November 26, 2025 • 40 mins
In this episode of The Evolving Leader, Jean and Scott sit down with Andy MacMillan, CEO of Alteryx, to explore how generative AI is reshaping work, leadership and organisational design. Andy shares a uniquely practical and grounded view of where companies really are with AI adoption, caught in a chaotic but exciting period where the opportunity is clear, but the scaffolding is not yet in place. He explains why AI transformation cannot be delegated to IT, how leaders should approach reimagining business processes, and why Theory of Constraints is one of the most powerful lenses for navigating the next wave of change.
Andy also opens up about the leadership challenges ahead: maintaining psychological safety in the midst of rapid technological shift, helping teams build confidence with new tools, and avoiding both complacency and panic. He shares candid lessons from his own leadership journey, the importance of transparency when organisations face change, and why the most impactful AI practices often happen at the “strategic altitude” rather than in day-to-day automation.
Further reading re. Andy MacMillan and Alteryx:
· How AI adoption is driving a new data era at Alteryx — An interview with Andy MacMillan discussing his role as CEO (appointed December 2024), Alteryx’s repositioning as an “AI Data Clearinghouse”, and his thoughts on shifting from siloed business-data systems to unified analytics.
· Alteryx Looks To Become An AI Data Powerhouse With New Unified Platform (CRN, May 2025) — MacMillan discusses the launch of the “Alteryx One” platform, a strategic move to unify analytics, data-prep and AI workflows under one roof.
· With agentic AI we are being sold on the idea of running before we can walk or crawl, says Alteryx’s Andy MacMillan (Tech Monitor, May 2025) — A leadership-focused piece where MacMillan emphasises the importance of starting with manageable AI projects, surfacing his mindset around experimentation, governance and human-machine synergy.
Other reading from Jean Gomes and Scott Allender:
Leading In A Non-Linear World (J Gomes, 2023)
The Enneagram of Emotional Intelligence (S Allender, 2023)
Social:
Instagram @evolvingleader
LinkedIn The Evolving Leader Podcast
Twitter @Evolving_Leader
Bluesky @evolvingleader.bsky.social
YouTube @evolvingleader
The Evolving Leader is researched, written and presented by Jean Gomes and Scott Allender with production by Phil Kerby. It is an Outside production.
Send a message to The Evolving Leader team
After more than 5 years of hosting The Evolving Leader, Jean Gomes and Scott Allender are launching their new show The Mindset Economy in January 2026. The new show will explore how to live better, work smarter and be connected in a more uncertain world where machines can think.
Available wherever you get your podcasts.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Jean Gomes (00:13):
What does a workplace look like when AI
becomes truly pervasive as wegrow more comfortable using
these tools for far more than anenhanced search engine, many of
us are beginning to see theirpotential to unlock more
meaningful work, sparkcreativity and amplify our
impact. Yes, automation willtake over routine tasks, but
(00:36):
that doesn't necessarily mean adystopian future. Instead, it's
a call to proactively imagineand design a better, more human
centred world of work. In thisepisode, we speak with Andy
McMillan, CEO of AI servicescompany altrex, about how to
think in this forward lookingway and how he's reshaping his
(00:59):
own organisation to make itreal. Join us one last time for
an essential conversation on theevolving leader.
Scott Allender (01:08):
Hi friends.
Welcome to the evolving leader.
The show born from the beliefthat we need deeper, more
accountable and more humanleadership to confront the
world's biggest challenges. I'mScott Allender
Jean Gomes (01:18):
and I'm Jean Gomes.
Scott Allender (01:20):
How are you feeling? Mr. Gomes,
Jean Gomes (01:21):
I'm feeling excited about the end of the year.
That's what I'm feeling. I'mfeeling it's been a really full,
full on year, with so muchhappening, so much happening in
the space that our guest isgoing to bring into this. And
I've been really immersed inimmersing myself, trying to
keep, you know, some somewhat,you know, wouldn't say ahead,
(01:42):
but somewhat in the game onthis, and really excited about
next year as well. So I'mfeeling more positive probably
since we started the evolvingleader, the level of kind of
clarity about where we're goingis great. And so yeah, I'm
feeling I'm feeling great. Howabout you? How are you feeling?
Scott Allender (02:02):
I'm feeling a mix of things today. I'm feeling
grateful that I've got some restcoming up here in about another
month, looking forward to somedowntime, already feeling really
energised about the next year,though, lots of good things on
the docket for the start of theyear, so I'm feeling really
optimistic and excited aboutthat stuff and and feeling a mix
(02:26):
of things about our show here.
As you know, we've been doingevolving leader for five years
now, and it's been incrediblyenriching and successful, and
now we're getting ready to pivotand launch our new show, the
mindset economy, in January 2026so lots to be excited about,
lots to be grateful for, and I'mreally grateful for our guests
(02:48):
joining us today, because it's agreat conversation to land
things with on the evolvingleader today, we're joined by
Andy McMillan and is a seasonedtech leader who has recently
taken on the CEO position ofAlteryx, an AI platform for
enterprise analytics used by8000 corporate customers
globally. Last year, Alteryxbecame a private company
(03:08):
following a $4.4 billion dealthat resulted in its delisting
from the New York StockExchange. So we have lots of
questions for Andy, and we'regoing to jump right in. But
first things first, Andy,welcome to the evolving leader.
Andy MacMillan (03:24):
Thanks for having me, guys, and congrats on
a heck of a run. Five years youguys have had incredible guests
and a great track record. Solooking forward to what's next
as well, but a privilege to bethe Capstone here on this great
run.
Jean Gomes (03:37):
Thank you, Andy, how are you feeling today?
Andy MacMillan (03:40):
Good. Feeling good. It's early for me in the
morning and getting ready to getgoing. Yeah, things are good.
Jean Gomes (03:46):
Cool. Can we start with you setting the scene for
what altrix does for itscustomers.
Andy MacMillan (03:52):
Sure, altrix has been around for over 20 years.
The company started by helpingpeople on their desktops work
with spreadsheets in moreeffective ways. So you can
imagine lots of people everymonth. Maybe you're doing
something like budget versusactuals comparisons, and you got
to pull two spreadsheetstogether and do a whole bunch of
work to kind of smash all thatdata together and make it work.
(04:13):
We helped people do that andthen automate that. And then as
we've grown and grown, westarted helping people do that
sort of an enterprise scale. Andso a lot of businesses have
essentially data automationsthat they run on Alteryx. And
now more and more, we're seeingpeople build those kind of data
workflows to power AI agents andllms and things like that. So
(04:33):
you can imagine, if you wantedto write some kind of cool
custom GPT or something thatwould maybe help you calculate
sales commissions or help youunderstand your travel expenses,
the first step would be, I haveto go pull all that data
together and have it make sense,and then be able to hand it to
the GPT. And so a business usercan do that in Alteryx. It's not
a coding platform. Think of itas being for somebody who's good
(04:55):
at Excel, not somebody who wantsto spend their nights and
evening writing Python code.
Yeah.
Scott Allender (05:02):
So I'd love to kind of start with getting your
sort of view on the sort ofstate of the nation, assessment
of how organisations arecurrently responding to genitive
AI, because. Getting a lot ofdifferent reports and
perspectives, and I'd love toget your insights here.
Andy MacMillan (05:19):
I think it's pretty chaotic, but also
exciting, and hopefully that'sthe kind of world people enjoy
living in, because, like, that'sthe kind of world we live in
now. I think the pace that wetake on change now in business
is completely different than,say, 20 years ago. I think the
kind of enabled by internet andmobile, all of a sudden when
(05:41):
things happen, they kind ofhappen in a snap, right? And and
so when something likegenerative AI comes out, I think
people can see the potential. Idon't think you have to spend a
lot of time in chat, GPT orGemini to sort of go, Wow. This
is different. This is somethingnew. And to think about how you
can apply that to your businessor your life. But then you have
the challenge of, okay, but allthe all the scaffolding is not
(06:02):
there yet, all theinfrastructure is not there yet.
However, I know that this isgoing to happen fast. I learned
from the last couple of bigwaves. This is going to happen
quickly. And so I think peopleare, are sort of scrambling. And
I mean that in a positive way, Ithink that's people leaning in
and trying to figure it out. ButI think we're kind of in the
scrambling phase right now aspeople try to move quickly, but
(06:22):
maybe don't like I said, haveall the scaffolding in place yet
to really know how to how toscale and find value, find ROI
and so that's the sort ofchaotic moment we're in right
Jean Gomes (06:33):
now. Given where on this, you know, there's a lot of
noise and kind of counteringviews and some fairly hysterical
kind of commentary on on AIadoption, when you are talking
to senior leaders, when you'retalking to CEOs and C suite
executives, what's your kind ofsense of what's most helpful for
(06:54):
them in building thatscaffolding and actually making
practical steps forward so theydon't fall into the traps,
perhaps, of some of the digitaltransformations of the past, I
think
Andy MacMillan (07:04):
one is looking for an outcome you're trying to
drive in your businessgenerally, and applying AI
towards it versus, you know, Igot aI I'm running around trying
to find what I'm going to dowith it. Can run people in a lot
of different directions. So Idefinitely think that's one. I
think the other is thinkingabout it systematically. I think
of AI as being this veryinteresting constraint remover,
(07:27):
and when you remove aconstraint, new constraints
arrive. And so I don't think itmeans that things scale
infinitely. But, you know, Ispent a lot of the time in my
business, thinking about, youknow, what if we can write code
10 times faster? Like, how willwe organise our teams
differently? Maybe my teams seemto be located closer together,
like, what changes about thebusiness? And so I think one,
you know, pointed at realproblems, and think about how to
(07:49):
work backwards towards solvingbusiness metrics that you
already care about. And I thinkthe second is, when you do that,
think about it as sort ofbusiness process reimagination
with different constraints. Ithink when we limit it only to a
personal productivity tool,you're not really reimagining
the business process. And ifdifferent people are going to be
(08:09):
different levels of newefficiency, you're going to have
to rebalance your teams andfigure it out, and you're right
back to business processreimagination. And so I think
for me, those two things are thekey pieces to really trying to
drive value when you're tryingto apply AI to your business,
not just sort of play aroundwith or learn a bit about AI.
