements to this.
One is how do we store data?
The second is what do we do with it?
And then the third is probablywhat happens if I leave or what
happens if I delete something.
So, how do we store it?
It's all a private instance andthat goes to every organization.
(17:40):
So, say you're a directorand you've got three boards.
Let's just say for our purpose is one ofthem is Apple and one of them's Coca-Cola.
And in the Apple instance you askabout Coca-Cola, it's going to have
no idea what that company is unlessCoca-Cola exists within the Apple
Board documents or equivalent.
So, we keep every company completelyseparate for our larger directors
(18:01):
such that the information doesn'tcross-contaminate, and then it also means
that the answers are a lot more refined.
So, you've all heard around AIhallucinations, that usually happens
when you've got massive data setsand they're very broad, and by
doing this we can narrow downthe context on which we're asking
questions and therefore reduce that.
So, that one is we store everythingprivately in individual instances for the
(18:25):
organizations on a private LLM so it'snot going back anywhere, so that's one.
The second is how do we use it?
And like I said before, we don't doany training on your data at all.
All of that architecture we've done inour side and then what we do with the
documents is effectively, some of yourlisteners might have heard about, it's
just RAG, retrieval augmented generation.
How do we chunk it up in the rightway, store that information such
(18:48):
that it can be retrieved for context.
And then the third one is,what do we do if you leave?
So, in that case, everything'spermanently deleted.
If you add in a board report,ask a question, delete it, ask
the same question, it won'tknow that board report exists.
And we've got bulk wipes.
So, if you want to wipe all of yournotes, you can do that with one click.
Jamie, what about liability?
(19:09):
If somebody comes to you and says,"Tutaki tool gave me certain information
that I relied on to do my function,and that turned out to be bad,
incorrect, or somehow erroneous, youshould be liable." How do you handle
the liability issue for Tutaki?
(19:30):
Yeah, the way we frame it is we are atool that you use, but ultimately, you
are the person that's still liable.
So, we help you do the thinking,but the thinking is still yours,
and ultimately, that's the end game.
I don't think we'd be in businessif we were liable for every single
decision that came out of our platform.
Look, we try and be as accurate aspossible on most things, but ultimately,
(19:51):
that liability sits with you.
So, I wonder about liability when itcomes to the concern that I've heard
many times, which is that as thisdevelops, and we'll talk about where
it might end, board members are lessand less likely to read the material.
I mean, if you're getting 400 pages andyour AI co-pilot can summarize it and
(20:13):
give you the context, I guess part ofthe question is, why would you read it?
Now, the AI may say, "Well, here arefour articles you can go read," and that
could make you a better board member.
But one concern that we've heard isthat board members will become more
and more reliant on their AI consultantand less involved in the materials
(20:38):
themselves, and then at some point,that liability will stick to them
for not doing their job properly.
They can say, "Well, I relied on this."
Certainly, from your company's point ofview, it's not your fault, it's the board
member, as you said, but now if boardmembers are not reading the materials,
and we have cited information in thepast that suggests that over 70% of board
(21:03):
members don't read all the materials.
So, that's a different issue, butit'll become more and more apparent.
And I just wonder about the roleof a human board member, who now
becomes very reliant for the threeor four boards he or she's on, to
just keep up with the information.
What kind of liability might thatgive rise to for the humans involved,
(21:30):
not the company and certainly notthe AI co-pilot who doesn't care.
What about that?
Yeah, look, I think you only need tolook at as far as schools today to
see probably a version of that impact.
You've got kids that are using AIto generate essays, and you've got
teachers that are using AI to gradethose essays, and what it creates is
(21:53):
this complete void of critical thinking,and I think what you want to avoid with
directors is helping them be betterand think better and get to the meeting
with more context and more prepared.
It doesn't necessarilyequate to reading more words.
It's how do you actually prepare better.
As long as they're still doingthe critical thinking and being
(22:13):
effective in those meetings,then that's a great outcome.
Now, I'm not saying that we shouldn'tread anything, and I think how it will
turn out is BoardPACs will get better.
