September 10, 2025 • 47 mins
In this episode of Let’s Talk Pricing, host Kevin Mitchell (President of PPS) sits down with Dr. Michael Wu, Chief AI Strategist at PROS and one of the most respected voices in AI and pricing.
Together, they explore how artificial intelligence is reshaping pricing strategy, from the rise of intelligent agents to the human skills that will matter most in an AI-driven world. Dr. Wu explains what’s changed with Generative AI, predictive models, and intelligent pricing agents—and how pricers can prepare for the next wave of transformation.
🎤 Key topics include:
Why Dr. Wu calls this moment a true (R)Evolution in pricing
How to prepare for the limitations of Generative AI
The future of human + machine collaboration in pricing
The mindset shift needed to move from dashboards to predictive and generative insights
👉 Don’t miss your chance to learn directly from Dr. Wu this fall! He’ll be live at PPS profitABLE in Las Vegas, October 21–24, 2025, leading a hands-on workshop on AI-powered pricing strategy (worth 1 CPP credit) and delivering a keynote on the rise of intelligent agents.
Reserve your spot at PricingSociety.com/ppslv25
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
You're listening to Let's Talk Pricing, your connection to the
voices, stories, and strategies shaping the pricing world.
Each episode, we go beyond theory into the practical,
timely conversations that help you lead with confidence and
drive results. Let's start the show.
(00:20):
Very good day everyone, and welcome to the Let's Talk
Pricing podcast. We are the official podcast of
the Professional Pricing Society.
My name is Kevin Mitchell from PPS and I'm very, very excited
about today's episode. We are going to be talking with
Doctor Michael Wu about becomingan agent of change in the age of
(00:42):
artificial intelligence. So we're going to be talking
about AI and machine learning and lots of interesting topics
there. And a lot of us already know
Michael. But for a quick introduction,
Doctor Michael Wu is the the chief AI strategist at Prose.
He is a world leading expert in artificial intelligence and
machine learning. He is, among other things, in
(01:05):
Advertiser, I'm sorry, an advisor and lecturer at Cal
Berkeley's extension AI programs.
And as part of his studies, he once used machine learning to
model the human brain. And also we're very, very
excited that Doctor Michael Wu is going to be with us at the
PPS Profitable conference in LasVegas in October.
(01:27):
He's going to have ACPP workshopon AI powered pricing strategy.
And he's also going to deliver akeynote address about the rise
of intelligent agents and the human skills that will matter
the most as AI reshapes pricing management, revenue growth
management, sales enablement andrelated fields.
(01:49):
So a very, very nice welcome to Doctor Michael Wu.
Michael, we are very excited to have you on.
How are you today, Sir? I'm good.
It's it's such a pleasure to be in conversation with you, Kevin.
Thank you so much, Michael. I appreciate that.
And we have a few questions for you.
And AI is an area where actually, believe it or not, I
(02:10):
should probably hang my head in shame as I say this.
But actually I go to PPS conferences to learn more about
AI from experts such as yourself.
It's an area that I do not know enough about yet.
So I'm looking forward to learning a lot from you.
And I know from your work and from being a follower and a big
(02:34):
fan of your work as well that you have described the shift to
strategic artificial intelligence, the shift to
strategic AI as a revolution in pricing.
So what's the key transformationthat's taking place right now in
pricing and revenue management and related fields?
Yeah, I think to answer the question, you know, you, we can
(02:58):
just look around and ask us what's happening right now in
the world, right? I mean, I think right now we all
know the world is a little bit chaotic.
It's it's, you know, and therefore in terms of, you know,
business and market, you know, it's highly volatile, right?
So it's changes all the time. And So what that means is that,
(03:18):
you know, pricing, what's the, what's happening in pricing
right now is we're trying to react to that world, right?
And the best way to deal with such a highly volatile and and
crazy unpredictable world is essentially to do real time
dynamic pricing, right? OK.
So what that really means is that you know, you know, you,
(03:41):
you have to kind of, you know, kind of compute these price
right in real time. And all that means is there's no
no more master price list, right?
So the price is actually generated at real time, right?
Because the minute you produce a, a price list, right, If it
takes time to actually produce aprice list for all the customer
(04:05):
and all the different product combination that you have,
right? And that takes a long time,
right? So, you know, you may do it very
fast, you may have AI help you to do, but that still takes
time, right? You know, even if it's just
like, you know, a week or a few day, right?
You know, it takes you time to produce that master price list,
right? And the minute you produce that
(04:25):
list, right, it's already a few days late, a weekly or monthly,
however long it takes you to produce them, right?
So that so that to get the real time like means right now,
right, you have to be able to compute and generate that time,
sorry, that price in real time, right?
Compute that. So, so, so that's why you know,
(04:46):
yeah, we we need to kind of moveon to this truly real time
dynamic pricing. I, I wanted to make a
distinction with the, the term dynamic pricing as well, because
I think if you very often when you hear real time dynamic
pricing, people say, oh, we, we,we, we do that.
You know what they really meant is that while we do dynamic
(05:06):
pricing and dynamic pricing justthat like your pricing changes
with something some, you know some indicator, some variable on
about the market, right, it changes with the but it doesn't
mean you change with that in real time.
Understood. Yeah.
And I like some of the things you said about there about our
chaotic volatile world. And of course, as pricing
(05:31):
professionals, we have a lot of externalities that come into
play with our decision making, market changes, government
changes, macroeconomic changes, customer changes, competitor
changes and things like that. And we have to take all those
into account. And so speed definitely is a a
very important variable there. And also, I like how you
(05:52):
describe that, what a lot of people think about dynamic
pricing, really they're just maybe pricing more quickly
instead of having that real, real time dynamic pricing where
everything is changing almost inan instant there, which of
course would be the goal. So if that's the goal, if having
a more dynamic type of process is our goal, what are the steps
(06:20):
that we need to take in order toget to that point?
Yeah. So I think to, to get to real
time dynamic pricing, right, I think one of the crucial thing
that's absolutely necessary is the ability to to kind of ingest
these real time predictors, right?
I mean, so you need the market condition right now, right?
(06:40):
Not at the day ago, not an hour ago, right, But right now, right
at this moment, right? So you need to be able to ingest
real time predictors, right? And then many of them, right?
There's, there's so many, you know, I mean, for, for, for, for
example, right, if you talk about tariffs, right?
I mean, it literally could change from one hour to the
(07:00):
next, right? I mean, it just takes, you know,
someone to say, OK, I'm going tochange the tariff and that's it.
Like it, it just done, right? And that moment on it just
changed, right? So, so, so there.
And there's so many of these different predictors about the
market conditions right in, in different country and different,
you know, different customers and so on.
So, so it really requires you tobe able to ingest not only the
(07:24):
real time being just so many of them, right?
So, So what that means is that you really need a neural network
based kind of pricing kind of system that you, you, you can't
use a traditional kind of segmentation based approach
because you know, the, the minute that you get more and
more of these predictors in, youwill run into data sparsity
(07:46):
challenges. I mean, you don't have enough,
you know, historical transactionthat with, with all this
different data attributes, you know, so you, you run into data
sparsity. So you have to have this what we
call segment free approach, you know, to, to, to to kind of
predict price, predict willingness to pay price, you
(08:06):
know, for yeah, for the various customer product combination and
neural network is has been really effective at doing that.
Understood. Thank you for the explanation
there. And I know because we've been
around for to put it politely for a little while and pricing
was a very, very tough job when as you mentioned a little while
(08:28):
ago when you might issue a pricelist once a quarter, once every
six months, once a year, I mean even that was a tough job.
But when we multiply that with all of the effects that we have
going on now with all of the chaos, with all of the
volatility, this really is a very, very tough critical job.
So of course we would encourage everyone to take advantage of
(08:50):
all the tools and resources and training and neural networks, as
you mentioned that are availableat your at your fingertips
because it is a different race than it used to be.
And if your competitors have rocket ships and you have a
tricycle, that's not going to work out too well for you in
the, in the long run. So you've been a, a lecturer at
(09:11):
Cal Berkeley and we are very, very excited that you're going
to lead a full day CPP workshop with Professional Pricing
Society at PPS profitable in LasVegas in October.
And I know from your abstract inour brochures that you mentioned
several limitations of generative AI.
(09:32):
So what are some of those limitations and how can business
professionals be prepared to cover the limitations within
Gen. AI?
Yeah, I think the, well, I mean,there's several, you know,
limitations that Gen. AI is very powerful.
You know, you can definitely do a lot of things, but there's
also a lot of things that Gen. AI cannot do right.
(09:54):
And one challenge that I think most people at least heard about
it is definitely the fact that Gen.
AI will hallucinate, right? They will make up things that
you know, that they're not sure of, you know, that you know, or
they don't know anything about, right?
So basically they would just create things, right?
And that's actually the feature of Gen.
AI, right? They are able to kind of
(10:15):
generate, right? That's why they call it generate
generated AI. So they generate things, right?