Jean Gomes (08:25):
That feels very different from some of these
digital transformations in thepast, where it kind of gets
driven down to the digitaldirector or the head of
transformation or the ITdepartment. This sounds like
it's a much more crossfunctional leadership required.
Andy MacMillan (08:40):
I think you can think about any aspect of a
business and imagine how AIchanges it. And I would posit
that the people that willultimately think about how to
change those parts of thebusiness will be those people in
that part of the business. So Iagree with you. Whereas digital
transformation, I think,initially started off very
technology driven, you know, oh,we need to stand up a website
(09:01):
and E commerce and things likethat, and we turn to our
technical teams to do that. Youthink of where that ended. We
have handed most of that stuffback to the business, right?
Most websites are, you know,yeah, the IT team runs the the
infrastructure, but themarketing team owns the website,
or the the merchandising teamowns the E commerce site. I
think with AI, you know, everyleader in every functional area
(09:24):
is going to have an opportunityto look at the thing they're
running and try to reimagine,how would I do this with this
new capability that I've neverhad access to so you're going to
be in the Office of Finance andthinking about, well, how would
I, how would I manage Travel andExpense differently? If I have
access to a real time agent,you're going to be in the legal
team and say, how would I run mylegal team differently if I can
(09:44):
red line a document using AIbefore I ever have my lawyers
actually look at it? And I thinkthat's not an IT problem. I
don't think that's going to, Imean, it is going to help.
They're going to provide againsta lot of that scaffolding. But I
think unlike these othertechnology transformations, I
think it's pretty easy toimagine a universe where, no
matter what function you're in,you need to be thinking about,
(10:07):
how does this new capability?
Forget about as a technology,just this capability. Let me
scale what I'm doing and.
Entirely differently, and Ithink that's where we're going
to see people over the next yearor two, really figuring out
almost function by function.
Like, what does this mean if Iwork in accounting? What does
this mean if I work in supplychain? What does this mean if
I'm in sales? And I thinkanybody who's not thinking about
(10:29):
that risks sort of having theirskill set get left behind.
Scott Allender (10:33):
What about beyond a year or two. Take take
me five years, 10 years down theroad. What do you think an
automated workplace looks
Andy MacMillan (10:41):
like? Yeah, I think companies look a lot
different. And I think again, ifyou if you imagine it from a
Theory of Constraintsstandpoint, and again, I think
it's helpful to almost thinkabout it by function. I'll give
you an example in softwaredevelopment over the past 20
years, one of the big trends hasbeen offshoring. And the reason
is, the main constraint whenyou're building software is how
(11:03):
many developers do I have accessto? Right? If I more developers,
I can write more software. Andso you would have your designers
and your product managers oftenclose to customers the biggest
markets, often the US. So thosepeople might be here in the US,
they would design things. Theywould send them offshore to a
lower cost location where youhad a lot of engineers. And they
would do these, you know, twoweeks sprints. So if you're not
in software development, a twoweek sprint is two weeks of
(11:25):
writing the code, and then yousort of pop up two weeks later
and go, Hey, here's a here's ademo of where we're at and how
it works. And you sort ofiterate. Imagine now if that
sprint only takes an hour? Well,I don't want to have that team
now on the other side of theworld, I might want to have them
really co located, iteratingwith that team. And so again, I
don't know if that becomes thenew constraint or not, how fast
(11:46):
I can iterate, but how manypeople I hire, where they're
located, how they work together,what that job entails. What does
it mean to be a softwaredeveloper if I'm using coding
assistant tools. So that jobchanges too. So I can imagine,
in five years, the entire modelaround software development
doesn't go away. I don't thinkit's all the robots doing it
like I don't think that's,that's the you know, I think
we're headed for that dystopianfuture. I think we are headed to
(12:09):
a place where we'll write bettersoftware faster. We'll do it. I
think with teams, it can iteratemuch more quickly. And so that
makes me start rethinking, well,how do these people work
together, and where are theylocated, and what are the new
constraints? Is the newconstraint, how fast they can
work versus how many developersI have? I don't know. I think
about that with with salespeople, we have a whole bunch of
(12:31):
people. We surround sales peoplewith that try to make them more
effective. We have inside salespeople, and we have renewal
managers and all these otherthings. And you start to wonder,
well, what parts of thosefunctions can AI help the
salesperson do more effectively?
So do I hire more sales peoplebecause they're more effective?
Do I have to hire fewersalespeople because they can get
more done? It's going to change.
(12:51):
And so I think leaders are goingto have to come at every part of
their business. And I think aCEO is having to work with every
one of their functional leadersto kind of come at it with an
open mind. And again, reallyoperate from like a first
principles. Again, I reallythink Theory of Constraints is
very helpful. When you thinkabout AI, it will just change
what the constraints are. And sothat means you different ratios
(13:13):
of roles, different skills inthose roles, and so I think
businesses will look a lotdifferent in five years.
Jean Gomes (13:20):
Let's pause here and on this theory of constraint
idea, because that's a veryhelpful lens to think about the
underlying assumptions that areplaying out in in in how your
business processes, businessmodel and so on is developing.
Can you walk us through anexample? How? I mean, you've
given us a few high levelexamples, but how you think
(13:41):
about applying that and how newconstraints start to emerge? How
do you what's the if you wereguiding a team thinking about
how to apply that thinking in amore detailed way? Can you just
walk us through an example tohelp us?
Andy MacMillan (13:56):
Yeah, I spend a lot of time with all of my teams
trying to think through, if Iwant this team to go faster, to
be more efficient, whatever thegoal is of the thing that I'm
working on, what is the onelever that changes the most?
Right? So if we were going todouble our sales, what's the
first thing everybody think ofthat changed? Was it that we had
twice as much pipeline? Was itpipeline? Was it that we had a
(14:17):
twice the win rate in ourcustomers? And simple questions
like that can lead to, what isthe core constraint? The fun
thing about constraints is, assoon as you fix a constraint,
something else becomes theconstraint like that's just the
way that works. But you startworking through what are the
biggest problems? And I thinkthat's a useful way to think
about rolling out AI, if I canapply AI to my biggest
constraints now, it starts toreally open up my business to
(14:40):
scale and do new things. Versusagain, I'm running around with
AI, you know, it summarisesstuff. Well, what can I go
summarise like that's not aparticularly interesting
business problem to go solve.
But when you think aboutconstraints in a business today,
very often, constraints comedown to head count. It is one of
the biggest expense items in alot of businesses, certainly a
lot of kind of software stylebusinesses. And if I start
(15:02):
applying AI to make peopledramatically more productive in
certain roles, and maybe that'snot as dramatic in other roles,
again, those constraints start.
To shift. My ratios start toshift. So if I think of
something, yeah, like my go tomarket team, it's very ratio
driven. You know, I have allthese regional teams, and in
each region, I have, you know, Xnumber of sales people, and I've
(15:23):
had X number of sales people. Ihave, you know, y number of
field marketers, and I've got,you know, z number of of
technical sales specialists thatare helping the sales people.
And so you look at that, that'sa pretty finely tuned machine.
It's a it's a pretty common inmy industry sort of software go
to market motion. I would guess,if I met with 10 other software
CEOs, our ratios look prettysimilar, you know, maybe a
(15:44):
little different here, a littledifferent there, but for the
most part, pretty similar. Andnow imagine AI makes one of
those groups three times moreproductive. It makes one of the
ones only 50% more productive?
Do I now change all the ratiosof how these teams work? Am I
hiring kind of a differentworkforce at that point? How do
they work together? That's goingto change. And so that's really
where I start to think about theimpact. It's not all our jobs go
(16:06):
away and the robots do all thework, it's more well, it's going
to change how we work togetherand what work we expect people
to get done. And I think that,to me, is, is interesting. I
think the other one, I think alot about, is there's this
debate, I don't know which sideof the spectrum I'm on, of, you
know, with AI do all entry leveljobs go away because the,
because the AI does all the theentry level work, or if the flip
(16:28):
side, you know, I can give AI tosomebody in the early part of
their career. And it's thisamazing amount of expertise that
can sort of apply to problems.
To problems. And so I think justthat friction sort of shows I'm
not sure the answer is one orthe other. I think I go back to
like, well, some barely, reallysenior roles might scale better,
because you can give them accessto AI. Maybe if my entry level
(16:48):
people are now way moreproductive, because I can give
them AI, might hire more entrylevel people too. I think a lot
about when something's betterand cheaper, people tend to buy
more of it. And so I also thinkabout employment that way. If
what I get for my dollar is asenior executive when I hire
someone who's way moreproductivity, I might hire more
people rather than just hiringfewer people, because I've got
(17:10):
productivity gains. So those arekind of the three things I think
about when I reimagine whatbusiness looks like over the
next couple of
Scott Allender (17:23):
years, what are some associated and emerging
leadership challenges andopportunities for inspiring and
motivating you know, economicactivity in the sort of new
world and the sort of new way ofdoing business.
Andy MacMillan (17:40):
I think a big one is acknowledging how much
change is coming. I think peopleget it wrong when they start
telling people, like, you know,don't worry, nothing's going to
change. It's like things aregoing to change. I told my whole
team, like, my my engineeringteam, for example, is starting
to use a lot of these codingassisting tools there. I have a
great engineering team. It'sobviously a little bit nerve
wracking when you start going,like, wow. Like, this, this
(18:01):
thing can write half the code Iwould have written before. Like,
that's an unsettling feeling.
And my advice to leaders is is,you know, you lean into that.
What I've been telling my teamis, let's take up these tools.
We've got a lot of stuff to gobuild so we're not maxed out on
capacity like I would. I wouldtake if you told me I got 1000
more developers at no cost, Iwould take them in a heartbeat.
