GenAI Pricing: Hype or High-Stakes Game Changer?

Machine learning has long enhanced pricing, but distinguishing Generative AI's practical applications from hype is challenging. To identify true GenAI opportunities, practitioners should consider: (1) untapped unstructured data sources for price decisions; (2) areas for content automation to enhance pricing; (3) uses for engaging chatbots to improve pricing processes.

Stephanie Yee is a Partner at Bain & Company, where she exclusively serves clients on the topics of pricing and profitability. She has led multiple successful pricing transformation programs and is a former pricing and sales senior executive at a Fortune 75 company. She has published and spoken extensively on the topic of Pricing.  Stephanie holds a Management Information Systems degree from Texas A&M University.

Episode Transcript

(00:00):
All right. Thank you all so much for tuning
into yet another episode of the Professional Pricing Society
podcast. My name is Terrence and today we
have a super special guest with us who is also going to be one
of the keynote speakers at our upcoming conference in Las Vegas
this fall, which is October 22ndthrough the 25th.
Her name is Stephanie Yee. She is a partner at Bain and

(00:22):
Company where she exclusively serves clients on the topics of
pricing and profitability. She has LED multiple successful
pricing transformation programs and is a former pricing and
sales senior executive at a Fortune 75 company.
Stephanie holds a Management Information Systems degree from
Texas A&M University as well. Miss Stephanie, how are we doing

(00:45):
today? I'm doing great.
How are you, Terrance? Doing very, very well, dear.
We're going to be talking today about Gen.
AI pricing hype or high stakes game changer.
And so you have a plethora of knowledge and you have a
plethora of experience, which iswhy you're going to be super
popular keynote for this year's conference.

(01:07):
And so I want to thank you firstof all for taking the time with
me to kind of share a bit of a teaser, if you will, on this
podcast for what we're going to be discussing in the fall.
Is that correct? That's right.
Terence, thanks for having me today.
I really appreciate the time. Yeah, not a problem at all.
So let's go ahead and jump into the conversation, you know, AI

(01:27):
and pricing. What is new about AI?
About G and AI specifically in your opinion?
Yeah, it's right because it's, it's interesting because AI is
feels like a very old topic, butalso a very new topic in the
space. So, you know, it's a matter of
fact, pricing has actually been in a functional area, one of the

(01:50):
most likely places across commercial functions to have
introduced some type of AI and ML capabilities.
So it's not necessarily new. Most of your audience I think,
knows that pricing actually has quite a bit of science behind
it. But there are some new
technologies available with Gen.AI that's bringing this

(02:11):
conversation to the forefront. You know, just to take a step
back, traditional AI can be broadly characterized as really
good at math solving for specific narrow task, you know,
typically requires a lot of datato build and train models.
And Jen AI actually differs fromtraditional AI in a couple of

(02:32):
different ways. Jen AI at the core of it is
based on large language models. And so therefore it's actually
really good at reading and synthesizing unstructured data.
So what does that mean? It means it's, it's really good
at like reading through call transcripts, articles, basically

(02:54):
text to understand and synthesize and summarize what
that means. And because it's able to do
that, it's all the, it's also really great at creating
unstructured data. And so with this capability, you
know, Jen AI, we're able to actually develop top tracks,
explanations, things like that. So it's actually quite good with

(03:17):
writing text among other things,but it's a little bit different
from the traditional AI capabilities that we've
historically used in pricing. And, you know, AI is just one of
those things that is continuing to evolve as time progresses.
And now that it's in the realm of pricers, in the realm of
pricing, you know, how do you see this, you know, new

(03:40):
technology being applied to to pricing in recent years?
And what do you foresee in the upcoming future?
Yeah, yeah. So we think Gen.
AI actually specifically will unlock new capabilities for
pricing that just really wasn't possible before.
And specifically we see 3 broad use cases that we're

(04:02):
particularly excited about. And I'm, I'm happy to talk
through, you know what these are.
You know, the first one is enabling price setting, which is
one of the core activities you do as a price, as a pricer.
We think this new capability will help bolster existing
traditional AIML capabilities where they have fallen short

(04:25):
when it comes to price optimization and price setting.
So I mean, it's no secret kind of if you look out into the
marketplace, there have been companies that are absolutely
been successful using traditional AIML approaches to
derive pricing. I mean, Amazon's probably the
best example that everybody knows both in B to B and in B to