So unfortunately in, you know, when the applications of each
and AI is in what we call a factbased type of application where
you need the accurate, precise, correct answer, right?
It's not very well suited for that, you know, so, so one is,
(10:36):
is definitely hallucination. And then the the second one is,
you know, really their inabilityto do with tabular data using
language alone. You know, I think that's a one
thing that Jenna I don't do verywell.
I mean, I, I didn't, you know, because Jenna, I typically what
they, the foundation that make allow them to, to work, you
(10:59):
know, so well with languages andimages and so on, is the fact
that in those data type, there'sa lot of internal correlation.
You know, by by seeing, for example, in the image, it's very
prevalent, right? If you see one pixel, a few
pixel, right, you immediately know what the neighboring pixel
will be, right? These are, they're highly
correlated, right? So that's why from seeing a few
(11:19):
pixel, you could generate the rest of it, right?
And same thing with language. If you see a few word, you sort
of know, you know in neighboringwhat the neighboring words will
be. And you could actually predict
what those words will be, right?So, but in tabular data, those
correlations are much, much weaker.
And now by seeing a few data points, right, you can't really
(11:41):
predict what the next data points or the neighboring data
points will be, right? I mean, it's actually not that
easy. I if that were easy, right,
then, you know, we will be able to predict the stock market
with, you know, precise accuracy, right?
And, and, and we know that we'redefinitely not there yet, right?
So, and in fact, you know, if you actually try to ask a
(12:02):
generative AI to, to analyse some, some data, like human data
set, right, the first thing theytry to do is they will try to
write code, you know, they will actually try to, you know, they
know that they're, they're, you know, their limitation as a
language model that they cannot deal with tablet data very well.
So they will just, yeah, start right by writing code to analyse
(12:22):
those. And these codes right now are
are, you know, relatively simple.
I don't get better. But yeah.
So, so those are, I think two major limitations of Gen.
AI. Yeah.
OK. Hey, we are learning lots about
Gen. AI, about machine learning and
about the pace of Business Todaywith Doctor Michael Wu.
(12:44):
Michael, I'm looking forward to following up with you more about
the hallucinations, about the dichotomy between how Gen.
I Gen. AI handles language versus data.
But we're going to take a quick break right now for a word from
our sponsors. But everyone, if you are
interested in what we're talkingabout today, make sure to take
(13:04):
advantage of Doctor Michael Wu'swork.
But right now, we're going to take a quick break.
We'll be right back with more. Let's talk pricing with Doctor
Michael Wu. This October 21st through 24th,
PPS Profitable in Las Vegas is all about you, the pricing
professional ready to navigate today's challenges and lead with
confidence. In just four days, you'll gain
(13:26):
the strategies, tools and connections to drive margin
growth, strengthen your influence and future proof your
career. Early Bird pricing ends
September 19th and our group rates make it easy to bring your
whole team so you can cover moreground and return with a shared
plan. Make the investment in yourself.
Visit pricingsociety.com/P PS (13:44):
LV25 to register.
All right, everyone. Welcome back to the Let's Talk
Pricing podcast. We are talking with Michael Wu
about becoming an agent of change in the age of artificial
intelligence. And Michael, before the break,
(14:06):
you mentioned 2 specific limitations of generative AI
Gen. AI. 1 is that it will
hallucinate, It will kind of be perhaps more creative than we
would expect it to be. And the other one was that it
can be very efficient at dealingwith language, predicting the
next word in a sentence, for example, but not so much with
(14:29):
data preventing predicting, I'm sorry, the next data point in a
series of data points or something like that.
So I remember seeing one of yourpresentations, I can't remember
where in the world it was, but you did a great demonstration of
an AI hallucination. You took a question from the
(14:50):
audience with someone that you were familiar with, and his or
her product group had a series of existing products, and AI
correctly identified several of these products when you gave it
a prompt, but it also invented afew products that did not
physically exist yet. So that was quite an eye opener
(15:15):
for me when I saw that presentation.
But in your view, what are the implications with AI around
hallucinations like that? Yeah, I think with respect to
pricing especially, right, I mean, I, I think that's, that's
a really critical mission critical function, right.
So you you can't tolerate, you know, these AI making up a price
(15:37):
that's not accurate or or. So what that means is really
that you know, if you are going to use Gen.
AI in these type of mission critical systems such as
pricing, right, they need to be grounded OK, so they they need
to be grounded with some other technology or some other help,
right? They need other technology to
(15:59):
help them, right? So, and, and the best, you know,
way to do that is you augment them with, for example, what we
talk about real time dynamic pricing, neural network pricing,
right? So all these are pricing systems
that can actually calculate and compute right, the precise price
for each customer product combination, right?
(16:21):
So, so once you know that, right, then you know, you can
ask, you can keep those data to the generative AI, then it can
basically present that to you orexplain it to you in a way that
that's you know, that that's useful for you, right, in the
particular context. So I think that is the best way
(16:42):
to do it is to is to use a combination of Gen.
AI and these more, I would say, you know, traditional pricing
system, right? I mean these are essentially use
them as tool right to ground kind of the actual kind of price
that these Jenny I might come upwith if you if you don't use
(17:03):
them otherwise, right. So yeah.
Understood. And if you've been around
pricing, you probably have seen the example from a little while
ago. There was a biology textbook, I
think it was about flies or spiders or something like that.
And I remember seeing that someone did a PowerPoint
(17:24):
presentation with this textbook that cost $23 million because
bookseller number one, who is inSeattle, said we are the leaders
in this field. So we're going to be 2% higher
than everybody else. Bookseller #2 who used to have
or probably still does have chains of bookstores around the
(17:44):
country and around the world, said we're going to match that
price until it became a stair step when it went all the way
up. But of course, 1 price in person
who was coherent and was awake that day would see the $23
million for a textbook does not make sense.
So we do have to have the guard rails as you mentioned there to
make sure that things stay wherewe want them to be.
(18:06):
So also, since this is an area that I need to learn a lot more
about, I'm going to ask a couplemore things about Gen.
AI and about predictive models and about so-called intelligent
pricing agents. So can you explain the changes
there? And also, can you talk about the
(18:26):
important things that we as pricing professionals that we as
revenue managers need to know about AI as we go from these
language models to more reasoning models to large
reasoning models? Yeah, I, I think, you know,
that's a a distinction that not not a lot of people understand.
(18:48):
So when you talk about Gen. AI, right, I mean, everybody is
familiar with touch PT, right? These are the large language
model. They're really good with
research strategy, you know, communication, right?
So, so it can help you with all those stuff, right.
So you can essentially now have an assistant that will help you
do market research, right? Analyse data, simple data from
(19:12):
the market. And what look like I said
before, right, these large language model, even though they
are not very good at analyzing data themselves, right?
They can actually use a tool, right?
They will use, for example, a Python compiled, they will write
the Python code. They're they're very good
language, right? So, and programming language is
a language, right? So they will write the code and
(19:34):
then, you know, have the code analyse the those data, right?
And so they will help you createstrategies, right?
If for whatever you know, condition, the market conditions
that you may have, you may need to be very detailed to specify
all all those market conditions,but it will help you create a
pricing strategy, OK to to deal with that condition.
(19:57):
And keep in mind, you know, that's, you know, because think
about these right, these language model, they have pretty
much read every single textbook about pricing.
Every single research paper you know that's been ever published
about pricing, right? So it probably knows all the
strategy in the world, right? So if you could describe your
(20:20):
condition, what you need, what you're trying to accomplish,
right? There is probably some pricing
strategy out there that is rightfor you, right?
And these large language model could surface doves and bring
that to your attention, right? You may some of them you may
heard of, right? Some of them you may not.
And for those who may not, you could just ask them explain how
this strategy work and you wouldjust tell you how it works,
(20:42):
right? So so you have this like
extremely powerful smart assistant that can essentially
help you do research, create strategy and also to help you
communicate, right? Sometimes you when you
community, when you do pricing, one of the critical role of
pricing analysts is you know, they sit cross functionally
(21:05):
within the organization. So they a lot of these cross
functional communication, right?And the way you communicate to
an executive team is not the same as the way you communicate
with, for example, a product team or marketing team or
finance team, right? So, so you have to have
different communication strategyto to maybe explain your
strategy or your research, rightresults, all that.
(21:27):
So, and a large language model can help you with all that, you
know, So, and, but as I said before, right, these large
English model, they will hallucinate, right?
So, so they, when you actually wanted to do some serious kind
of analysis, right? It cannot actually do that very
(21:47):
effectively, except you will actually leverage the ability to
write code, right? And so recently I think there's
a new kind of essentially generative AI.
So these are the what we call the large reasoning model,
right? So these are LRM instead of LLM.
They're LRM. So the large reasoning models,
(22:10):
that is, you know, the difference about them is like,
you know, I'm not. When you ask them a question,
not only did it actually give you the answer, right?
They will also reason through itto arrive at that answer.