So there's a lot of room togrow, but for themselves,
(18:22):
personally, your career strategyat this point can't be I'm going
to hope my current employerdoesn't know or use AI, and
maybe for the next 10 years,I'll just find other employers
that don't know about or use AI,like, that's not a career plan.
Yeah, and so, so sort ofembracing that with your team.
Hey, we're on a missiontogether. This change is
happening. You're going to workat a place that's going to
(18:44):
support you going through this,and we're going to learn as we
do it. And, you know, embracethat this change is coming. I
think that's really important.
And then to be listening, I'mnot also just directing my team
go do this, and I don't want tohear about the results. I'm
trying to lean in and learn howare each of these functions
changing? What concerns dopeople have as these ratios
(19:05):
start to shift, starting tothink about, how are we going to
rebalance our teams? And ourmessage internally has been,
we're going to slow hiring alittle bit right now. And what
I'm trying to do is create spacein our business. So as we
rebalance, we can apply the headcount as we're rebalancing. So
trying to be really upfront withpeople, like, yes, there will be
change that comes. We can kindof both manage that change as a
(19:26):
leadership team, so that all theimpact isn't just hitting the
front lines as we, you know,hire a bunch of people and go,
Oh, oops, we should hire themover here. Like nobody wants to
work at a company that doesthat. Does that. So sort of
acknowledge, yep, we're being alittle a little cautious with
hiring, until we see, kind ofsome of the dust settle on some
of these ratios, but we're goingto lean in. And I think that's
really important. I think peoplewant to feel led through change,
(19:49):
not that change is being managedfor them. I think a lot of
leaders make a big mistaketrying to obfuscate change and
challenge trying to obfuscateinto the change is coming. And I
do the opposite. I get in frontof a say, this is happening. You
read about every single day,every conference you go to, you
know this is going to be a placethat sort of helps you go
through that change.
Jean Gomes (20:10):
How are you I mean, I love the kind of. Spirit that
you're creating in in thisbecause, you know, psychological
safety disappears when peopleare in in an environment where
the technology is threateningthem in that way. And that's
always been the case, but nowyou kind of a whole new level of
threat, because it's, it's it'sintelligence, not just
(20:30):
automation. What? What are youdoing to help people at a
practical level, to be able toquickly lean into all of these
things. What have you learnedabout what works?
Andy MacMillan (20:44):
Well, part of it's giving people access and
opportunity to play around andwork with this stuff. So I did
this, not only here, but at mylast company, very early on. We
licenced an LLM non training,you know, private use, but
company wide, gave people accessto that, gave them training. So
we run a regular training onhow, what are the rules for
(21:05):
using this? How can you use it?
You know, what data can you putin it? What can you be doing?
We've been showcasing on our allhands on things that people have
built using AI and so, reallytrying to make it a place to
sort of learn and let peoplesort of evolve and build stuff.
So I think that's a big part ofit, and sort of celebrating
that, and again, just continuingto to highlight, you know, for
us, we're trying to efficientlygrow our business. You know, as
(21:27):
you mentioned in the theopening, you know, we just went
through a pretty substantialtake private we did a financial
restructuring of the company,got us to a very healthy place
on our balance sheet, and wewant to drive even more growth.
And so I've been telling thecompany, this is the perfect
place to apply AI. We want todrive growth while we manage
having a strong balance sheet.
(21:48):
So from the top level of thecompany, I've given a vision of
how AI is going to help us dothis in a way that brings
everybody along, I've then triedto provide them the tools to
start to, frankly, play with thetechnology. And I've told them,
you know, you are free to startapplying this to how you do your
job, and let's start sharingwhat works. And so I think those
things are all part of it, andit's just a continual topic. I
(22:09):
think this is something thatcomes up on almost every one of
our all hands. I write a weeklynote to the whole company. Every
Sunday night, I just write alittle email. I talk about AI
transformation quite a bit inthose emails, what I'm doing,
what I'm seeing other people do.
I built a little GPT myselfusing Alteryx to prepare some
data for something I was workingon. And I was like, I'll just
show this to the company. So onone of our all hands, I
(22:30):
literally just got out my littlething, and I'm like, here's what
I'm doing. And, you know, andjust trying to let people know,
you know, this is kind of howyou learn. You just sort of
iterate and play
Scott Allender (22:39):
with stuff. When I echo John's sentiments about
the sort of spirit that you'resetting and the tone that you're
bringing, and your sort ofintentionality, which is really
inspiring as you talk about thisstuff, I'd like to take a slight
tangent and kind of hear abouthow you got here, like, what are
some of your sort of mostimportant leadership lessons
(23:01):
that has readied you to leadthis organisation and prepare
people for the future?
Andy MacMillan (23:06):
I think one of them, Scott, that stands out
specific to kind of thisapproach was, you know, when you
when you become a hired CEO. Onething I tell people, very rarely
does somebody say, hey, thisbusiness is running perfectly.
Why don't you come run it? Sothere's usually stuff to go in
and do. And one of the things Ifound earlier, one of the first
(23:27):
companies that I joined as CEO,was the more people I openly
shared problems with, the morepeople wanted to help with those
problems. And I always feltlike, wherever the line got
drawn, like, if I you know that,I think I started too small with
who knew about what we weretrying to do. There were some
real problems in the firstbusiness I joined. I mean real,
real issues. And those issueshadn't been shared widely in the
(23:49):
company. And I sort of foundthat as those as I started to
bring more people, kind of intothe tent. People were like, Oh,
yeah. Like, actually, that makessense. Why things have been
like, the like, how do I help?
But then wherever the line hadstopped to the next level was
like, what's going on? Like, I'mreally frustrated. Why are
people and so I've just foundthe way to get everybody working
on something big change, orwhatever is, to bring as many
(24:10):
people as possible into theprocess. And if you're really
open with people again, inpositive I'm a naturally
optimistic person, so I'm alwayslike, Okay, well, here's the
problem, but we're going to fixit. You get people helping you
fix it. When you try to fix theproblem, before you tell people
about it. And I think this canbe something leaders try to do.
You think, oh my gosh, if, ifeverybody finds out that we have
this big challenge, you know,people will quit, people will
(24:31):
leave. You know, people can go.
They'll vote with their feet,and so then they're on the
outside, looking at theleadership team, going, seems
like something's not working.
Nobody's telling us what's goingon. They must not know what to
do. That's unsettling. Versus,hey guys, we've identified the
problem. Here's what it is. It'sa big problem. We're working on
it, and let's figure it out.
(24:52):
Most people, even if, like, Idon't know this might be really
hard. They sort of grab onto therope and they start pulling, and
people like feeling like they'rehelping work on the thing that
you're working on. So I thinkthat's one of the big leadership
lessons for me. I thinkthroughout my career, I came up
through the product managementside of the world as a
developer, early on in mycareer, I've always liked the
problems the company works on. Ijust find that interesting. And
(25:15):
so I think one of the thingsthat served me well as a leader
is, strangely, I never managedfrontline people. My first
leadership Job was a VP leveljob, and I managed people that
were 10 years older than me, andI sort of got this big like
career accelerator. So So one,I've never been a micromanager,
in part because I've never hadto lead people in the very first
stage of their career, which iswhere I think you have to do the
most hands on management. SoI've gotten management. So I've
(25:37):
gotten to be kind of a littlebit of a senior leader, leading
leaders early. And then theother is, while doing that, I've
always enjoyed leaning intoproblems, but realising when
you're a senior leader, it's notalways your problem to solve
entirely. But I like to sort of,hey, let me I kind of call it
participant leadership, like, Ilike to, hey, we're having a
(25:57):
design meeting. I'm going tocome to the design meeting as
the CEO. I maybe have somecontext of what's going on in
the market and the company, butI'm not running the design
meeting. I'm not the ultimateapprover. I'm just sort of there
to participate, and I reallyenjoy that, and that has served
me well, I think throughout mycareer, is to be somebody who
just likes to, you know, workwith different teams trying to
solve problems around thecompany, but not feel like I'm
(26:19):
the the ultimate solver. I'msort of an enabler.
Jean Gomes (26:23):
Coming back to, you know, the kind of challenges of
embracing AI, we hear a lot ofevidence now coming that it's
making people de skilled. It'screating cognitive complacency.
People's memory is going theycan't remember what they've
created on AI and so on. Whathave you learned? I mean, either
(26:45):
in your own use or in helpingyour your teams to adopt it so
that they get smarter, notdumber, using it.
Andy MacMillan (26:52):
Um, I don't know. I think there's some, you
know, like brainwave, kind offancy studies on all this. But
my analogy is a little bit, youknow, I think they said that
about calculators and computers.
And I'm not really sure wheredumber as a society, because we
use computers. I remember mydad, when I was growing up,
always wanted to teach me to domath on his slide rule, because
that was the way to reallyunderstand it. And I don't think
I ever understood how to use theslide rule, and I don't think
(27:12):
I'm any dumber for not havingfigured that out. So I feel a
little bit like it's a newcapability. How do we learn how
to use it? I don't think thatmeans we don't think anymore. I
think it means, pretty soon, welearn how to work with this much
smarter set of intelligence thatcan help us learn a whole lot
more. I feel like I learn atonne using AI every day. I find
myself even just when I getinterested in something, I can
(27:34):
go into chat, GPT or Gemini, andsort of work through a problem
in such depth, where I learnedso much. So I'm not entirely
convinced that this is somethingthat's going to make us dumber.