(04:46):
C, they use a data-driven approach.
They run tests and experiments. They ingest data from many
different places to really optimize price.
But when we take this example ofusing AIML to drive price
optimization and we survey companies who have undertaken,

(05:06):
you know the project and and theprogram to actually develop
these capabilities through AIML.We actually see that laggards as
compared to market leaders are 2.5 times more likely to still
lack effective pricing guidance when they use these traditional
AIML approaches. And what we learned is that many

(05:31):
of them failed to get the full value of the program.
And this actually happens for three reasons.
One is that AIML approaches typically use a lot of
historical internal data to develop the right price
recommendation, which altogetherisn't wrong or bad because those
historical price points are actually tested in the

(05:52):
marketplace. But sometimes it doesn't really
account for things that are happening here and right now and
other external data sources thatactually might improve your
outcome. The second is we still see that
there is a disconnect between pricing and sales.
This is an age-old issue in the world of pricing where sales

(06:12):
doesn't fully trust the guidance.
So then therefore they're not using it.
And we also see that there's disparage and fragmented data
across the ecosystem that can beused in pricing systematically,
but is not usually incorporated because it's difficult to use.
And so the capability that is unlock through Gen.

(06:34):
AI can really help address each of these shortcomings.
And, and here's how. So this first one I talked about
where traditional AIML really focuses on historical data.
You know, imagine a world in which you're actually able to
bolster your price recommendations or more on more

(06:55):
recent data that could impact price.
So say for example, you know, tomorrow out in the news, your
competitor announces that they're going to build new
supply, you know, or if there's a supply chain disruption
somewhere, you know, in the marketplace.
As an example, we worked with a chemicals client that previously

(07:17):
invested a significant amount oftime in building a pricing
guidance tool based on historical AIML capabilities.
But through the pandemic and theUkraine war, the supply demand
dynamics changed massively, as you can imagine.
And the price recommendations based on past data just wasn't

(07:38):
good. You know, because it doesn't,
it's not relevant. There's new things that are
happening in the marketplace that actually should drive a
different price recommendation. So they ended up, you know, kind
of moving away from their a AIMLtool and started doing things in
in spreadsheets in Excel to try to like, you know, really
account for these latest trends.Well, fast forwarded now with

(08:01):
this capability with Gen. AI, they're actually able to
capture data from some of these external sources.
So imagine being able to bring that in Gen.
AI as a language model, being able to synthesize, hey, this is
happening in the marketplace. There's new capacity coming up.
There's competitor price actionswe, you know, that that we're
now learning about and ingestingthat can help them basically

(08:25):
alter their price recommendationthat they would have
historically provided to really kind of answer questions around
is the market going to be long or short?
What's this company's position in the marketplace?
Should they be pushing towards like a spot deal or should they
actually tie in volumes on contract because things are
going to be long, you know what,what should the prices really

(08:47):
be? And so by ingesting some of this
external data that was hard to really synthesize and, and, and
then bring into the pricing recommendations, able to
actually improve, improve the quality of their recommendations
by bringing both, you know, the traditional AI and, and some of
these newer jet AI capabilities together.

(09:10):
I'll give you another example ofone that we've been working on
with a different client where they're actually using their own
customer service data to inform pricing decisions.
So we've recently worked with a client and they're using their
Gen. AI capabilities to expand their
data set and bring an insight oncustomer service issues and

(09:32):
delays because actually when you're doing pricing, very often
times it's a reflection of the value prop and the service that
you provide. And so if there's been issues
with the services that you're providing, that context is
actually really important. And so we're giving those kinds
of information to the account executives so they have a much
clearer picture of the negotiation landscape when they

(09:55):
go to talk about pricing and do their negotiations.
And so as I said before, we're finding that using both
traditional AIML techniques to lean the best you can from the
past, do the math to get a data-driven decisions, but also
combining that with some of these newer capabilities with
these language models really provides the most powerful

(10:18):
outcomes for price optimization.The second thing is any price is
like, it's not enough just to set prices.
You actually have to work on getting the prices.
And this is especially, especially relevant in B to B
where there's typically, you know, some kind of salesperson
that sits between, you know, theprice and negotiating with the

(10:38):
customer. And so you don't always actually
get the prices that you set because if some of that value
gets negotiated away. So the second use case we see
that Jenny, I can really help out with is, is actually with
price getting. So one of the greatest sources
of margin leakage comes from contract non compliance.