So what that means is that you can use these large reasoning
model. They can also help you write
code, right? They can help you do analysis,
right? So, so that's like, you know, a
(22:32):
lot of the the work that I wouldsay a pricing analyst
traditional have to do, right? Whether it's research, analysis,
strategy, communication, right? You know, using generative AI,
it can help you pretty much do all those things.
One thing that I find interesting is in a lot of ways,
(22:54):
the AI revolution kind of seems like the Internet squared.
And by that I mean, I remember when the Internet was first
really starting and it was described as all of the
libraries, all of the information in the world is
available to you, but it's thrown on the floor and you have
to search through and find it yourself.
(23:15):
But it seems like now with the AI revolution that all the
information is out there. But now it can do a better job
of sorting that information, finding the information that you
need. So you're not searching through
all of the libraries on the world, in the world on your
floor. You can get focused on the ones
that you need for your strategies, for your tactics,
(23:36):
for your next steps, and it can help you compare and contrast
and pick the right ones for you in a way.
So it's interesting how you know, maybe history doesn't
repeat itself, but sometimes it doesn't rhyme a little bit, I
guess would be one way to say it.
The thing about the large, the LRMS, the large reasoning models
(23:58):
there and how they can write code and do analysis for you.
And ideally, I can certainly seethis as a tool that will let
pricing professionals focus moreon strategy, focus more on
perhaps the old Pareto principlewhere if 80% of the good stuff
of your profit and your margins come from 20% of your offerings
(24:19):
and you can concentrate on thoseand have these AI assistants do
a lot of work on the other 80%, that might be 20% of your end
product there. So it's a a very interesting
revolution, as you say, that we're going through right now.
Yeah, totally. And I think, you know, you
mentioned that, you know, now that the RM can help you write
code, right? You you can focus on strategy,
(24:42):
but the LRM can also help you dostrategy, right.
So the how the the large language model can help you.
Just remember they read every research paper on strategy, how
to you know, and, and and textbook about pricing out
there, right. So, so it can really help you do
a lot of your work really and make you a much more productive
pricing analyst. And we are going to take another
(25:05):
break for some messaging from our sponsors, but when we come
back, we are going to talk even more about what it takes to lead
in a world of intelligent systems.
We're going to talk about and strengthening the human skills
that AI cannot replace with Doctor Michael Wu, the Let's
Talk Pricing podcast. We'll be right back after these
(25:25):
messages. Thank you.
We're heading to Las Vegas this October 21st through 24th for
PPS Profitable Pricing's preeminent conference.
Four days of bold strategies, practical tools, and real world
solutions to help you boost margins, lead with confidence,
and accelerate your career. Connect with pricing leaders,
(25:46):
sharpen your skills, and leave ready to make an immediate
impact. Secure your spot and take
advantage of early bird pricing through September 19th now at
pricingsociety.com/P PS (25:54):
LV25 All right, let's get back to the
show. Welcome back to Let's Talk
Pricing. We're talking with Doctor
Michael Wu about becoming an agent of change in the age of
artificial intelligence. We've had a great discussion
where we talked about Gen. AI, we talked about large
reasoning models. We talked about how we can
(26:16):
overcome some of the limitationsof artificial intelligence.
And Mike, we have a few more questions here.
So as AI is going to take over alot of the routine work in a lot
of our jobs and in a lot of our organizations, how do human
roles evolve in pricing? Where is the place for the human
(26:40):
being with our Gray matter up here?
What is our role going to be in this going forward?
Yeah, I think that's a that's a good question.
I mean, I think it is a little bit scary to think about how
much that AI can do today. And so, and I think, you know, I
think the best way to think about this, right, is to think
(27:00):
that, you know, if you hire a stellar employee, right?
And one day it, he or she learn to, to do everything that you do
and maybe to do it better than you, what do you do?
Like what, what would you do? I mean, a good like manager, a
good boss, good leader will promote them, right?
(27:20):
You promote them, right? So, and I think that's exactly
what should happen, right? You're, and you're basically,
you know, human pricing professionals will essentially
evolve their role, right, to be a much more strategic role,
right? You'll be managing, you know, a
lot more things, right? It could be people or a bunch of
(27:44):
AI systems, right, working to oreven a bunch of AI agents,
right. You know, you may have a agent
to help you do pricing in a particular area, right?
Or maybe a particular type of product, a particular type of
packaging of of of product, right.
Or, you know, yeah, so you have different agents, different
people, right? So you'd be a manager, right?
(28:06):
To make sure that they actually work together well to achieve
the goal, right. And you'll be aligning these,
you know, different AI systems and people with the moral
ethical standard, right? For example, like, you know, AI,
it's it's highly, sometimes highly controversial.
You need to be responsible on how you price, right?
And you need to make sure that you are complying to whatever
(28:30):
domain you're in because, you know, every single company has
pricing. I say every single company has
some product or services that would they would sell, right?
Every company would need to put some price on, on price tag on
those product and services, right?
So, but you know, every company in every different industry,
(28:51):
right, had different regulations, right?
And so you, yeah, you need to make sure that these AI or
people or AI agents, right, are,you know, compliant and aligned
to the standard, the moral standard, ethical standard of
whichever country that you may be in.
(29:13):
Yes. And I can see that as kind of a
quantum physics type of a leap where you go from A to B, but
really you might not exist any place in between A&B along that
along that journey. And of course, we all know
within pricing, the change management is always very
important. And you talked about this a
little bit earlier as well. But how we use language with the
(29:38):
different groups, even the internal groups that we have to
deal with as pricing leaders, aspricing managers, as pricing
directors, that can be very different.
For example, how we would talk about a price change to a 25
year sales veteran would be different than how we would talk
about it with a brand new finance MBA from a top school.
(30:01):
And that would be different fromhow we would talk about it with
our C-Suite or our senior managers groups, all in the same
company. But those have to be 3 very,
very different discussions. And of course, we have to
influence without authority up, down and sideways along with all
of that. So I can certainly see some of
the strategic elements and some of the change management
(30:21):
elements in this quantum shift that that might be happening
there. So tell us about some of the
other mindset shifts that might need to happen with us as
pricing managers, with our pricing teams as we move from
business intelligence dashboardsto predictive to Gen. generative
(30:42):
AI insights. What are some of the things that
we're going to need to change with us as human beings along
this journey here in order to continue to thrive?
Yeah, I think you know, yeah, very good question.
I, I think, you know, most of usprobably have heard the phrase
that, you know, AI will replace you, but some other human using
(31:04):
AI will, right. So so that's the the mindset
that we have to really truly internalized.
And what does that actually mean?
What does that actually mean? You know, so as I, as we already
kind of, you know, discuss on the question before, right?
Like in to some extent, right, holding on tightly to what you
(31:27):
do, right, will actually sometimes make you more
replaceable, right? You need to essentially allow,
you know, allow this role, your role to, to kind of shift and
evolve into this, this more strategic role, right, Even
though you may be very differentfrom what you actually.
And I think that, you know, in the pricing world, I I'm, I'm,
(31:50):
as most of you probably know that I'm, I don't have a pricing
background, right? I'm AAI machine learning guy
coming to the to the pricing world, right?
And I've, I've learned over time, over the years, I've
learned that, you know, like pricing people are a little bit,
you know, they like to be in control of things, right?
They like to do things that are,you know, and, and sometimes
(32:13):
that actually could almost, I would say, hinder you from
moving on to something more strategic, right?
If you, you, you, you, I, I justwant to do my calculation all by
myself And, and I want to know exactly what's, what's going on
exactly which calculation is being made to make this change
of price, right? So I so if you are so tied to
(32:38):
it, so, so you want to be in control so much that you, you
would not let the AI take over that, that those calculation,
right? Then you'll be doing all that
yourself, right? And very, very often is that
like, if you keep doing that yourself, right, then someone
who is using AI who could do that much faster, right?
And maybe maybe even not as goodas you are, right?
(33:01):
But they could do it so much faster and they can iterate much
faster, right? So even though they, they may
not do it as good as you are in one pass, if they do it in like
5 pass or 10 pass, right, 10 iteration, like they could still
be better than than you, right? Those people who use AI who
could, you know, do this much, much faster either way, much,
much faster, we'll replace you right?
(33:22):
So, so in to some extent, right.The mindset change that we must
have is we must learn to trust AI as in some way seed control
to AI and let AI help you. We could see this in just in
business in general, right? I mean, the best leaders are not
the one who who kind of do everything themselves, right?
(33:43):
They actually delegate right to the appropriate people and
that's why they become leader, right.
So, so I think that's the the exact same mindset that pricing
professionals needs to have. Of course.
Got it. And you're 100% corrected.
Speed is a big currency within pricing.
(34:03):
And if you hang around PPS conferences enough, you'll hear
someone eventually say that a really good price right now
might be better than the optimumprice 2 weeks down the road.
So speed certainly has a big value to our customers, to our
marketplaces there. Speed is a very big customer and
(34:25):
I know I've talked with people who've done win loss analysis
and they found that the speed ofgetting a quote is one of the
biggest factors and sometimes even bigger than whether or not
the price was high or low was how quickly you were able to
respond there. So of course, we can use AI
agents and the things that we'retalking about today to become
quicker, to become more reactivewith our challenges and with our
(34:50):
opportunities there because speed definitely matters.