I think it's something we'regoing to have to learn to work
with differently. It mightchange how we operate. I would
argue maybe the advent of like,my kids just get to use
(27:56):
calculators, even when they takethe AC, T and stuff. So, yeah,
maybe they don't do off the topof their head arithmetic as fast
as my generation did. I don'tknow that that's going to set
them back, but, you know, maybeit changes. So I think that's my
my mindset is, I don't thinkthis is a replacement for
cognitive capabilities. I thinkit's a new skill that we're
going to learn, and I think it'sgoing to mean that people can do
(28:18):
more things more quickly, and dothem better. I think about the
even around the house, like Ican go into Gemini and get help
on doing a home repair. I'mterrible at home repair. I'm a
lot better with Gemini. So doesthat mean I'm dumber or smarter
that I know how to use Geminibefore I start, you know,
smashing holes in the walls. Ithink it makes me smarter than I
know how to use it, but I sortof get the argument of, like,
Well, my brain didn't figure itout. Like, Well, okay, but my
(28:41):
brain wasn't doing a very goodjob figuring out home repair
before I started to have
Jean Gomes (28:46):
access to this. So I think, you know, I think it
probably comes down to how youuse it, and as a, as somebody
who's, you know, grown up as a,as a product manager and
probably has a quite goodunderstanding of how to think
about solving problems. Itreally helps you. I'm, I'm more
concerned about, you know, howpeople who don't know how to use
it are sort of relying upon itlike a search engine to get easy
(29:09):
answers, who don't dive into thedetail, who don't use it to
think together, you know, might,might fall into the trap of, you
know, losing their abilities. I
Andy MacMillan (29:20):
do think we're starting to learn what it's good
at and not good at. I mean,we've all had the experience
where you you ask the LLM aquestion, it gives you an
answer, and you sort of go, areyou sure? It goes, Oh, you're
right. This is totally wrong.
How about this idea? And so Ithink we're learning also sort
of what it's good at, what it'snot good at, to your point,
which is, maybe we're not takingeverything always at face value.
Sometimes there's a little bitmore connecting the dots that we
(29:41):
want to be do, doing, ratherthan just, you know, offloading
all of this. I think everystudent I know has learned the
lesson of, you know, oh, I usedone of these tools to summarise
the thing I was supposed to readfor class. And I got to class
and realised I didn't, didn'tknow what was going on, right?
So you, I think there's sort ofa where to use it, how to use
it, that we're all stillfiguring out when to trust it,
when to when to drill deeper. SoI agree with you. I think
(30:03):
that's, that's, again, kind of askill we're learning.
Jean Gomes (30:11):
What are your top tips in using it? Well, I mean,
what? The things that you'velearned about it.
Andy MacMillan (30:16):
I think, as most people say, I think learning
about how to prompt and ask forvery specific things is
important. I'm also learning alot about the responses come
back with full confidence, likethe dial is always set to 211 on
confidence, regardless of howconfident it actually is. And so
you can query into like, Howsure are you, and where did you
get this from? And so I think,sort of, again, that level of
(30:38):
interrogation on things thatmatter, I think, is really
helpful. I think we're seeingmore of this now getting cited.
So I'm finding now often, if Iask, maybe my personal life, a
more general question, and nowit shows a little bit of its
work. And sometimes we'll showsources, I'll find myself
clicking through the sources. Soit's like, okay, like, yeah,
it's summarised for me. Andeverybody felt like, Oh, this is
(31:00):
going to be the death of, youknow, search engines with click
through results. And now I'mfinding like, no, no, I've read
the summary, but now I want toclick through, you know, oh,
here's the thing, and there'sthree references. And I'll look
at the middle and go, Oh, thatlooks interesting. And I'll kind
of click through and and read. Ithink that's new. I don't think
I was doing a lot of that, youknow. Six months ago, I was sort
of taking sort of taking it atface value. And so I think those
kinds of things, you know,learning to prompt, learning to
(31:21):
drill in, is is a big one. Theother is it work. I've learned
this pattern of trying to donarrower things. And so what
I'll do is, again, maybe,because I work at Alteryx, I'll
pull together a set of data,right? Like so, for example, I
was analysing my sales pipelinerecently. I went into Alteryx. I
pull in my Salesforce data, I doa little calculation on my
(31:44):
pipeline, and then I hand thatto a narrowly scoped GPT and
say, Okay, you're a pipelineexpert. Help me work through
this. I think that's reallyinteresting, too. You're not
doing this like, broad generalyou know, I think of it as like,
are you asking a smart friend aquestion, or are you asking a
specialist a question? And moreoften, I'm realising, you know,
AI is sort of a smart generalfriend, but I can sort of tell
(32:06):
it to be a specialist. I cangive it a narrower set of data
and a narrower set ofinstructions, and I think I get
real value when I do that in mybusiness work, right? And that's
important, because I think whenyou ask it generalist questions,
you get generalist answers. Andit worked. That's not that
interesting. If I say to to chatGPT, you know, what will my
sales commission be if I closethis deal and it goes well,
(32:27):
here's how companies normallycalculate sales commissions. It
might be something like this.
That's not the right answer.
Like, the answer I'm looking foris, like, what will it actually
be? So again, if I it work, cancreate a small data pipeline of
how we actually do that, and Ican tell it, hey, only work in
this thing. Do this work. Ithink that's really useful too.
So I think also learning tothink about scoping, you know,
(32:50):
in personal life too, like I'vedone that with. I wanted to
analyse some some stocksrecently, and so again, I gave
it a narrow data set and anarrow set of instructions and
asked it to do something, versusjust asking, you know, the front
page of Gemini or something, togo do the work. So I think that
scoping is a skill set that willemerge to
Scott Allender (33:10):
those are good suggestions. Are there any other
sort of, maybe five to 10 minutea day practices for people
listening saying, I've been kindof avoiding this. I don't really
know how to use Gemini or chatGPT the right way, anything that
you'd say. You do this 510minutes a day, and you'll start
to really build your your musclearound AI, yeah,
Andy MacMillan (33:30):
I sort of, I like to think of it as, like,
imagine two very discretealtitudes that you want to
interact with an app. One issort of the, you know, do some
leg work for me, kind of stuff,you know, on the ground cover,
summarise this email kind ofstuff, right? Like, it's great,
helpful, but it's pretty kind oflow level, like, hey, grind out
some work for me. I think thebig unlock is when you start to
(33:53):
think of it more like, hey, whatif I could hire a McKinsey
consultant to look at this?
Normally, I wouldn't. But like,I can. I can ask this thing to
act as an expert thatunderstands audit and finance,
and go through this thing reallyquite fascinating. What you can
get out of it? I've been doingthings like for my last board
meeting, I get my financial sentto me by my CFO, and I can go
(34:14):
into, again, our private nontraining instance of an LLM, and
I can say, imagine you are a BCGor McKinsey consultant advising
a CEO in a PE backed company onthe following board meeting.
What trends in the businesswould you recognise? What kind
of questions do you think theboard would ask? Now, I would
argue, 80% of what came back,like, I do this for a living.
(34:36):
I'm like, yeah, those are thethings. But you get a couple
nuggets here. They're like,okay, that took that took two
minutes, and I got a couple ofit's point. I should think about
that. One of the was a reallyinteresting idea. Didn't come up
my board meeting, but I waslike, I should take that and go
talk to my team about that idea.
And so I think again, using itat that higher altitude, again,
with a scope set of questions,of, you know, hey, act as an
(34:59):
expert in software development.
You know, look at this project.
What would you say are the holesin this project? What other
areas would you advise? It takestwo minutes, and even if you
know, you only get a reallyinteresting answer one out of
three times. That is not a lotof. Effort for a lot of material
that you can get back. And so Ithink that's really useful, I
think especially for seniorexecutives that are trying to
(35:21):
operate across a large set ofinformation coming to you. I'm
finding more and more often I'mtaking something sent to me and
I'm just dragging it into theLLM, and I'm not doing the
summarise for me. I'm doing theHey act in this role and help me
really think through what thisdoes. And then again, you can go
back and forth with it. Hey,that's a really good point. Why
did you bring that up? What areother companies doing around
that, you know, prepare a twopager for me on how I would go
(35:43):
about, you know, looking intothat area, it's a really
interesting brainstormingexercise, kind of at that level
that I think people should domore of.
Jean Gomes (35:52):
And, you know, this is probably one of these kind of
questions or thought experimentsthat comes and goes with every
new trend. But one thing thatcame out probably six, eight
months ago was the idea that atsome point, you know, an
organisation might actually havea, you know, an AI CEO I'm just
(36:13):
interested in, you know, likeyour thoughts on, on that kind
of idea.
Andy MacMillan (36:19):
I mean, someone said recently that AI makes very
confident decisions on verylittle data. And I'm like, well,
that's most of my job. So maybe,maybe they could do that quite
well. You know, I don't know. Igo back to, I think there's a,
there's an amazing capabilityset emerging. But I still think
businesses are going to bepeople figuring out what they
want to go do, why they want togo do it, and sort of using
(36:40):
these tools to think about howto go do that. So I definitely
think you could imagine having aseries of agents acting as
really skilled consultants andadvisors and a whole bunch of
areas in the business, and youmight turn to that more often
than not. I still think there'sa level of accountability that
exists in the job, in a lot ofroles, not just CEO, but also
CFO and things like that. Ithink we're a long way from
(37:02):
somebody saying, you know, thefinancials are good because the
robots signed off on it. I thinkwe're gonna say who signed the
financials to say that. That'sright, a little bit like that.
You know, there was a Waymorecently they got pulled over
here in the Bay Area, andthere's this great video the cop
walking up to the window, andthere's nobody in the driver's
seat, and they're trying tofigure like, Well, what happens
now? Because the Waymo made anillegal U turn, I don't know
(37:23):
that we're ready for that to belike a company that did
something wrong or a companythat needs to make a hard
decision. So I think there's alot of play still in this world
for people having ownership overwhat we're doing, but I do think
the role changes. I think it'llbe hard to imagine being a CEO a
couple years from now and notbeing regularly advised and
(37:44):
interacting with, you know, datadriven AI, I think that's,
that's the reality of wherewe're headed. What's
Jean Gomes (37:49):
the thing you're most excited about for next
year? The
Andy MacMillan (37:52):
thing I'm most excited about for next year is
we have just started to enablein our product the ability to
let people build their own datapipelines and put them directly
into whatever kind of LLMthey're using. And so for me, I
have a whole model I use to runthe business, a whole set of
meetings, cross functionally andstuff like that. And I'm
basically creating customer gptsaimed at every one of those
(38:15):
meetings. And I think I'm goingto live in a universe next year
where before we go to every oneof those meetings, we can have a
report coming to us, generatedby our AI agent telling us all
of the things that we're goingto be focused on in that
meeting. And I think that'sgoing to make life a lot of a
lot more fun. Like, I'm kind ofexcited about the pace at which
I think I can run the businessusing AI kind of blended into my
(38:36):
current leadership model. And Idon't think this is a three year
off thing. I think this would besomething that I'll be able to
do at altrix in the first halfof next year, and I think
that'll be pretty I know I likechange, so it'd be interesting
to see, like, how does thathow's that work?