So, it's not that you are reading200-page BoardPACs, it's that the AI
is actually creating a draft BoardPAC.
It's linked into all of your companyinputs, and this is where we want
to get to is we've got your HubSpot,we've got your Xero, we've got all your
(22:35):
accounting systems, your SAP, whateverit is, here's your draft BoardPAC
based on your company objectives.
The CEO and management team thengo through and add their context,
and instead of a 200-page BoardPAC,it's a targeted 40 pages that
you can actually digest at scale.
I think the power of all this stuffis getting you the context in the way
that's meaningful as opposed to justdumping hundreds of pages of information
(22:56):
that you have to sift through.
I don't know about you but if Iread a book, I forget the first
page by the time I'm on the last.
So, even if you're reading all 200 pages,it's a challenge to get the right context.
Great point, because actually it isall about how do you better prepare,
and if the result of this is thatBoardPACs become 40 or 50 pages, I think
(23:21):
everyone would agree that that is apositive development in the boardroom.
There's just no question about thatbecause increasingly, I think we all know
that management puts more and more outthere so they can't be criticized for
holding something back from the board,and that's just not the best way to do it.
It's just not.
(23:42):
Well, I was going to say, it's soineffective for you because if they put
200 pages in a BoardPAC because they'vedone the work and they want you to see
it, that liability now sits on you.
So, they've put in all this workand shared with you information that
probably isn't relevant and now it'son you to read it all which just
is not effective for anyone really.
But Joe, the point that youbrought up actually make me think
(24:03):
that the implication of all thison the business judgment rule.
In the US, like the basic theme is there'sno law against a company dying or not
successful at the business level as long.
As the directors made a good judgmentbased on the information that was provided
at that time, they are actually not liablebased on the business judgment rules.
(24:26):
So, I'm like, "Wow, does this changethis where it's saying, 'Well, now you
become complacent and you don't readit because you think the AI is telling
you, 'Are you making good judgment?'"I think Jamie's tool need to add more
Socratic method questions to challenge thereader, not just be a passive assistant,
(24:48):
but be like, "Whoa, are you thinkingabout it the right way?" Or actually
ask them questions to say the least.
I don't know, but this islike this is fascinating.
We'll see where it goes.
So, I want to talk a little bitabout the actual user experience.
because when we talked the last time,you put on the screen what a user might
see, and I wrote down a bunch of it.
(25:08):
because I think it's reallytelling what you're going to see.
So, inconsistency analysis, getinsights, ask questions, company
retrospective, board meetingpre-read, so obviously, a summary
of the pre-read, make me look smart.
That was one of them.
And I thought, "Wow, you know what,that's the one I'm going to first,
(25:31):
right?" So, talk about how you tailorit for each individual or how that
comes about and how you make itinteractive, easy to interact with.
Yeah.
And I'll have to show you guys a new demo.
There's been a whole lot of additionssince, since the last time we caught up.
(25:52):
But if you think about it inthree ways, one is, what do you
want to see as an individual?
The second is, how do you wantto engage with that information?
And then the third is, at what level?
So, what do you want to see?
We ask all our directors, whatare their key interest points?
What are the topics thatthey want to follow?
And that then feeds into, like I saidat the start, that architecture of how
(26:14):
we gather and chunk up information.
So, if you're a financial directorthat only wants to see key metrics,
how they're performing, what are thekey changes in the train lines, then
we'll index more heavily in that.
If you're a director and moreoperational and you are looking at like
manufacturing plan and that's your keyfocus, we'll index more heavily on that.
So, that that first piecetailors the user experience.
(26:36):
And then the second one,how do you engage with it?
So, the general flow, and we're tryingnot to upset the cart too much in terms
of how directors engage, but allowdirectors to choose their own journey.
So, like I said before, mostdirectors get the BoardPAC.
They read it to some level, some allthe way down to every bullet point, and
then some will skim it or quote unquoteskim, which is basically, relying on gut.
(26:59):
And then we use that process andallow you to tailor it slightly.
So, the first step is you get in,here's your prepared company dashboard
based on your topics of interestof what's changed since the last
meeting, what are key red flags in thefinancial area, risk area, compliance,
what are your competitors doing?