So that that's very interesting.So I appreciate the information
there. So we have quite a few
challenges now. And for pricing people, there
are going to be things that we as the Nexus of a lot of
different functions within our organizations need to do to
(35:11):
strengthen. And one of the things that you
mentioned was that pricing as a job certainly will change with
these new tools and we might be looking at using our strategic
insights to have more commercialinfluence within our
organizations. So what's something that a
(35:32):
pricing professional can do right now to kind of strengthen
that Venn diagram overlap, that connection between their pricing
knowledge and between their commercial influence, whether it
be with sales, with senior management, with financial
marketing or product management,all these other fields that come
into play, that kind of touch pricing as well.
(35:53):
What can we do to get better at strengthening those connections?
Yeah, I think there's, you know,a couple of things, right.
I mean, I think 1 is definitely be relevant, right?
And be relevant, meaning you have to literally adapt as
quickly as the market is changing, right?
So, and as I said before, right,the, the best way to do that is
(36:14):
to allow AI to help you because they will, they will, AI will
always be faster than than us, than than me, which is so.
And then the the other way, the other way that AI can help you
there is, you know, it's also berelevant, be relevant in the way
you communicate, right? As we mentioned before, the way
you talk to, you know, sales team, it's not the same as you
(36:37):
talk to finance team, right, or a, a product team or right.
So so you want and you need to essentially have different
communication strategy with these different teams.
And, and you know, previously, because it just takes you so
much time to just come, even come up with a communication
plan that you would just have one version that you just have
to you just kind of expect people to kind of, you know,
(37:00):
understand you right and be and be supportive of of your, your
strategy, right. But now, because you have
generated, you could, you could create these communication
strategy communication plans so quickly, so fast, right?
You can literally create, you know, A1 for each department
that you need to talk to and still have extra time leftover,
right? So, so, so, so that's why I
(37:24):
think it's crucial, as I said before, right?
But that mindset change, right? You need to learn to trust AI,
let AI help you, right? Seed control, keep, you know,
relinquish some control, right? Not all control, right?
But you need to at some point relinquish some control, seed
control to to this AI system. Yeah.
OK, understood. Thank you very much for the
(37:44):
explanation there. So as human beings, I know that
there are quite a few things andwe've talked about some of them
already. What are some of the other
strengths that we as human beings can bring to pricing and
revenue management that AI is not going to be able to replace?
What are some things that we arestill better at than the tools,
(38:06):
than the models, than Jen AI at doing within our fields?
Yeah, I think they're certainly,you know, this, this, this space
is definitely shrinking. Isn't so, so far I, I would say
like creative problem solving isstill, you know, the domain of
of that we do better human do better, right?
(38:28):
And, and I think, you know, and probably one thing that is not
going to be easily replaced by AI is the human connection,
right? How do you talk to your, your,
your colleague, right, and, and your customers or your so all
that communication, right? Maybe education, right and, and
(38:50):
teaching people like what you'reunderstand, making them
understand your strategy, right?So that required to some extent,
you know, not just the information, right, is how you
present, how you kind of help them understand, right, kind of
guide them right on this journeyto to understand your, your
strategy. And so that human connection,
(39:10):
right, probably is very hard to to be replaced, right?
And yeah, but I, I also think that this really, you know, is
don't see this as a a threat. You know, like kind of, you
know, I, I think this is actually opportunity.
In fact, I think Chrysler can actually make themselves
(39:31):
indispensable, you know, like, you know, today by being that
perfect kind of change agent forthe inevitable future, right, of
AI that that will come, right? We all know that, you know,
sometime in the future, right? Everything in a company, in the
organization will have AI, you know, to be augmented by AI
(39:56):
today. It's just like computers 50
years ago, right? Imagine what, what would the
world be, right? You know, 50 years ago, right?
And there's like, oh, computers are going to take my job and
everything, right? But you know it, yeah, it will
take some job, right? But The thing is like.
Every company will adopt computer right to to help them,
(40:16):
you know, do all kinds of stuff,right.
So this future of AI everywhere in a company will come, it will
come, right? So every company will have to
kind of adapt to this, you know,to this future right there.
There's no, no like, you know, escape from this, right.
Every company will have to learnto use AI in everything they do,
(40:39):
right? So, and I think that you know,
pricers actually have you know, are in in this perfect kind of
position to actually help them help this company move on to
this feature, right, because this feature is inevitable,
right? And every company have to move
there. And if you make yourself the
agent that could help them get to that future, you are
(41:01):
indismissible, right? You are completely
indispensable, right? Let let me turn the question
back to you, Kevin, right? And, and understand like why
pricing analysts and pricing professional are the perfect
people for this job, right? I mean, So what are the, I would
say, you know, the, the, the, I would say the strength, right.
That a or a kind of quality of agood pricing profession like let
(41:26):
me ask you that question, yeah. Yeah, absolutely.
No, that's a great question. Pricing professionals have to be
able to influence without authority in a lot of cases.
And also we have a job that is often, in spite of its
importance, it's often under appreciated, under compensated,
(41:46):
under trained. But in spite of all of that, we
basically are the Nexus for everything that goes to a
company's profits, margins, to the bottom line.
You know, every, I say every dollar, euro, yen won, whatever
your currency is, it comes into your organization has a pricing
decision attached to it. It's a tough job, a very
(42:09):
important job, and one that often a lot of people think that
they can do better than you do it themselves.
But it really is a case where some of the things that you
mentioned that we as human beings have that AI can't
replace, such as a creative problem solving, such as game
theory, such as that human connection, such as having all
these different languages. And I don't mean actual
(42:31):
languages, I mean the language you would talk to a sales
veteran versus a finance person versus a senior manager versus a
product manager and so on and soforth.
We have the ability to educate people to correct when
necessary, when we can find deals and say this deal was
great and I know you got a big bonus because of it.
(42:52):
But if we replicated this deal 1000 times, we would lose jobs.
So we can't do that. We have to be able to deliver
those messages. We have to be change agents.
We have to be change management specialists, artists and
scientists and all of the above.So I do agree with you, it is
can be a little bit scary. But hey, I'm sure a couple, 100
years ago, typewriters were scary if you were a scribe.
(43:15):
It just changes it. It will change the job.
As you mentioned, computers change jobs, the Internet change
jobs. AI is going to change jobs.
A few years ago, we heard that every company is a software
company whether they know it or not.
And according to our guests on the Let's Talk Pricing podcast
today, every company is an AI company right now, whether they
(43:37):
know it or not, it is coming. It is not going to stop.
And so we have to be in a place where we can deal with these new
tools where we as pricing professionals can use the
influencing without authority that the skills, the strategies,
the tactics that we have in order to be good at our jobs.
And so, yeah, I agree with you. In a lot of cases, we are really
(43:57):
well placed to take advantage ofthese tools to have them be a
wait for us to elevate our strategic thinking and to really
strengthen our performance to move more quickly, to move
better in a lot of cases as well.
So, yeah, I I agree with the thestatement there.
And also, I think that as a pricing person and as someone
(44:22):
who deals with thousands of great pricing revenue managers
from around the world, a lot of us do have a lot of those human
elements that you talked about that will enable us to use this
as we should as a tool in order to do our jobs even better.
Yeah, totally. And I think like all the quality
that you mentioned that makes a pricing a great, you know,
(44:44):
pricing professionals, right, isprecisely what's needed, right,
to evolve a company from its current state to this inevitable
future that will come, right? So pricing professionals are
actually, you know, if you are, are, you know, willing to kind
of change, of course, that change of mindset is, is
crucial, right? You need to essentially let AI
(45:05):
help you, right, see a little control to AI and to elevate
your role, right? Then you can actually make
yourself completely indispensable in this world of
AI. Yeah, absolutely.
And I think about pricing, people think of the changes that
we've been through, not I wouldn't even say in the last
decade and less than that, as wewere talking about earlier, it
(45:28):
may have been standard operatingprocedure for a lot of company
to have one price list once a quarter if they were advanced
once every six months, once a year or whatever.
But during the COVID pandemic, when we had the extra pressure
on our supply chains, when we had inflation creeping into
parts of the world where we are not used to inflation, that got
(45:51):
thrown out of the window. We adjusted.
We adjusted quickly. We used all of these elements
that we're talking about in order to move our companies
forward and to continue to thrive and do better and exceed
our expectations there. So this is going to be the next
big change. Totally, Yeah.
All right, So Doctor Michael Wu.Michael, thank you so much for
(46:12):
our conversation today. Learned a lot and thank you for
helping us better understand what's coming with Jen AI, with
LRMS, continuing LLMS and so much more.