Scott Allender (38:50):
Thank you, Andy, this was really insightful and a
really practical way for us towrap five years of incredible
conversations on the evolvingleader. But don't worry, folks,
the evolving leader isn't goingaway. It's just evolving. There
is a new multi trillion dollarglobal economy that's rapidly
emerging, spurred bytransformational advances in our
(39:11):
understanding of how our brainsand bodies work and the question
of, what are humans for in anautomated world, how we feel,
how we think and how we see,will profoundly influence our
economic success and our wellbeing and our capacity to
embrace uncertainty. So thisJanuary, January 2026, a new
(39:33):
show is coming to you from uscalled the mindset economy, and
our goal, as always, is to forgea more human world. So we'll see
you there. And until then, I askyou, for one final time, the
world is evolving. Are you? You?
You.
What does a workplace look like when AI
becomes truly pervasive as wegrow more comfortable using
these tools for far more than anenhanced search engine, many of
us are beginning to see theirpotential to unlock more
meaningful work, sparkcreativity and amplify our
impact. Yes, automation willtake over routine tasks, but
(00:36):
that doesn't necessarily mean adystopian future. Instead, it's
a call to proactively imagineand design a better, more human
centred world of work. In thisepisode, we speak with Andy
McMillan, CEO of AI servicescompany altrex, about how to
think in this forward lookingway and how he's reshaping his
(00:59):
own organisation to make itreal. Join us one last time for
an essential conversation on theevolving leader.
Scott Allender (01:08):
Hi friends.
Welcome to the evolving leader.
The show born from the beliefthat we need deeper, more
accountable and more humanleadership to confront the
world's biggest challenges. I'mScott Allender
Jean Gomes (01:18):
and I'm Jean Gomes.
Scott Allender (01:20):
How are you feeling? Mr. Gomes,
Jean Gomes (01:21):
I'm feeling excited about the end of the year.
That's what I'm feeling. I'mfeeling it's been a really full,
full on year, with so muchhappening, so much happening in
the space that our guest isgoing to bring into this. And
I've been really immersed inimmersing myself, trying to
keep, you know, some somewhat,you know, wouldn't say ahead,
(01:42):
but somewhat in the game onthis, and really excited about
next year as well. So I'mfeeling more positive probably
since we started the evolvingleader, the level of kind of
clarity about where we're goingis great. And so yeah, I'm
feeling I'm feeling great. Howabout you? How are you feeling?
Scott Allender (02:02):
I'm feeling a mix of things today. I'm feeling
grateful that I've got some restcoming up here in about another
month, looking forward to somedowntime, already feeling really
energised about the next year,though, lots of good things on
the docket for the start of theyear, so I'm feeling really
optimistic and excited aboutthat stuff and and feeling a mix
(02:26):
of things about our show here.
As you know, we've been doingevolving leader for five years
now, and it's been incrediblyenriching and successful, and
now we're getting ready to pivotand launch our new show, the
mindset economy, in January 2026so lots to be excited about,
lots to be grateful for, and I'mreally grateful for our guests
(02:48):
joining us today, because it's agreat conversation to land
things with on the evolvingleader today, we're joined by
Andy McMillan and is a seasonedtech leader who has recently
taken on the CEO position ofAlteryx, an AI platform for
enterprise analytics used by8000 corporate customers
globally. Last year, Alteryxbecame a private company
(03:08):
following a $4.4 billion dealthat resulted in its delisting
from the New York StockExchange. So we have lots of
questions for Andy, and we'regoing to jump right in. But
first things first, Andy,welcome to the evolving leader.
Andy MacMillan (03:24):
Thanks for having me, guys, and congrats on
a heck of a run. Five years youguys have had incredible guests
and a great track record. Solooking forward to what's next
as well, but a privilege to bethe Capstone here on this great
run.
Jean Gomes (03:37):
Thank you, Andy, how are you feeling today?
Andy MacMillan (03:40):
Good. Feeling good. It's early for me in the
morning and getting ready to getgoing. Yeah, things are good.
Jean Gomes (03:46):
Cool. Can we start with you setting the scene for
what altrix does for itscustomers.
Andy MacMillan (03:52):
Sure, altrix has been around for over 20 years.
The company started by helpingpeople on their desktops work
with spreadsheets in moreeffective ways. So you can
imagine lots of people everymonth. Maybe you're doing
something like budget versusactuals comparisons, and you got
to pull two spreadsheetstogether and do a whole bunch of
work to kind of smash all thatdata together and make it work.
(04:13):
We helped people do that andthen automate that. And then as
we've grown and grown, westarted helping people do that
sort of an enterprise scale. Andso a lot of businesses have
essentially data automationsthat they run on Alteryx. And
now more and more, we're seeingpeople build those kind of data
workflows to power AI agents andllms and things like that. So
(04:33):
you can imagine, if you wantedto write some kind of cool
custom GPT or something thatwould maybe help you calculate
sales commissions or help youunderstand your travel expenses,
the first step would be, I haveto go pull all that data
together and have it make sense,and then be able to hand it to
the GPT. And so a business usercan do that in Alteryx. It's not
a coding platform. Think of itas being for somebody who's good
(04:55):
at Excel, not somebody who wantsto spend their nights and
evening writing Python code.
Yeah.
Scott Allender (05:02):
So I'd love to kind of start with getting your
sort of view on the sort ofstate of the nation, assessment
of how organisations arecurrently responding to genitive
AI, because. Getting a lot ofdifferent reports and
perspectives, and I'd love toget your insights here.
Andy MacMillan (05:19):
I think it's pretty chaotic, but also
exciting, and hopefully that'sthe kind of world people enjoy
living in, because, like, that'sthe kind of world we live in
now. I think the pace that wetake on change now in business
is completely different than,say, 20 years ago. I think the
kind of enabled by internet andmobile, all of a sudden when
(05:41):
things happen, they kind ofhappen in a snap, right? And and
so when something likegenerative AI comes out, I think
people can see the potential. Idon't think you have to spend a
lot of time in chat, GPT orGemini to sort of go, Wow. This
is different. This is somethingnew. And to think about how you
can apply that to your businessor your life. But then you have
the challenge of, okay, but allthe all the scaffolding is not
(06:02):
there yet, all theinfrastructure is not there yet.
However, I know that this isgoing to happen fast. I learned
from the last couple of bigwaves. This is going to happen
quickly. And so I think peopleare, are sort of scrambling. And
I mean that in a positive way, Ithink that's people leaning in
and trying to figure it out. ButI think we're kind of in the
scrambling phase right now aspeople try to move quickly, but
(06:22):
maybe don't like I said, haveall the scaffolding in place yet
to really know how to how toscale and find value, find ROI
and so that's the sort ofchaotic moment we're in right
Jean Gomes (06:33):
now. Given where on this, you know, there's a lot of
noise and kind of counteringviews and some fairly hysterical
kind of commentary on on AIadoption, when you are talking
to senior leaders, when you'retalking to CEOs and C suite
executives, what's your kind ofsense of what's most helpful for
(06:54):
them in building thatscaffolding and actually making
practical steps forward so theydon't fall into the traps,
perhaps, of some of the digitaltransformations of the past, I
think
Andy MacMillan (07:04):
one is looking for an outcome you're trying to
drive in your businessgenerally, and applying AI
towards it versus, you know, Igot aI I'm running around trying
to find what I'm going to dowith it. Can run people in a lot
of different directions. So Idefinitely think that's one. I
think the other is thinkingabout it systematically. I think
of AI as being this veryinteresting constraint remover,
(07:27):
and when you remove aconstraint, new constraints
arrive. And so I don't think itmeans that things scale
infinitely. But, you know, Ispent a lot of the time in my
business, thinking about, youknow, what if we can write code
10 times faster? Like, how willwe organise our teams
differently? Maybe my teams seemto be located closer together,
like, what changes about thebusiness? And so I think one,
you know, pointed at realproblems, and think about how to
(07:49):
work backwards towards solvingbusiness metrics that you
already care about. And I thinkthe second is, when you do that,
think about it as sort ofbusiness process reimagination
with different constraints. Ithink when we limit it only to a
personal productivity tool,you're not really reimagining
the business process. And ifdifferent people are going to be
(08:09):
different levels of newefficiency, you're going to have
to rebalance your teams andfigure it out, and you're right
back to business processreimagination. And so I think
for me, those two things are thekey pieces to really trying to
drive value when you're tryingto apply AI to your business,
not just sort of play aroundwith or learn a bit about AI.
Jean Gomes (08:25):
That feels very different from some of these
digital transformations in thepast, where it kind of gets
driven down to the digitaldirector or the head of
transformation or the ITdepartment. This sounds like
it's a much more crossfunctional leadership required.
Andy MacMillan (08:40):
I think you can think about any aspect of a
business and imagine how AIchanges it. And I would posit
that the people that willultimately think about how to
change those parts of thebusiness will be those people in
that part of the business. So Iagree with you. Whereas digital
transformation, I think,initially started off very
technology driven, you know, oh,we need to stand up a website
(09:01):
and E commerce and things likethat, and we turn to our
technical teams to do that. Youthink of where that ended. We
have handed most of that stuffback to the business, right?
Most websites are, you know,yeah, the IT team runs the the
infrastructure, but themarketing team owns the website,
or the the merchandising teamowns the E commerce site. I
think with AI, you know, everyleader in every functional area
(09:24):
is going to have an opportunityto look at the thing they're
running and try to reimagine,how would I do this with this
new capability that I've neverhad access to so you're going to
be in the Office of Finance andthinking about, well, how would
I, how would I manage Travel andExpense differently? If I have
access to a real time agent,you're going to be in the legal
team and say, how would I run mylegal team differently if I can
(09:44):
red line a document using AIbefore I ever have my lawyers
actually look at it? And I thinkthat's not an IT problem. I
don't think that's going to, Imean, it is going to help.