(11:03):
And so historically, you know, when you're working with a
customer, you know, if you've got contracts, you develop
contracts and inside of these contracts you'll have different
terms and details. And most of this stuff is
actually locked up in PD, FS andWord documents.
And that makes it really hard toknow if the customers are

(11:24):
compliant against your agreed upon terms.
So terms like payment, you know,you're supposed to pay in a
certain amount of days, you havethe ability to do price
escalations if you know input costs raises above a certain
level or there's like right delivery terms, it's OK, I'm
able to charge you if I need to,you know, rush deliver something

(11:44):
to you. Using both AIML and Gen.
AI, we're actually now able to systematically read through
these contracts and extract those terms out.
And that's really powerful because it's much easier to
analyze whether or not there's compliance against the terms

(12:04):
once you've extracted that. Now I can take that and compare
that to like, well, how many times have I charged you for
freight? Am I, you know, getting the full
value out of it when it's in a Word document, it's very
difficult at scale to do that. But when you're able to extract
those terms, imagine to like an Excel or something like that,
then it becomes a lot easier to say, hey, these sets of

(12:25):
customers, we agreed to these terms, but they're not following
it. And therefore have this much
margin dollars that I could be getting that I'm not getting.
So as an example of this, I recently worked with a
healthcare client and we helped them identify 300 bits of
improvement, uplift a money owedto them to contract clients.

(12:48):
It was collecting on late fees that they could have, making
sure that people were paying on time, things like that.
And when you identify this kind of value, we were, it actually
be able to do 2 things with them.
One is what we call kind of likeringing the cash register.
So it's like, hey, the your customers owe you money on these
things. Like actually go get that.

(13:09):
That's like money that drops straight to the bottom line.
But the second thing we were able to do from a longer term
perspective was say, Oh, well, you have these really beneficial
terms, but it's only in these five, you know, contracts or 10
contracts. Like why shouldn't you be
thinking about applying it to all of your contracts, you know,

(13:30):
and how do you in your negotiation process and as you
work with that customer, move them towards these beneficial
terms or at least drive like a give, get conversation on that.
And so we find that most of the times our clients are pretty
inconsistent in the way that they apply these beneficial
clauses in their contracts. And this exercise really helped

(13:54):
bring to life where there could be more consistent and we were
able to like communicate, you know what that value of being
more consistent would be. So that's another exciting use
case that Gen. AII think will unlock in the in
terms of price getting. And then the last and the third

(14:14):
use case is with sales enablement.
And this is where I think Gen. AI actually really shines and
can help in several different right ways.
Most pricers will tell you that getting you know sales to trust
and use pricing guidance is one of the toughest changes to make
in an organization. And one of the things that Jenny

(14:36):
I can do is provide because of that text language capability
summaries and explanations of price that can help with seller
gain confidence in price recommendations.
So there's a lot of focus right now on how do you develop sales
Co pilots that help them really like improve and be better at

(14:57):
the things that they're doing intheir job.
Well, you can imagine a world inwhich these capabilities
actually help a seller understand why is it priced this
way. Going back to some of the stuff
I said earlier, what's happeningin the marketplace that's
driving these prices, you know, and it's able to have a two way
almost chat like dialogue to say, hey, give me a summary of,

(15:19):
of why we've we've, why are prices what it is, you know,
what are the talking points and all those good things.
And in today's world, a lot of that stuff is very manual.
Some pricing or sales OPS team is trying to build that stuff
and it's not very dynamic, you know?
And so you can imagine a world in which you can actually

(15:40):
greatly increase like the trust and understanding of pricing
with the seller through some of these capabilities.
And this really creates what we call a democratization of
insights, which is really the fancier way of just saying that
they have access to insights that previously they would have
had to go to a pricing analyst or somebody like that to get and

(16:02):
to understand. And now they can, you know,
self-serve on, on, on, on some of these capabilities.
The second thing I think that Jenny I can do that really
supports sales is Jenny I can develop really compelling
marketing and sales collateral that really speaks to the value

(16:24):
proposition of the product, the service that's in line with the
prices paid. So you can imagine, you know, if
a seller is able to compellinglyarticulate the value of the
product or service that they're selling, then you know, the
customer feels good about the pricing that they're actually

(16:45):
getting. It makes sense.
The prices paid are consistent with the value that they think
that they're getting. And this capability is not only
better with Jen AI, it's a lot faster.
So in fact, we recently worried what worked with a client to
increase the speed at which they're able to create these

(17:06):
good sales and marketing collateral.
And think about this in terms oflike the emails you need to
send, you know, the, the PowerPoint presentations, all of
those good things. Their original turn around time
for initial copy was basically reduced from 5 days using a
marketing agency to two days. Wow.