Definitely learned a lot today. And for those of us who joined
us for the Left Stock Pricing podcast, Michael Wu has a lot
(46:33):
more to teach us. So make sure to check out his
sessions at PPS. Profitable will be in Las Vegas
from October 21st to 24th. Michael, I'm looking forward to
seeing you there, of course. And for our listeners out there,
you can visit pricingsociety.comto check out our information
(46:53):
about our upcoming events in LasVegas and Barcelona, also our
online training and our other offerings there.
And we will see everyone next time on the Let's Talk Pricing
podcast. Thank you so much for joining us
today.
You're listening to Let's Talk Pricing, your connection to the
voices, stories, and strategies shaping the pricing world.
Each episode, we go beyond theory into the practical,
timely conversations that help you lead with confidence and
drive results. Let's start the show.
(00:20):
Very good day everyone, and welcome to the Let's Talk
Pricing podcast. We are the official podcast of
the Professional Pricing Society.
My name is Kevin Mitchell from PPS and I'm very, very excited
about today's episode. We are going to be talking with
Doctor Michael Wu about becomingan agent of change in the age of
(00:42):
artificial intelligence. So we're going to be talking
about AI and machine learning and lots of interesting topics
there. And a lot of us already know
Michael. But for a quick introduction,
Doctor Michael Wu is the the chief AI strategist at Prose.
He is a world leading expert in artificial intelligence and
machine learning. He is, among other things, in
(01:05):
Advertiser, I'm sorry, an advisor and lecturer at Cal
Berkeley's extension AI programs.
And as part of his studies, he once used machine learning to
model the human brain. And also we're very, very
excited that Doctor Michael Wu is going to be with us at the
PPS Profitable conference in LasVegas in October.
(01:27):
He's going to have ACPP workshopon AI powered pricing strategy.
And he's also going to deliver akeynote address about the rise
of intelligent agents and the human skills that will matter
the most as AI reshapes pricing management, revenue growth
management, sales enablement andrelated fields.
(01:49):
So a very, very nice welcome to Doctor Michael Wu.
Michael, we are very excited to have you on.
How are you today, Sir? I'm good.
It's it's such a pleasure to be in conversation with you, Kevin.
Thank you so much, Michael. I appreciate that.
And we have a few questions for you.
And AI is an area where actually, believe it or not, I
(02:10):
should probably hang my head in shame as I say this.
But actually I go to PPS conferences to learn more about
AI from experts such as yourself.
It's an area that I do not know enough about yet.
So I'm looking forward to learning a lot from you.
And I know from your work and from being a follower and a big
(02:34):
fan of your work as well that you have described the shift to
strategic artificial intelligence, the shift to
strategic AI as a revolution in pricing.
So what's the key transformationthat's taking place right now in
pricing and revenue management and related fields?
Yeah, I think to answer the question, you know, you, we can
(02:58):
just look around and ask us what's happening right now in
the world, right? I mean, I think right now we all
know the world is a little bit chaotic.
It's it's, you know, and therefore in terms of, you know,
business and market, you know, it's highly volatile, right?
So it's changes all the time. And So what that means is that,
(03:18):
you know, pricing, what's the, what's happening in pricing
right now is we're trying to react to that world, right?
And the best way to deal with such a highly volatile and and
crazy unpredictable world is essentially to do real time
dynamic pricing, right? OK.
So what that really means is that you know, you know, you,
(03:41):
you have to kind of, you know, kind of compute these price
right in real time. And all that means is there's no
no more master price list, right?
So the price is actually generated at real time, right?
Because the minute you produce a, a price list, right, If it
takes time to actually produce aprice list for all the customer
(04:05):
and all the different product combination that you have,
right? And that takes a long time,
right? So, you know, you may do it very
fast, you may have AI help you to do, but that still takes
time, right? You know, even if it's just
like, you know, a week or a few day, right?
You know, it takes you time to produce that master price list,
right? And the minute you produce that
(04:25):
list, right, it's already a few days late, a weekly or monthly,
however long it takes you to produce them, right?
So that so that to get the real time like means right now,
right, you have to be able to compute and generate that time,
sorry, that price in real time, right?
Compute that. So, so, so that's why you know,
(04:46):
yeah, we we need to kind of moveon to this truly real time
dynamic pricing. I, I wanted to make a
distinction with the, the term dynamic pricing as well, because
I think if you very often when you hear real time dynamic
pricing, people say, oh, we, we,we, we do that.
You know what they really meant is that while we do dynamic
(05:06):
pricing and dynamic pricing justthat like your pricing changes
with something some, you know some indicator, some variable on
about the market, right, it changes with the but it doesn't
mean you change with that in real time.
Understood. Yeah.
And I like some of the things you said about there about our
chaotic volatile world. And of course, as pricing
(05:31):
professionals, we have a lot of externalities that come into
play with our decision making, market changes, government
changes, macroeconomic changes, customer changes, competitor
changes and things like that. And we have to take all those
into account. And so speed definitely is a a
very important variable there. And also, I like how you
(05:52):
describe that, what a lot of people think about dynamic
pricing, really they're just maybe pricing more quickly
instead of having that real, real time dynamic pricing where
everything is changing almost inan instant there, which of
course would be the goal. So if that's the goal, if having
a more dynamic type of process is our goal, what are the steps
(06:20):
that we need to take in order toget to that point?
Yeah. So I think to, to get to real
time dynamic pricing, right, I think one of the crucial thing
that's absolutely necessary is the ability to to kind of ingest
these real time predictors, right?
I mean, so you need the market condition right now, right?
(06:40):
Not at the day ago, not an hour ago, right, But right now, right
at this moment, right? So you need to be able to ingest
real time predictors, right? And then many of them, right?
There's, there's so many, you know, I mean, for, for, for, for
example, right, if you talk about tariffs, right?
I mean, it literally could change from one hour to the
(07:00):
next, right? I mean, it just takes, you know,
someone to say, OK, I'm going tochange the tariff and that's it.
Like it, it just done, right? And that moment on it just
changed, right? So, so, so there.
And there's so many of these different predictors about the
market conditions right in, in different country and different,
you know, different customers and so on.
So, so it really requires you tobe able to ingest not only the
(07:24):
real time being just so many of them, right?
So, So what that means is that you really need a neural network
based kind of pricing kind of system that you, you, you can't
use a traditional kind of segmentation based approach
because you know, the, the minute that you get more and
more of these predictors in, youwill run into data sparsity
(07:46):
challenges. I mean, you don't have enough,
you know, historical transactionthat with, with all this
different data attributes, you know, so you, you run into data
sparsity. So you have to have this what we
call segment free approach, you know, to, to, to to kind of
predict price, predict willingness to pay price, you
(08:06):
know, for yeah, for the various customer product combination and
neural network is has been really effective at doing that.
Understood. Thank you for the explanation
there. And I know because we've been
around for to put it politely for a little while and pricing
was a very, very tough job when as you mentioned a little while
(08:28):
ago when you might issue a pricelist once a quarter, once every
six months, once a year, I mean even that was a tough job.
But when we multiply that with all of the effects that we have
going on now with all of the chaos, with all of the
volatility, this really is a very, very tough critical job.
So of course we would encourage everyone to take advantage of
(08:50):
all the tools and resources and training and neural networks, as
you mentioned that are availableat your at your fingertips
because it is a different race than it used to be.
And if your competitors have rocket ships and you have a
tricycle, that's not going to work out too well for you in
the, in the long run. So you've been a, a lecturer at
(09:11):
Cal Berkeley and we are very, very excited that you're going
to lead a full day CPP workshop with Professional Pricing
Society at PPS profitable in LasVegas in October.
And I know from your abstract inour brochures that you mentioned
several limitations of generative AI.
(09:32):
So what are some of those limitations and how can business
professionals be prepared to cover the limitations within
Gen. AI?
Yeah, I think the, well, I mean,there's several, you know,
limitations that Gen. AI is very powerful.
You know, you can definitely do a lot of things, but there's
also a lot of things that Gen. AI cannot do right.
(09:54):
And one challenge that I think most people at least heard about
it is definitely the fact that Gen.
AI will hallucinate, right? They will make up things that
you know, that they're not sure of, you know, that you know, or
they don't know anything about, right?
So basically they would just create things, right?
And that's actually the feature of Gen.
AI, right? They are able to kind of
(10:15):
generate, right? That's why they call it generate
generated AI. So they generate things, right?
So unfortunately in, you know, when the applications of each
and AI is in what we call a factbased type of application where
you need the accurate, precise, correct answer, right?
It's not very well suited for that, you know, so, so one is,
(10:36):
is definitely hallucination. And then the the second one is,
you know, really their inabilityto do with tabular data using
language alone. You know, I think that's a one
thing that Jenna I don't do verywell.
I mean, I, I didn't, you know, because Jenna, I typically what
they, the foundation that make allow them to, to work, you
(10:59):
know, so well with languages andimages and so on, is the fact
that in those data type, there'sa lot of internal correlation.
You know, by by seeing, for example, in the image, it's very
prevalent, right? If you see one pixel, a few
pixel, right, you immediately know what the neighboring pixel
will be, right? These are, they're highly
correlated, right? So that's why from seeing a few
(11:19):
pixel, you could generate the rest of it, right?