They're going to provide againsta lot of that scaffolding. But I
think unlike these othertechnology transformations, I
think it's pretty easy toimagine a universe where, no
matter what function you're in,you need to be thinking about,
(10:07):
how does this new capability?
Forget about as a technology,just this capability. Let me
scale what I'm doing and.
Entirely differently, and Ithink that's where we're going
to see people over the next yearor two, really figuring out
almost function by function.
Like, what does this mean if Iwork in accounting? What does
this mean if I work in supplychain? What does this mean if
I'm in sales? And I thinkanybody who's not thinking about
(10:29):
that risks sort of having theirskill set get left behind.
Scott Allender (10:33):
What about beyond a year or two. Take take
me five years, 10 years down theroad. What do you think an
automated workplace looks
Andy MacMillan (10:41):
like? Yeah, I think companies look a lot
different. And I think again, ifyou if you imagine it from a
Theory of Constraintsstandpoint, and again, I think
it's helpful to almost thinkabout it by function. I'll give
you an example in softwaredevelopment over the past 20
years, one of the big trends hasbeen offshoring. And the reason
is, the main constraint whenyou're building software is how
(11:03):
many developers do I have accessto? Right? If I more developers,
I can write more software. Andso you would have your designers
and your product managers oftenclose to customers the biggest
markets, often the US. So thosepeople might be here in the US,
they would design things. Theywould send them offshore to a
lower cost location where youhad a lot of engineers. And they
would do these, you know, twoweeks sprints. So if you're not
in software development, a twoweek sprint is two weeks of
(11:25):
writing the code, and then yousort of pop up two weeks later
and go, Hey, here's a here's ademo of where we're at and how
it works. And you sort ofiterate. Imagine now if that
sprint only takes an hour? Well,I don't want to have that team
now on the other side of theworld, I might want to have them
really co located, iteratingwith that team. And so again, I
don't know if that becomes thenew constraint or not, how fast
(11:46):
I can iterate, but how manypeople I hire, where they're
located, how they work together,what that job entails. What does
it mean to be a softwaredeveloper if I'm using coding
assistant tools. So that jobchanges too. So I can imagine,
in five years, the entire modelaround software development
doesn't go away. I don't thinkit's all the robots doing it
like I don't think that's,that's the you know, I think
we're headed for that dystopianfuture. I think we are headed to
(12:09):
a place where we'll write bettersoftware faster. We'll do it. I
think with teams, it can iteratemuch more quickly. And so that
makes me start rethinking, well,how do these people work
together, and where are theylocated, and what are the new
constraints? Is the newconstraint, how fast they can
work versus how many developersI have? I don't know. I think
about that with with salespeople, we have a whole bunch of
(12:31):
people. We surround sales peoplewith that try to make them more
effective. We have inside salespeople, and we have renewal
managers and all these otherthings. And you start to wonder,
well, what parts of thosefunctions can AI help the
salesperson do more effectively?
So do I hire more sales peoplebecause they're more effective?
Do I have to hire fewersalespeople because they can get
more done? It's going to change.
(12:51):
And so I think leaders are goingto have to come at every part of
their business. And I think aCEO is having to work with every
one of their functional leadersto kind of come at it with an
open mind. And again, reallyoperate from like a first
principles. Again, I reallythink Theory of Constraints is
very helpful. When you thinkabout AI, it will just change
what the constraints are. And sothat means you different ratios
(13:13):
of roles, different skills inthose roles, and so I think
businesses will look a lotdifferent in five years.
Jean Gomes (13:20):
Let's pause here and on this theory of constraint
idea, because that's a veryhelpful lens to think about the
underlying assumptions that areplaying out in in in how your
business processes, businessmodel and so on is developing.
Can you walk us through anexample? How? I mean, you've
given us a few high levelexamples, but how you think
(13:41):
about applying that and how newconstraints start to emerge? How
do you what's the if you wereguiding a team thinking about
how to apply that thinking in amore detailed way? Can you just
walk us through an example tohelp us?
Andy MacMillan (13:56):
Yeah, I spend a lot of time with all of my teams
trying to think through, if Iwant this team to go faster, to
be more efficient, whatever thegoal is of the thing that I'm
working on, what is the onelever that changes the most?
Right? So if we were going todouble our sales, what's the
first thing everybody think ofthat changed? Was it that we had
twice as much pipeline? Was itpipeline? Was it that we had a
(14:17):
twice the win rate in ourcustomers? And simple questions
like that can lead to, what isthe core constraint? The fun
thing about constraints is, assoon as you fix a constraint,
something else becomes theconstraint like that's just the
way that works. But you startworking through what are the
biggest problems? And I thinkthat's a useful way to think
about rolling out AI, if I canapply AI to my biggest
constraints now, it starts toreally open up my business to
(14:40):
scale and do new things. Versusagain, I'm running around with
AI, you know, it summarisesstuff. Well, what can I go
summarise like that's not aparticularly interesting
business problem to go solve.
But when you think aboutconstraints in a business today,
very often, constraints comedown to head count. It is one of
the biggest expense items in alot of businesses, certainly a
lot of kind of software stylebusinesses. And if I start
(15:02):
applying AI to make peopledramatically more productive in
certain roles, and maybe that'snot as dramatic in other roles,
again, those constraints start.
To shift. My ratios start toshift. So if I think of
something, yeah, like my go tomarket team, it's very ratio
driven. You know, I have allthese regional teams, and in
each region, I have, you know, Xnumber of sales people, and I've
(15:23):
had X number of sales people. Ihave, you know, y number of
field marketers, and I've got,you know, z number of of
technical sales specialists thatare helping the sales people.
And so you look at that, that'sa pretty finely tuned machine.
It's a it's a pretty common inmy industry sort of software go
to market motion. I would guess,if I met with 10 other software
CEOs, our ratios look prettysimilar, you know, maybe a
(15:44):
little different here, a littledifferent there, but for the
most part, pretty similar. Andnow imagine AI makes one of
those groups three times moreproductive. It makes one of the
ones only 50% more productive?
Do I now change all the ratiosof how these teams work? Am I
hiring kind of a differentworkforce at that point? How do
they work together? That's goingto change. And so that's really
where I start to think about theimpact. It's not all our jobs go
(16:06):
away and the robots do all thework, it's more well, it's going
to change how we work togetherand what work we expect people
to get done. And I think that,to me, is, is interesting. I
think the other one, I think alot about, is there's this
debate, I don't know which sideof the spectrum I'm on, of, you
know, with AI do all entry leveljobs go away because the,
because the AI does all the theentry level work, or if the flip
(16:28):
side, you know, I can give AI tosomebody in the early part of
their career. And it's thisamazing amount of expertise that
can sort of apply to problems.
To problems. And so I think justthat friction sort of shows I'm
not sure the answer is one orthe other. I think I go back to
like, well, some barely, reallysenior roles might scale better,
because you can give them accessto AI. Maybe if my entry level
(16:48):
people are now way moreproductive, because I can give
them AI, might hire more entrylevel people too. I think a lot
about when something's betterand cheaper, people tend to buy
more of it. And so I also thinkabout employment that way. If
what I get for my dollar is asenior executive when I hire
someone who's way moreproductivity, I might hire more
people rather than just hiringfewer people, because I've got
(17:10):
productivity gains. So those arekind of the three things I think
about when I reimagine whatbusiness looks like over the
next couple of
Scott Allender (17:23):
years, what are some associated and emerging
leadership challenges andopportunities for inspiring and
motivating you know, economicactivity in the sort of new
world and the sort of new way ofdoing business.
Andy MacMillan (17:40):
I think a big one is acknowledging how much
change is coming. I think peopleget it wrong when they start
telling people, like, you know,don't worry, nothing's going to
change. It's like things aregoing to change. I told my whole
team, like, my my engineeringteam, for example, is starting
to use a lot of these codingassisting tools there. I have a
great engineering team. It'sobviously a little bit nerve
wracking when you start going,like, wow. Like, this, this
(18:01):
thing can write half the code Iwould have written before. Like,
that's an unsettling feeling.
And my advice to leaders is is,you know, you lean into that.
What I've been telling my teamis, let's take up these tools.
We've got a lot of stuff to gobuild so we're not maxed out on
capacity like I would. I wouldtake if you told me I got 1000
more developers at no cost, Iwould take them in a heartbeat.
So there's a lot of room togrow, but for themselves,
(18:22):
personally, your career strategyat this point can't be I'm going
to hope my current employerdoesn't know or use AI, and
maybe for the next 10 years,I'll just find other employers
that don't know about or use AI,like, that's not a career plan.
Yeah, and so, so sort ofembracing that with your team.
Hey, we're on a missiontogether. This change is
happening. You're going to workat a place that's going to
(18:44):
support you going through this,and we're going to learn as we
do it. And, you know, embracethat this change is coming. I
think that's really important.
And then to be listening, I'mnot also just directing my team
go do this, and I don't want tohear about the results. I'm
trying to lean in and learn howare each of these functions
changing? What concerns dopeople have as these ratios
(19:05):
start to shift, starting tothink about, how are we going to
rebalance our teams? And ourmessage internally has been,
we're going to slow hiring alittle bit right now. And what
I'm trying to do is create spacein our business. So as we
rebalance, we can apply the headcount as we're rebalancing. So
trying to be really upfront withpeople, like, yes, there will be
change that comes. We can kindof both manage that change as a
(19:26):
leadership team, so that all theimpact isn't just hitting the
front lines as we, you know,hire a bunch of people and go,
Oh, oops, we should hire themover here. Like nobody wants to
work at a company that doesthat. Does that. So sort of
acknowledge, yep, we're being alittle a little cautious with
hiring, until we see, kind ofsome of the dust settle on some
of these ratios, but we're goingto lean in. And I think that's
really important. I think peoplewant to feel led through change,
(19:49):
not that change is being managedfor them. I think a lot of
leaders make a big mistaketrying to obfuscate change and
challenge trying to obfuscateinto the change is coming. And I
do the opposite. I get in frontof a say, this is happening. You
read about every single day,every conference you go to, you
know this is going to be a placethat sort of helps you go
through that change.