Give me a 10-minute scan ofwhat I should be interested in.
(27:20):
Okay, cool.
Now, when I get to reading the BoardPAC,I have a lot of context on where
I want to focus my time, and alsocontext on the meaning behind some
of the things that are being said.
So, the next step is then obviouslyreading your BoardPACs, we align them
all to agenda items, and then we havethat 24/7 context as you read it.
So, you're reading through adocument, here's a project.
(27:42):
We actually put a red flag next to thatfor you, "Hey, this project's moved three
times in the last three board meetings,which you might not be aware of. The due
dates changed by six months and now, it'sat risk of completely missed delivery."
So, as you are reading the packet, youdon't have to go through all documents.
It's all there for you.
So, that's the second interms of how you engage.
Then the other one is form factor,and we're seeing a lot now is the
(28:05):
typical ways you read documents,but a lot more is, do you want to
actually speak to your documents?
I'm going for a walk, I'm sittingin a cab, I'm going to have a
conversation about my BoardPACs.
And that's kind of this next valuepoint of AI is having to be able to
have a live conversation at depthon topics and across your BoardPACs.
So, like we're having right now, I couldjust discuss with Tutaki the BoardPAC, ask
(28:30):
questions about the last three months andgo into a deep conversation about it in
a cab, and that's a new kind of avenue.
And then the third is, at whatlevel do you want to engage on it?
And like I said, you can read the wholeBoardPAC, you can jump from the dashboard
to your run sheet, which goes these arethe key questions that you've highlighted
as important for the meeting and thenall of that is tailorable to you, so
(28:51):
you can run all of it or some of it.
We know that the process foreach director is very different.
So, that's basically the userexperience and the button of
make me smarter hopefully isembedded throughout the process.
We can't tap into your brain, butbasically, if you think about the core of
it, it's help me ask the right questionson the right stuff with the right context,
(29:11):
and that's what we're trying to do really.
Yeah, that's a good way to think about it.
So, one thing we did when wetalked before, and this was a great
conversation that we had, which is,what does the future look like for AI?
And Raza mentioned we have the OnBoards AI Board Advisor Maturity Model.
(29:33):
So, on this model, here'swhat we talked about.
Level one is AI isassisting one board member.
Two is it's assisting the board.
Third is its AI is in the boardroom.
Next, AI is a non-voting board member.
(29:55):
Next level is AI is a voting board member.
Next is AI is the dominant voiceon the board, meaning there'd be
more AI than human in the room.
And the highest level, level seven is.
AI is the only voice on the board.
You get rid of the human elementand all the things that people
(30:17):
forget and mistake, and you reallyjust clean up the whole thing.
Now, I'm sort of kidding aboutthat, but the reality is, that
seems like a logical progression.
It seems like there's so much tosay about the advantages as AI
develops and as you develop thismodel, it's going to get better.
(30:39):
It's going to make fewer mistakes,it's going to understand better
what a board member needs to do.
All of the things that you trainboard members to do, AI will
remember and it won't forget them.
So, on this AI Board AdvisorMaturity Model, we're someplace at
(30:59):
number two, AI assisting the board.
I think that's where Tutaki is.
Do you think we'll get to levelseven anytime in our lifetime?
It's funny you asked this.
We did a talk and there wereabout 150 directors there, and
a similar question was asked.
It was like, "Am I going to be replacedat some point?" The scary thing is
this is every nature of our job, right?
(31:22):
And I think where it lacks nowis there's a lot of context that
sits outside of documentation.
You have a conversation betweentwo people over a coffee, and that
becomes context for then how youthink in a meeting and what questions
you ask and how you answer them.
If you had an AI tool listening to everyconversation you had, every email, every
document for the last three years onthis board, I guarantee it would be a
(31:45):
pretty damn good director and it would beborderline being able to sit on the board.
So, until we solve that informationgap, I think it'll be difficult, but
I think it will get there and I thinkthe adoption curve will look a lot
steeper in these smaller organizations.
Myself, I would love to have afull board there that I can share
my BoardPAC with and go grill me.