And same thing with language. If you see a few word, you sort
of know, you know in neighboringwhat the neighboring words will
be. And you could actually predict
what those words will be, right?So, but in tabular data, those
correlations are much, much weaker.
And now by seeing a few data points, right, you can't really
(11:41):
predict what the next data points or the neighboring data
points will be, right? I mean, it's actually not that
easy. I if that were easy, right,
then, you know, we will be able to predict the stock market
with, you know, precise accuracy, right?
And, and, and we know that we'redefinitely not there yet, right?
So, and in fact, you know, if you actually try to ask a
(12:02):
generative AI to, to analyse some, some data, like human data
set, right, the first thing theytry to do is they will try to
write code, you know, they will actually try to, you know, they
know that they're, they're, you know, their limitation as a
language model that they cannot deal with tablet data very well.
So they will just, yeah, start right by writing code to analyse
(12:22):
those. And these codes right now are
are, you know, relatively simple.
I don't get better. But yeah.
So, so those are, I think two major limitations of Gen.
AI. Yeah.
OK. Hey, we are learning lots about
Gen. AI, about machine learning and
about the pace of Business Todaywith Doctor Michael Wu.
(12:44):
Michael, I'm looking forward to following up with you more about
the hallucinations, about the dichotomy between how Gen.
I Gen. AI handles language versus data.
But we're going to take a quick break right now for a word from
our sponsors. But everyone, if you are
interested in what we're talkingabout today, make sure to take
(13:04):
advantage of Doctor Michael Wu'swork.
But right now, we're going to take a quick break.
We'll be right back with more. Let's talk pricing with Doctor
Michael Wu. This October 21st through 24th,
PPS Profitable in Las Vegas is all about you, the pricing
professional ready to navigate today's challenges and lead with
confidence. In just four days, you'll gain
(13:26):
the strategies, tools and connections to drive margin
growth, strengthen your influence and future proof your
career. Early Bird pricing ends
September 19th and our group rates make it easy to bring your
whole team so you can cover moreground and return with a shared
plan. Make the investment in yourself.
Visit pricingsociety.com/P PS (13:44):
LV25 to register.
All right, everyone. Welcome back to the Let's Talk
Pricing podcast. We are talking with Michael Wu
about becoming an agent of change in the age of artificial
intelligence. And Michael, before the break,
(14:06):
you mentioned 2 specific limitations of generative AI
Gen. AI. 1 is that it will
hallucinate, It will kind of be perhaps more creative than we
would expect it to be. And the other one was that it
can be very efficient at dealingwith language, predicting the
next word in a sentence, for example, but not so much with
(14:29):
data preventing predicting, I'm sorry, the next data point in a
series of data points or something like that.
So I remember seeing one of yourpresentations, I can't remember
where in the world it was, but you did a great demonstration of
an AI hallucination. You took a question from the
(14:50):
audience with someone that you were familiar with, and his or
her product group had a series of existing products, and AI
correctly identified several of these products when you gave it
a prompt, but it also invented afew products that did not
physically exist yet. So that was quite an eye opener
(15:15):
for me when I saw that presentation.
But in your view, what are the implications with AI around
hallucinations like that? Yeah, I think with respect to
pricing especially, right, I mean, I, I think that's, that's
a really critical mission critical function, right.
So you you can't tolerate, you know, these AI making up a price
(15:37):
that's not accurate or or. So what that means is really
that you know, if you are going to use Gen.
AI in these type of mission critical systems such as
pricing, right, they need to be grounded OK, so they they need
to be grounded with some other technology or some other help,
right? They need other technology to
(15:59):
help them, right? So, and, and the best, you know,
way to do that is you augment them with, for example, what we
talk about real time dynamic pricing, neural network pricing,
right? So all these are pricing systems
that can actually calculate and compute right, the precise price
for each customer product combination, right?
(16:21):
So, so once you know that, right, then you know, you can
ask, you can keep those data to the generative AI, then it can
basically present that to you orexplain it to you in a way that
that's you know, that that's useful for you, right, in the
particular context. So I think that is the best way
(16:42):
to do it is to is to use a combination of Gen.
AI and these more, I would say, you know, traditional pricing
system, right? I mean these are essentially use
them as tool right to ground kind of the actual kind of price
that these Jenny I might come upwith if you if you don't use
(17:03):
them otherwise, right. So yeah.
Understood. And if you've been around
pricing, you probably have seen the example from a little while
ago. There was a biology textbook, I
think it was about flies or spiders or something like that.
And I remember seeing that someone did a PowerPoint
(17:24):
presentation with this textbook that cost $23 million because
bookseller number one, who is inSeattle, said we are the leaders
in this field. So we're going to be 2% higher
than everybody else. Bookseller #2 who used to have
or probably still does have chains of bookstores around the
(17:44):
country and around the world, said we're going to match that
price until it became a stair step when it went all the way
up. But of course, 1 price in person
who was coherent and was awake that day would see the $23
million for a textbook does not make sense.
So we do have to have the guard rails as you mentioned there to
make sure that things stay wherewe want them to be.
(18:06):
So also, since this is an area that I need to learn a lot more
about, I'm going to ask a couplemore things about Gen.
AI and about predictive models and about so-called intelligent
pricing agents. So can you explain the changes
there? And also, can you talk about the
(18:26):
important things that we as pricing professionals that we as
revenue managers need to know about AI as we go from these
language models to more reasoning models to large
reasoning models? Yeah, I, I think, you know,
that's a a distinction that not not a lot of people understand.
(18:48):
So when you talk about Gen. AI, right, I mean, everybody is
familiar with touch PT, right? These are the large language
model. They're really good with
research strategy, you know, communication, right?
So, so it can help you with all those stuff, right.
So you can essentially now have an assistant that will help you
do market research, right? Analyse data, simple data from
(19:12):
the market. And what look like I said
before, right, these large language model, even though they
are not very good at analyzing data themselves, right?
They can actually use a tool, right?
They will use, for example, a Python compiled, they will write
the Python code. They're they're very good
language, right? So, and programming language is
a language, right? So they will write the code and
(19:34):
then, you know, have the code analyse the those data, right?
And so they will help you createstrategies, right?
If for whatever you know, condition, the market conditions
that you may have, you may need to be very detailed to specify
all all those market conditions,but it will help you create a
pricing strategy, OK to to deal with that condition.
(19:57):
And keep in mind, you know, that's, you know, because think
about these right, these language model, they have pretty
much read every single textbook about pricing.
Every single research paper you know that's been ever published
about pricing, right? So it probably knows all the
strategy in the world, right? So if you could describe your
(20:20):
condition, what you need, what you're trying to accomplish,
right? There is probably some pricing
strategy out there that is rightfor you, right?
And these large language model could surface doves and bring
that to your attention, right? You may some of them you may
heard of, right? Some of them you may not.
And for those who may not, you could just ask them explain how
this strategy work and you wouldjust tell you how it works,
(20:42):
right? So so you have this like
extremely powerful smart assistant that can essentially
help you do research, create strategy and also to help you
communicate, right? Sometimes you when you
community, when you do pricing, one of the critical role of
pricing analysts is you know, they sit cross functionally
(21:05):
within the organization. So they a lot of these cross
functional communication, right?And the way you communicate to
an executive team is not the same as the way you communicate
with, for example, a product team or marketing team or
finance team, right? So, so you have to have
different communication strategyto to maybe explain your
strategy or your research, rightresults, all that.
(21:27):
So, and a large language model can help you with all that, you
know, So, and, but as I said before, right, these large
English model, they will hallucinate, right?
So, so they, when you actually wanted to do some serious kind
of analysis, right? It cannot actually do that very
(21:47):
effectively, except you will actually leverage the ability to
write code, right? And so recently I think there's
a new kind of essentially generative AI.
So these are the what we call the large reasoning model,
right? So these are LRM instead of LLM.
They're LRM. So the large reasoning models,
(22:10):
that is, you know, the difference about them is like,
you know, I'm not. When you ask them a question,
not only did it actually give you the answer, right?
They will also reason through itto arrive at that answer.
So what that means is that you can use these large reasoning
model. They can also help you write
code, right? They can help you do analysis,
right? So, so that's like, you know, a
(22:32):
lot of the the work that I wouldsay a pricing analyst
traditional have to do, right? Whether it's research, analysis,
strategy, communication, right? You know, using generative AI,
it can help you pretty much do all those things.
One thing that I find interesting is in a lot of ways,
(22:54):
the AI revolution kind of seems like the Internet squared.
And by that I mean, I remember when the Internet was first
really starting and it was described as all of the
libraries, all of the information in the world is
available to you, but it's thrown on the floor and you have
to search through and find it yourself.
(23:15):
But it seems like now with the AI revolution that all the
information is out there. But now it can do a better job
of sorting that information, finding the information that you
need. So you're not searching through
all of the libraries on the world, in the world on your
floor. You can get focused on the ones
that you need for your strategies, for your tactics,
(23:36):
for your next steps, and it can help you compare and contrast
and pick the right ones for you in a way.