Jean Gomes (20:10):
How are you I mean, I love the kind of. Spirit that
you're creating in in thisbecause, you know, psychological
safety disappears when peopleare in in an environment where
the technology is threateningthem in that way. And that's
always been the case, but nowyou kind of a whole new level of
threat, because it's, it's it'sintelligence, not just
(20:30):
automation. What? What are youdoing to help people at a
practical level, to be able toquickly lean into all of these
things. What have you learnedabout what works?
Andy MacMillan (20:44):
Well, part of it's giving people access and
opportunity to play around andwork with this stuff. So I did
this, not only here, but at mylast company, very early on. We
licenced an LLM non training,you know, private use, but
company wide, gave people accessto that, gave them training. So
we run a regular training onhow, what are the rules for
(21:05):
using this? How can you use it?
You know, what data can you putin it? What can you be doing?
We've been showcasing on our allhands on things that people have
built using AI and so, reallytrying to make it a place to
sort of learn and let peoplesort of evolve and build stuff.
So I think that's a big part ofit, and sort of celebrating
that, and again, just continuingto to highlight, you know, for
us, we're trying to efficientlygrow our business. You know, as
(21:27):
you mentioned in the theopening, you know, we just went
through a pretty substantialtake private we did a financial
restructuring of the company,got us to a very healthy place
on our balance sheet, and wewant to drive even more growth.
And so I've been telling thecompany, this is the perfect
place to apply AI. We want todrive growth while we manage
having a strong balance sheet.
(21:48):
So from the top level of thecompany, I've given a vision of
how AI is going to help us dothis in a way that brings
everybody along, I've then triedto provide them the tools to
start to, frankly, play with thetechnology. And I've told them,
you know, you are free to startapplying this to how you do your
job, and let's start sharingwhat works. And so I think those
things are all part of it, andit's just a continual topic. I
(22:09):
think this is something thatcomes up on almost every one of
our all hands. I write a weeklynote to the whole company. Every
Sunday night, I just write alittle email. I talk about AI
transformation quite a bit inthose emails, what I'm doing,
what I'm seeing other people do.
I built a little GPT myselfusing Alteryx to prepare some
data for something I was workingon. And I was like, I'll just
show this to the company. So onone of our all hands, I
(22:30):
literally just got out my littlething, and I'm like, here's what
I'm doing. And, you know, andjust trying to let people know,
you know, this is kind of howyou learn. You just sort of
iterate and play
Scott Allender (22:39):
with stuff. When I echo John's sentiments about
the sort of spirit that you'resetting and the tone that you're
bringing, and your sort ofintentionality, which is really
inspiring as you talk about thisstuff, I'd like to take a slight
tangent and kind of hear abouthow you got here, like, what are
some of your sort of mostimportant leadership lessons
(23:01):
that has readied you to leadthis organisation and prepare
people for the future?
Andy MacMillan (23:06):
I think one of them, Scott, that stands out
specific to kind of thisapproach was, you know, when you
when you become a hired CEO. Onething I tell people, very rarely
does somebody say, hey, thisbusiness is running perfectly.
Why don't you come run it? Sothere's usually stuff to go in
and do. And one of the things Ifound earlier, one of the first
(23:27):
companies that I joined as CEO,was the more people I openly
shared problems with, the morepeople wanted to help with those
problems. And I always feltlike, wherever the line got
drawn, like, if I you know that,I think I started too small with
who knew about what we weretrying to do. There were some
real problems in the firstbusiness I joined. I mean real,
real issues. And those issueshadn't been shared widely in the
(23:49):
company. And I sort of foundthat as those as I started to
bring more people, kind of intothe tent. People were like, Oh,
yeah. Like, actually, that makessense. Why things have been
like, the like, how do I help?
But then wherever the line hadstopped to the next level was
like, what's going on? Like, I'mreally frustrated. Why are
people and so I've just foundthe way to get everybody working
on something big change, orwhatever is, to bring as many
(24:10):
people as possible into theprocess. And if you're really
open with people again, inpositive I'm a naturally
optimistic person, so I'm alwayslike, Okay, well, here's the
problem, but we're going to fixit. You get people helping you
fix it. When you try to fix theproblem, before you tell people
about it. And I think this canbe something leaders try to do.
You think, oh my gosh, if, ifeverybody finds out that we have
this big challenge, you know,people will quit, people will
(24:31):
leave. You know, people can go.
They'll vote with their feet,and so then they're on the
outside, looking at theleadership team, going, seems
like something's not working.
Nobody's telling us what's goingon. They must not know what to
do. That's unsettling. Versus,hey guys, we've identified the
problem. Here's what it is. It'sa big problem. We're working on
it, and let's figure it out.
(24:52):
Most people, even if, like, Idon't know this might be really
hard. They sort of grab onto therope and they start pulling, and
people like feeling like they'rehelping work on the thing that
you're working on. So I thinkthat's one of the big leadership
lessons for me. I thinkthroughout my career, I came up
through the product managementside of the world as a
developer, early on in mycareer, I've always liked the
problems the company works on. Ijust find that interesting. And
(25:15):
so I think one of the thingsthat served me well as a leader
is, strangely, I never managedfrontline people. My first
leadership Job was a VP leveljob, and I managed people that
were 10 years older than me, andI sort of got this big like
career accelerator. So So one,I've never been a micromanager,
in part because I've never hadto lead people in the very first
stage of their career, which iswhere I think you have to do the
most hands on management. SoI've gotten management. So I've
(25:37):
gotten to be kind of a littlebit of a senior leader, leading
leaders early. And then theother is, while doing that, I've
always enjoyed leaning intoproblems, but realising when
you're a senior leader, it's notalways your problem to solve
entirely. But I like to sort of,hey, let me I kind of call it
participant leadership, like, Ilike to, hey, we're having a
(25:57):
design meeting. I'm going tocome to the design meeting as
the CEO. I maybe have somecontext of what's going on in
the market and the company, butI'm not running the design
meeting. I'm not the ultimateapprover. I'm just sort of there
to participate, and I reallyenjoy that, and that has served
me well, I think throughout mycareer, is to be somebody who
just likes to, you know, workwith different teams trying to
solve problems around thecompany, but not feel like I'm
(26:19):
the the ultimate solver. I'msort of an enabler.
Jean Gomes (26:23):
Coming back to, you know, the kind of challenges of
embracing AI, we hear a lot ofevidence now coming that it's
making people de skilled. It'screating cognitive complacency.
People's memory is going theycan't remember what they've
created on AI and so on. Whathave you learned? I mean, either
(26:45):
in your own use or in helpingyour your teams to adopt it so
that they get smarter, notdumber, using it.
Andy MacMillan (26:52):
Um, I don't know. I think there's some, you
know, like brainwave, kind offancy studies on all this. But
my analogy is a little bit, youknow, I think they said that
about calculators and computers.
And I'm not really sure wheredumber as a society, because we
use computers. I remember mydad, when I was growing up,
always wanted to teach me to domath on his slide rule, because
that was the way to reallyunderstand it. And I don't think
I ever understood how to use theslide rule, and I don't think
(27:12):
I'm any dumber for not havingfigured that out. So I feel a
little bit like it's a newcapability. How do we learn how
to use it? I don't think thatmeans we don't think anymore. I
think it means, pretty soon, welearn how to work with this much
smarter set of intelligence thatcan help us learn a whole lot
more. I feel like I learn atonne using AI every day. I find
myself even just when I getinterested in something, I can
(27:34):
go into chat, GPT or Gemini, andsort of work through a problem
in such depth, where I learnedso much. So I'm not entirely
convinced that this is somethingthat's going to make us dumber.
I think it's something we'regoing to have to learn to work
with differently. It mightchange how we operate. I would
argue maybe the advent of like,my kids just get to use
(27:56):
calculators, even when they takethe AC, T and stuff. So, yeah,
maybe they don't do off the topof their head arithmetic as fast
as my generation did. I don'tknow that that's going to set
them back, but, you know, maybeit changes. So I think that's my
my mindset is, I don't thinkthis is a replacement for
cognitive capabilities. I thinkit's a new skill that we're
going to learn, and I think it'sgoing to mean that people can do
(28:18):
more things more quickly, and dothem better. I think about the
even around the house, like Ican go into Gemini and get help
on doing a home repair. I'mterrible at home repair. I'm a
lot better with Gemini. So doesthat mean I'm dumber or smarter
that I know how to use Geminibefore I start, you know,
smashing holes in the walls. Ithink it makes me smarter than I
know how to use it, but I sortof get the argument of, like,
Well, my brain didn't figure itout. Like, Well, okay, but my
(28:41):
brain wasn't doing a very goodjob figuring out home repair
before I started to have
Jean Gomes (28:46):
access to this. So I think, you know, I think it
probably comes down to how youuse it, and as a, as somebody
who's, you know, grown up as a,as a product manager and
probably has a quite goodunderstanding of how to think
about solving problems. Itreally helps you. I'm, I'm more
concerned about, you know, howpeople who don't know how to use
it are sort of relying upon itlike a search engine to get easy
(29:09):
answers, who don't dive into thedetail, who don't use it to
think together, you know, might,might fall into the trap of, you
know, losing their abilities. I
Andy MacMillan (29:20):
do think we're starting to learn what it's good
at and not good at. I mean,we've all had the experience
where you you ask the LLM aquestion, it gives you an
answer, and you sort of go, areyou sure? It goes, Oh, you're
right. This is totally wrong.
How about this idea? And so Ithink we're learning also sort
of what it's good at, what it'snot good at, to your point,
which is, maybe we're not takingeverything always at face value.