(32:06):
Give me all the questions thatyou've got so that I can answer
them and tailor my material beforeI get to the actual board meeting.
Like the value of that is crazily high.
And then for a board meeting, youmight see it similar to Wimbledon
that's just had AI line judges.
There are a couple of kinks they needto work out, but it might sit in the
board meeting and be kind of that devil'sadvocate or give an opposition view on
(32:27):
certain topics so that you don't havethis groupthink element and can actually
discuss things in a different way.
So, I think that'll be the starting pointand then when we get to this step seven
of AI, everything, I think we'll be at thepoint in time where we've got AI boards
talking to AI companies, talking to AIconsumers, so that'll be completely a
(32:47):
paradigm shift in how we approach that.
Would there be a non-human voting member?
That's actually a moreinteresting question.
Like even one vote from AI boardmember makes it very, very interesting,
and that can be foreseeable in maybeless timeframe than the level seven.
(33:09):
But it would be fascinating to seewhether the law and legal frameworks
and all of those catch up to itand say, "Oh, yeah, you are allowed
to have one voting AI member."
And we talk about independentdirectors and non-independent
directors, the argument would be thatthis AI is even more independent.
Really, it's not beholden tothe CEO or to the board chair.
(33:33):
It doesn't care about compensation.
It doesn't care about beingrenewed for another term.
It doesn't care about, did itembarrass itself at the last meeting?
It's truly independent, notperfect, but independent.
And I could see that first vote, asyou said, I agree, that's the most
likely step that we're going to go.
(33:54):
That's a really great point, Joe.
That would be the prediction thing.
The independent tie-breaking,odd number director.
Jamie didn't take me out on his yacht lastweek and try to influence me on my vote.
I didn't go out to dinner at someguy's castle on a Greek island.
It's just, I looked at the stuff,I read all the emails, I saw every
(34:17):
board minute for the last fiveyears, and this is what I think.
And I have the humanities expert knowledgeon this industry on my fingertips, and
I'm saying, I will vote this or that.
Yes.
So, I would not be surprised if ithappens even in my lifetime, but I'm
positive it'll happen in yours, Jamie.
(34:37):
And I think the other thing for thefuture of AI in the boardroom, I think
what you mentioned that management hashad now to do a lot more disclosure and
provide a lot of materials, and I thinkwhat you alluded to as, well, even for
management, that's actually quite a bitof work to come up with these materials,
(35:00):
and you could foresee a set of tools thatcan gather all that primary sources of
information, as detailed as it can be inthe company down to, as you mentioned, the
transactions in the accounting system, butbring it up into a manner for the board
that is much more readable and so on.
(35:21):
So, hopefully, that also happenssoon because we are indeed inundated.
I hope that happens right away.
That's something everyone'sgoing to welcome.
Yeah, I mean, I was talking with oneof our customers who wants to use it
for the exact purpose across theirmanagement team's sub-teams, they
spend up to two weeks a month and it'slike a million dollars of people time.
(35:43):
Every month they do that, and thenby the time the board meeting's
over, they're starting the next one.
So, from a cost point of view, it'smassive, and these documents are huge.
So, when a director gets them,it's hard to actually digest them.
So, you've got arguably the moststrategic group in the company
struggling to ingest this stuff.
And arguably the most useful operatorsin the company spending half the year
(36:06):
creating those documents, which arehard to read, so it's a massive pain
point for the organization to solve.
Absolutely.
Jamie, this has been afascinating conversation.
Thank you so much for joining us today,and thank you all for listening to On
Boards with our guest, Jamie Green.
Please visit our websiteat OnBoardsPodcast.com.
(36:28):
That's OnBoardsPodcast.com.
We'd love to hear your comments,suggestions, and feedback.
If you're not already a subscriber, pleasebe sure to subscribe at Apple Podcast,
Spotify, or wherever you get your podcast.
And remember to leaveus a five star review.
And please tune in for thenext episode of On Boards.
(36:48):
Thanks.
Popular Podcasts
My Favorite Murder with Karen Kilgariff and Georgia Hardstark
My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
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