So it's interesting how you know, maybe history doesn't
repeat itself, but sometimes it doesn't rhyme a little bit, I
guess would be one way to say it.
The thing about the large, the LRMS, the large reasoning models
(23:58):
there and how they can write code and do analysis for you.
And ideally, I can certainly seethis as a tool that will let
pricing professionals focus moreon strategy, focus more on
perhaps the old Pareto principlewhere if 80% of the good stuff
of your profit and your margins come from 20% of your offerings
(24:19):
and you can concentrate on thoseand have these AI assistants do
a lot of work on the other 80%, that might be 20% of your end
product there. So it's a a very interesting
revolution, as you say, that we're going through right now.
Yeah, totally. And I think, you know, you
mentioned that, you know, now that the RM can help you write
code, right? You you can focus on strategy,
(24:42):
but the LRM can also help you dostrategy, right.
So the how the the large language model can help you.
Just remember they read every research paper on strategy, how
to you know, and, and and textbook about pricing out
there, right. So, so it can really help you do
a lot of your work really and make you a much more productive
pricing analyst. And we are going to take another
(25:05):
break for some messaging from our sponsors, but when we come
back, we are going to talk even more about what it takes to lead
in a world of intelligent systems.
We're going to talk about and strengthening the human skills
that AI cannot replace with Doctor Michael Wu, the Let's
Talk Pricing podcast. We'll be right back after these
(25:25):
messages. Thank you.
We're heading to Las Vegas this October 21st through 24th for
PPS Profitable Pricing's preeminent conference.
Four days of bold strategies, practical tools, and real world
solutions to help you boost margins, lead with confidence,
and accelerate your career. Connect with pricing leaders,
(25:46):
sharpen your skills, and leave ready to make an immediate
impact. Secure your spot and take
advantage of early bird pricing through September 19th now at
pricingsociety.com/P PS (25:54):
LV25 All right, let's get back to the
show. Welcome back to Let's Talk
Pricing. We're talking with Doctor
Michael Wu about becoming an agent of change in the age of
artificial intelligence. We've had a great discussion
where we talked about Gen. AI, we talked about large
reasoning models. We talked about how we can
(26:16):
overcome some of the limitationsof artificial intelligence.
And Mike, we have a few more questions here.
So as AI is going to take over alot of the routine work in a lot
of our jobs and in a lot of our organizations, how do human
roles evolve in pricing? Where is the place for the human
(26:40):
being with our Gray matter up here?
What is our role going to be in this going forward?
Yeah, I think that's a that's a good question.
I mean, I think it is a little bit scary to think about how
much that AI can do today. And so, and I think, you know, I
think the best way to think about this, right, is to think
(27:00):
that, you know, if you hire a stellar employee, right?
And one day it, he or she learn to, to do everything that you do
and maybe to do it better than you, what do you do?
Like what, what would you do? I mean, a good like manager, a
good boss, good leader will promote them, right?
(27:20):
You promote them, right? So, and I think that's exactly
what should happen, right? You're, and you're basically,
you know, human pricing professionals will essentially
evolve their role, right, to be a much more strategic role,
right? You'll be managing, you know, a
lot more things, right? It could be people or a bunch of
(27:44):
AI systems, right, working to oreven a bunch of AI agents,
right. You know, you may have a agent
to help you do pricing in a particular area, right?
Or maybe a particular type of product, a particular type of
packaging of of of product, right.
Or, you know, yeah, so you have different agents, different
people, right? So you'd be a manager, right?
(28:06):
To make sure that they actually work together well to achieve
the goal, right. And you'll be aligning these,
you know, different AI systems and people with the moral
ethical standard, right? For example, like, you know, AI,
it's it's highly, sometimes highly controversial.
You need to be responsible on how you price, right?
And you need to make sure that you are complying to whatever
(28:30):
domain you're in because, you know, every single company has
pricing. I say every single company has
some product or services that would they would sell, right?
Every company would need to put some price on, on price tag on
those product and services, right?
So, but you know, every company in every different industry,
(28:51):
right, had different regulations, right?
And so you, yeah, you need to make sure that these AI or
people or AI agents, right, are,you know, compliant and aligned
to the standard, the moral standard, ethical standard of
whichever country that you may be in.
(29:13):
Yes. And I can see that as kind of a
quantum physics type of a leap where you go from A to B, but
really you might not exist any place in between A&B along that
along that journey. And of course, we all know
within pricing, the change management is always very
important. And you talked about this a
little bit earlier as well. But how we use language with the
(29:38):
different groups, even the internal groups that we have to
deal with as pricing leaders, aspricing managers, as pricing
directors, that can be very different.
For example, how we would talk about a price change to a 25
year sales veteran would be different than how we would talk
about it with a brand new finance MBA from a top school.
(30:01):
And that would be different fromhow we would talk about it with
our C-Suite or our senior managers groups, all in the same
company. But those have to be 3 very,
very different discussions. And of course, we have to
influence without authority up, down and sideways along with all
of that. So I can certainly see some of
the strategic elements and some of the change management
(30:21):
elements in this quantum shift that that might be happening
there. So tell us about some of the
other mindset shifts that might need to happen with us as
pricing managers, with our pricing teams as we move from
business intelligence dashboardsto predictive to Gen. generative
(30:42):
AI insights. What are some of the things that
we're going to need to change with us as human beings along
this journey here in order to continue to thrive?
Yeah, I think you know, yeah, very good question.
I, I think, you know, most of usprobably have heard the phrase
that, you know, AI will replace you, but some other human using
(31:04):
AI will, right. So so that's the the mindset
that we have to really truly internalized.
And what does that actually mean?
What does that actually mean? You know, so as I, as we already
kind of, you know, discuss on the question before, right?
Like in to some extent, right, holding on tightly to what you
(31:27):
do, right, will actually sometimes make you more
replaceable, right? You need to essentially allow,
you know, allow this role, your role to, to kind of shift and
evolve into this, this more strategic role, right, Even
though you may be very differentfrom what you actually.
And I think that, you know, in the pricing world, I I'm, I'm,
(31:50):
as most of you probably know that I'm, I don't have a pricing
background, right? I'm AAI machine learning guy
coming to the to the pricing world, right?
And I've, I've learned over time, over the years, I've
learned that, you know, like pricing people are a little bit,
you know, they like to be in control of things, right?
They like to do things that are,you know, and, and sometimes
(32:13):
that actually could almost, I would say, hinder you from
moving on to something more strategic, right?
If you, you, you, you, I, I justwant to do my calculation all by
myself And, and I want to know exactly what's, what's going on
exactly which calculation is being made to make this change
of price, right? So I so if you are so tied to
(32:38):
it, so, so you want to be in control so much that you, you
would not let the AI take over that, that those calculation,
right? Then you'll be doing all that
yourself, right? And very, very often is that
like, if you keep doing that yourself, right, then someone
who is using AI who could do that much faster, right?
And maybe maybe even not as goodas you are, right?
(33:01):
But they could do it so much faster and they can iterate much
faster, right? So even though they, they may
not do it as good as you are in one pass, if they do it in like
5 pass or 10 pass, right, 10 iteration, like they could still
be better than than you, right? Those people who use AI who
could, you know, do this much, much faster either way, much,
much faster, we'll replace you right?
(33:22):
So, so in to some extent, right.The mindset change that we must
have is we must learn to trust AI as in some way seed control
to AI and let AI help you. We could see this in just in
business in general, right? I mean, the best leaders are not
the one who who kind of do everything themselves, right?
(33:43):
They actually delegate right to the appropriate people and
that's why they become leader, right.
So, so I think that's the the exact same mindset that pricing
professionals needs to have. Of course.
Got it. And you're 100% corrected.
Speed is a big currency within pricing.
(34:03):
And if you hang around PPS conferences enough, you'll hear
someone eventually say that a really good price right now
might be better than the optimumprice 2 weeks down the road.
So speed certainly has a big value to our customers, to our
marketplaces there. Speed is a very big customer and
(34:25):
I know I've talked with people who've done win loss analysis
and they found that the speed ofgetting a quote is one of the
biggest factors and sometimes even bigger than whether or not
the price was high or low was how quickly you were able to
respond there. So of course, we can use AI
agents and the things that we'retalking about today to become
quicker, to become more reactivewith our challenges and with our
(34:50):
opportunities there because speed definitely matters.
So that that's very interesting.So I appreciate the information
there. So we have quite a few
challenges now. And for pricing people, there
are going to be things that we as the Nexus of a lot of
different functions within our organizations need to do to
(35:11):
strengthen. And one of the things that you
mentioned was that pricing as a job certainly will change with
these new tools and we might be looking at using our strategic
insights to have more commercialinfluence within our
organizations. So what's something that a
(35:32):
pricing professional can do right now to kind of strengthen
that Venn diagram overlap, that connection between their pricing
knowledge and between their commercial influence, whether it
be with sales, with senior management, with financial
marketing or product management,all these other fields that come
into play, that kind of touch pricing as well.