Sometimes there's a little bitmore connecting the dots that we
(29:41):
want to be do, doing, ratherthan just, you know, offloading
all of this. I think everystudent I know has learned the
lesson of, you know, oh, I usedone of these tools to summarise
the thing I was supposed to readfor class. And I got to class
and realised I didn't, didn'tknow what was going on, right?
So you, I think there's sort ofa where to use it, how to use
it, that we're all stillfiguring out when to trust it,
when to when to drill deeper. SoI agree with you. I think
(30:03):
that's, that's, again, kind of askill we're learning.
Jean Gomes (30:11):
What are your top tips in using it? Well, I mean,
what? The things that you'velearned about it.
Andy MacMillan (30:16):
I think, as most people say, I think learning
about how to prompt and ask forvery specific things is
important. I'm also learning alot about the responses come
back with full confidence, likethe dial is always set to 211 on
confidence, regardless of howconfident it actually is. And so
you can query into like, Howsure are you, and where did you
get this from? And so I think,sort of, again, that level of
(30:38):
interrogation on things thatmatter, I think, is really
helpful. I think we're seeingmore of this now getting cited.
So I'm finding now often, if Iask, maybe my personal life, a
more general question, and nowit shows a little bit of its
work. And sometimes we'll showsources, I'll find myself
clicking through the sources. Soit's like, okay, like, yeah,
it's summarised for me. Andeverybody felt like, Oh, this is
(31:00):
going to be the death of, youknow, search engines with click
through results. And now I'mfinding like, no, no, I've read
the summary, but now I want toclick through, you know, oh,
here's the thing, and there'sthree references. And I'll look
at the middle and go, Oh, thatlooks interesting. And I'll kind
of click through and and read. Ithink that's new. I don't think
I was doing a lot of that, youknow. Six months ago, I was sort
of taking sort of taking it atface value. And so I think those
kinds of things, you know,learning to prompt, learning to
(31:21):
drill in, is is a big one. Theother is it work. I've learned
this pattern of trying to donarrower things. And so what
I'll do is, again, maybe,because I work at Alteryx, I'll
pull together a set of data,right? Like so, for example, I
was analysing my sales pipelinerecently. I went into Alteryx. I
pull in my Salesforce data, I doa little calculation on my
(31:44):
pipeline, and then I hand thatto a narrowly scoped GPT and
say, Okay, you're a pipelineexpert. Help me work through
this. I think that's reallyinteresting, too. You're not
doing this like, broad generalyou know, I think of it as like,
are you asking a smart friend aquestion, or are you asking a
specialist a question? And moreoften, I'm realising, you know,
AI is sort of a smart generalfriend, but I can sort of tell
(32:06):
it to be a specialist. I cangive it a narrower set of data
and a narrower set ofinstructions, and I think I get
real value when I do that in mybusiness work, right? And that's
important, because I think whenyou ask it generalist questions,
you get generalist answers. Andit worked. That's not that
interesting. If I say to to chatGPT, you know, what will my
sales commission be if I closethis deal and it goes well,
(32:27):
here's how companies normallycalculate sales commissions. It
might be something like this.
That's not the right answer.
Like, the answer I'm looking foris, like, what will it actually
be? So again, if I it work, cancreate a small data pipeline of
how we actually do that, and Ican tell it, hey, only work in
this thing. Do this work. Ithink that's really useful too.
So I think also learning tothink about scoping, you know,
(32:50):
in personal life too, like I'vedone that with. I wanted to
analyse some some stocksrecently, and so again, I gave
it a narrow data set and anarrow set of instructions and
asked it to do something, versusjust asking, you know, the front
page of Gemini or something, togo do the work. So I think that
scoping is a skill set that willemerge to
Scott Allender (33:10):
those are good suggestions. Are there any other
sort of, maybe five to 10 minutea day practices for people
listening saying, I've been kindof avoiding this. I don't really
know how to use Gemini or chatGPT the right way, anything that
you'd say. You do this 510minutes a day, and you'll start
to really build your your musclearound AI, yeah,
Andy MacMillan (33:30):
I sort of, I like to think of it as, like,
imagine two very discretealtitudes that you want to
interact with an app. One issort of the, you know, do some
leg work for me, kind of stuff,you know, on the ground cover,
summarise this email kind ofstuff, right? Like, it's great,
helpful, but it's pretty kind oflow level, like, hey, grind out
some work for me. I think thebig unlock is when you start to
(33:53):
think of it more like, hey, whatif I could hire a McKinsey
consultant to look at this?
Normally, I wouldn't. But like,I can. I can ask this thing to
act as an expert thatunderstands audit and finance,
and go through this thing reallyquite fascinating. What you can
get out of it? I've been doingthings like for my last board
meeting, I get my financial sentto me by my CFO, and I can go
(34:14):
into, again, our private nontraining instance of an LLM, and
I can say, imagine you are a BCGor McKinsey consultant advising
a CEO in a PE backed company onthe following board meeting.
What trends in the businesswould you recognise? What kind
of questions do you think theboard would ask? Now, I would
argue, 80% of what came back,like, I do this for a living.
(34:36):
I'm like, yeah, those are thethings. But you get a couple
nuggets here. They're like,okay, that took that took two
minutes, and I got a couple ofit's point. I should think about
that. One of the was a reallyinteresting idea. Didn't come up
my board meeting, but I waslike, I should take that and go
talk to my team about that idea.
And so I think again, using itat that higher altitude, again,
with a scope set of questions,of, you know, hey, act as an
(34:59):
expert in software development.
You know, look at this project.
What would you say are the holesin this project? What other
areas would you advise? It takestwo minutes, and even if you
know, you only get a reallyinteresting answer one out of
three times. That is not a lotof. Effort for a lot of material
that you can get back. And so Ithink that's really useful, I
think especially for seniorexecutives that are trying to
(35:21):
operate across a large set ofinformation coming to you. I'm
finding more and more often I'mtaking something sent to me and
I'm just dragging it into theLLM, and I'm not doing the
summarise for me. I'm doing theHey act in this role and help me
really think through what thisdoes. And then again, you can go
back and forth with it. Hey,that's a really good point. Why
did you bring that up? What areother companies doing around
that, you know, prepare a twopager for me on how I would go
(35:43):
about, you know, looking intothat area, it's a really
interesting brainstormingexercise, kind of at that level
that I think people should domore of.
Jean Gomes (35:52):
And, you know, this is probably one of these kind of
questions or thought experimentsthat comes and goes with every
new trend. But one thing thatcame out probably six, eight
months ago was the idea that atsome point, you know, an
organisation might actually havea, you know, an AI CEO I'm just
(36:13):
interested in, you know, likeyour thoughts on, on that kind
of idea.
Andy MacMillan (36:19):
I mean, someone said recently that AI makes very
confident decisions on verylittle data. And I'm like, well,
that's most of my job. So maybe,maybe they could do that quite
well. You know, I don't know. Igo back to, I think there's a,
there's an amazing capabilityset emerging. But I still think
businesses are going to bepeople figuring out what they
want to go do, why they want togo do it, and sort of using
(36:40):
these tools to think about howto go do that. So I definitely
think you could imagine having aseries of agents acting as
really skilled consultants andadvisors and a whole bunch of
areas in the business, and youmight turn to that more often
than not. I still think there'sa level of accountability that
exists in the job, in a lot ofroles, not just CEO, but also
CFO and things like that. Ithink we're a long way from
(37:02):
somebody saying, you know, thefinancials are good because the
robots signed off on it. I thinkwe're gonna say who signed the
financials to say that. That'sright, a little bit like that.
You know, there was a Waymorecently they got pulled over
here in the Bay Area, andthere's this great video the cop
walking up to the window, andthere's nobody in the driver's
seat, and they're trying tofigure like, Well, what happens
now? Because the Waymo made anillegal U turn, I don't know
(37:23):
that we're ready for that to belike a company that did
something wrong or a companythat needs to make a hard
decision. So I think there's alot of play still in this world
for people having ownership overwhat we're doing, but I do think
the role changes. I think it'llbe hard to imagine being a CEO a
couple years from now and notbeing regularly advised and
(37:44):
interacting with, you know, datadriven AI, I think that's,
that's the reality of wherewe're headed. What's
Jean Gomes (37:49):
the thing you're most excited about for next
year? The
Andy MacMillan (37:52):
thing I'm most excited about for next year is
we have just started to enablein our product the ability to
let people build their own datapipelines and put them directly
into whatever kind of LLMthey're using. And so for me, I
have a whole model I use to runthe business, a whole set of
meetings, cross functionally andstuff like that. And I'm
basically creating customer gptsaimed at every one of those
(38:15):
meetings. And I think I'm goingto live in a universe next year
where before we go to every oneof those meetings, we can have a
report coming to us, generatedby our AI agent telling us all
of the things that we're goingto be focused on in that
meeting. And I think that'sgoing to make life a lot of a
lot more fun. Like, I'm kind ofexcited about the pace at which
I think I can run the businessusing AI kind of blended into my
(38:36):
current leadership model. And Idon't think this is a three year
off thing. I think this would besomething that I'll be able to
do at altrix in the first halfof next year, and I think
that'll be pretty I know I likechange, so it'd be interesting
to see, like, how does thathow's that work?
Scott Allender (38:50):
Thank you, Andy, this was really insightful and a
really practical way for us towrap five years of incredible
conversations on the evolvingleader. But don't worry, folks,
the evolving leader isn't goingaway. It's just evolving. There
is a new multi trillion dollarglobal economy that's rapidly
emerging, spurred bytransformational advances in our
(39:11):
understanding of how our brainsand bodies work and the question
of, what are humans for in anautomated world, how we feel,
how we think and how we see,will profoundly influence our
economic success and our wellbeing and our capacity to
embrace uncertainty. So thisJanuary, January 2026, a new
(39:33):
show is coming to you from uscalled the mindset economy, and
our goal, as always, is to forgea more human world. So we'll see
you there. And until then, I askyou, for one final time, the
world is evolving. Are you? You?
You.
Popular Podcasts
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.
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.
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