(35:53):
What can we do to get better at strengthening those connections?
Yeah, I think there's, you know,a couple of things, right.
I mean, I think 1 is definitely be relevant, right?
And be relevant, meaning you have to literally adapt as
quickly as the market is changing, right?
So, and as I said before, right,the, the best way to do that is
(36:14):
to allow AI to help you because they will, they will, AI will
always be faster than than us, than than me, which is so.
And then the the other way, the other way that AI can help you
there is, you know, it's also berelevant, be relevant in the way
you communicate, right? As we mentioned before, the way
you talk to, you know, sales team, it's not the same as you
(36:37):
talk to finance team, right, or a, a product team or right.
So so you want and you need to essentially have different
communication strategy with these different teams.
And, and you know, previously, because it just takes you so
much time to just come, even come up with a communication
plan that you would just have one version that you just have
to you just kind of expect people to kind of, you know,
(37:00):
understand you right and be and be supportive of of your, your
strategy, right. But now, because you have
generated, you could, you could create these communication
strategy communication plans so quickly, so fast, right?
You can literally create, you know, A1 for each department
that you need to talk to and still have extra time leftover,
right? So, so, so, so that's why I
(37:24):
think it's crucial, as I said before, right?
But that mindset change, right? You need to learn to trust AI,
let AI help you, right? Seed control, keep, you know,
relinquish some control, right? Not all control, right?
But you need to at some point relinquish some control, seed
control to to this AI system. Yeah.
OK, understood. Thank you very much for the
(37:44):
explanation there. So as human beings, I know that
there are quite a few things andwe've talked about some of them
already. What are some of the other
strengths that we as human beings can bring to pricing and
revenue management that AI is not going to be able to replace?
What are some things that we arestill better at than the tools,
(38:06):
than the models, than Jen AI at doing within our fields?
Yeah, I think they're certainly,you know, this, this, this space
is definitely shrinking. Isn't so, so far I, I would say
like creative problem solving isstill, you know, the domain of
of that we do better human do better, right?
(38:28):
And, and I think, you know, and probably one thing that is not
going to be easily replaced by AI is the human connection,
right? How do you talk to your, your,
your colleague, right, and, and your customers or your so all
that communication, right? Maybe education, right and, and
(38:50):
teaching people like what you'reunderstand, making them
understand your strategy, right?So that required to some extent,
you know, not just the information, right, is how you
present, how you kind of help them understand, right, kind of
guide them right on this journeyto to understand your, your
strategy. And so that human connection,
(39:10):
right, probably is very hard to to be replaced, right?
And yeah, but I, I also think that this really, you know, is
don't see this as a a threat. You know, like kind of, you
know, I, I think this is actually opportunity.
In fact, I think Chrysler can actually make themselves
(39:31):
indispensable, you know, like, you know, today by being that
perfect kind of change agent forthe inevitable future, right, of
AI that that will come, right? We all know that, you know,
sometime in the future, right? Everything in a company, in the
organization will have AI, you know, to be augmented by AI
(39:56):
today. It's just like computers 50
years ago, right? Imagine what, what would the
world be, right? You know, 50 years ago, right?
And there's like, oh, computers are going to take my job and
everything, right? But you know it, yeah, it will
take some job, right? But The thing is like.
Every company will adopt computer right to to help them,
(40:16):
you know, do all kinds of stuff,right.
So this future of AI everywhere in a company will come, it will
come, right? So every company will have to
kind of adapt to this, you know,to this future right there.
There's no, no like, you know, escape from this, right.
Every company will have to learnto use AI in everything they do,
(40:39):
right? So, and I think that you know,
pricers actually have you know, are in in this perfect kind of
position to actually help them help this company move on to
this feature, right, because this feature is inevitable,
right? And every company have to move
there. And if you make yourself the
agent that could help them get to that future, you are
(41:01):
indismissible, right? You are completely
indispensable, right? Let let me turn the question
back to you, Kevin, right? And, and understand like why
pricing analysts and pricing professional are the perfect
people for this job, right? I mean, So what are the, I would
say, you know, the, the, the, I would say the strength, right.
That a or a kind of quality of agood pricing profession like let
(41:26):
me ask you that question, yeah. Yeah, absolutely.
No, that's a great question. Pricing professionals have to be
able to influence without authority in a lot of cases.
And also we have a job that is often, in spite of its
importance, it's often under appreciated, under compensated,
(41:46):
under trained. But in spite of all of that, we
basically are the Nexus for everything that goes to a
company's profits, margins, to the bottom line.
You know, every, I say every dollar, euro, yen won, whatever
your currency is, it comes into your organization has a pricing
decision attached to it. It's a tough job, a very
(42:09):
important job, and one that often a lot of people think that
they can do better than you do it themselves.
But it really is a case where some of the things that you
mentioned that we as human beings have that AI can't
replace, such as a creative problem solving, such as game
theory, such as that human connection, such as having all
these different languages. And I don't mean actual
(42:31):
languages, I mean the language you would talk to a sales
veteran versus a finance person versus a senior manager versus a
product manager and so on and soforth.
We have the ability to educate people to correct when
necessary, when we can find deals and say this deal was
great and I know you got a big bonus because of it.
(42:52):
But if we replicated this deal 1000 times, we would lose jobs.
So we can't do that. We have to be able to deliver
those messages. We have to be change agents.
We have to be change management specialists, artists and
scientists and all of the above.So I do agree with you, it is
can be a little bit scary. But hey, I'm sure a couple, 100
years ago, typewriters were scary if you were a scribe.
(43:15):
It just changes it. It will change the job.
As you mentioned, computers change jobs, the Internet change
jobs. AI is going to change jobs.
A few years ago, we heard that every company is a software
company whether they know it or not.
And according to our guests on the Let's Talk Pricing podcast
today, every company is an AI company right now, whether they
(43:37):
know it or not, it is coming. It is not going to stop.
And so we have to be in a place where we can deal with these new
tools where we as pricing professionals can use the
influencing without authority that the skills, the strategies,
the tactics that we have in order to be good at our jobs.
And so, yeah, I agree with you. In a lot of cases, we are really
(43:57):
well placed to take advantage ofthese tools to have them be a
wait for us to elevate our strategic thinking and to really
strengthen our performance to move more quickly, to move
better in a lot of cases as well.
So, yeah, I I agree with the thestatement there.
And also, I think that as a pricing person and as someone
(44:22):
who deals with thousands of great pricing revenue managers
from around the world, a lot of us do have a lot of those human
elements that you talked about that will enable us to use this
as we should as a tool in order to do our jobs even better.
Yeah, totally. And I think like all the quality
that you mentioned that makes a pricing a great, you know,
(44:44):
pricing professionals, right, isprecisely what's needed, right,
to evolve a company from its current state to this inevitable
future that will come, right? So pricing professionals are
actually, you know, if you are, are, you know, willing to kind
of change, of course, that change of mindset is, is
crucial, right? You need to essentially let AI
(45:05):
help you, right, see a little control to AI and to elevate
your role, right? Then you can actually make
yourself completely indispensable in this world of
AI. Yeah, absolutely.
And I think about pricing, people think of the changes that
we've been through, not I wouldn't even say in the last
decade and less than that, as wewere talking about earlier, it
(45:28):
may have been standard operatingprocedure for a lot of company
to have one price list once a quarter if they were advanced
once every six months, once a year or whatever.
But during the COVID pandemic, when we had the extra pressure
on our supply chains, when we had inflation creeping into
parts of the world where we are not used to inflation, that got
(45:51):
thrown out of the window. We adjusted.
We adjusted quickly. We used all of these elements
that we're talking about in order to move our companies
forward and to continue to thrive and do better and exceed
our expectations there. So this is going to be the next
big change. Totally, Yeah.
All right, So Doctor Michael Wu.Michael, thank you so much for
(46:12):
our conversation today. Learned a lot and thank you for
helping us better understand what's coming with Jen AI, with
LRMS, continuing LLMS and so much more.
Definitely learned a lot today. And for those of us who joined
us for the Left Stock Pricing podcast, Michael Wu has a lot
(46:33):
more to teach us. So make sure to check out his
sessions at PPS. Profitable will be in Las Vegas
from October 21st to 24th. Michael, I'm looking forward to
seeing you there, of course. And for our listeners out there,
you can visit pricingsociety.comto check out our information
(46:53):
about our upcoming events in LasVegas and Barcelona, also our
online training and our other offerings there.
And we will see everyone next time on the Let's Talk Pricing
podcast. Thank you so much for joining us
today.
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.
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
Betrayal: Weekly
Betrayal Weekly is back for a brand new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. Please join our Substack for additional exclusive content, curated book recommendations and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience and healing. Your voice matters! Be a part of our Betrayal journey on Substack. And make sure to check out Seasons 1-4 of Betrayal, along with Betrayal Weekly Season 1.