(17:28):
So great efficiency gains. And you know, if you kind of
read up on Gen. AI, they'll say that one of the
most compelling capabilities is that they are really good at
developing just comprehensive and compelling arguments for,
you know, whatever it is that that you've prompted them to do.
And then I think the third way in which an AI is really helping

(17:52):
enable sellers is helping them prepare through negotiation
training. So a lot of pricing value is
actually eroded away during the negotiation process with the
customer, as you can imagine. And now there are actually AI
assisted self learning modules that can take into account a

(18:13):
sales reps like previous responses as they're going
through this training and it'll generate a customer response for
them to practice with. So I think in all of these
different ways, Jen AI is actually going to really help
upskill the seller, which in turn will actually increase.
I think you know the customer value at what the you know what

(18:37):
the customer value is to the work.
OK. So essentially the new
technology being applied in pricing, specifically with
generative AI is super beneficial.
And according to you, you know, it helps with different time
efficiency, marketing, collateral, negotiation,

(18:59):
trainings. I mean, there's just a a
plethora of things that this is going to help pricers out with
moving forward. And you know, as as companies
continue to really grab a hold of utilizing this tool as best
they possibly can and is their task a little bit more time
efficient as far as completing those tasks.

(19:19):
My question is, you know, as miraculous and as awesome as
something like Gin AI is, how doyou suggest or what do you
advise to, to get started in working in artificial
intelligence? Yeah, that's right.
And so we think that there's four key steps to getting
started, OK. So the first is identifying

(19:43):
where the money and the value isin the business to unlock.
You know, on this first step, weactually strongly encourage that
pricing teams look beyond just the specific pricing use cases
and actually think more broadly about the business outcomes to
unlock that would be of most value to the organization.

(20:05):
So think of this not as like. Hey, I want to set my sight on
just improving price recommendations, but a more
aspirational goal that is looks more like actually want to
increase win rates by X percent for the organization.
We want to increase renewal rates by X percent.
We want to reduce bid response time by this amount of time and

(20:29):
pricing no doubt will be a component of that solve, but
there will also be other capabilities required.
And we think it's important to set your sights a little bit
broader to create the win and the organizational energy behind
the effort. Because I mean, just frankly, if
you touch US executive, but I want to improve my how I set

(20:52):
prices versus hey, I want to increase our win rates, but it's
just a different level of engagement that you get from
those conversations. So we think the first thing is
know where the value is. The second step is you got to
figure out where you are in terms of your org readiness to

(21:12):
be able to utilize some of thesetools.
So make no mistake, as great as AI and Gen.
AI and all these new technologies are, they are
tools. Tools enable a strategy.
It doesn't make a strategy. So if you don't know how you

(21:32):
want a price in the future as a business, a tool's not going to
fix that. You know, you will end up
codifying your same old pricing practices if you don't think
through and kind of define what that future state should look
like. And then you'll be left
wondering, OK, well, why didn't that tool work?
Well, it's because you codified your all your old practices that

(21:55):
actually wasn't already working with this new technology.
You definitely will need new data sets as we can have talked
about. You'll need new tools with Gen.
AI and other things like that. You'll need new architecture for
how that data interacts with allof your systems.
You'll need actually probably different talent to enable this

(22:15):
work. And you'll actually need
commitment from your leadership to drive this change through and
to get actually the resourcing that you need to do this right.
So it's really important that you know, you identify the
value, which is the first step, but the second step is you need
to have a clear vision of what your starting point is so that
you know where you can go. You know, what's really more

(22:39):
immediate next step versus long term aspirational in your road
map. So once you know where the value
is and what your organizational capabilities are, then you can
start to prioritize well, what things can I actually tackle
near term and what things are probably kind of a little bit
longer in the road map. So we want to get to those

(23:01):
things, but there's more foundational things we need to
do 1st. And there are so many use cases
you can choose from. And so prioritization is really
paramount. We see organizations who are
early pioneers of this work falling into a couple of traps.
And so when you think about prioritization, one of the

(23:24):
things we see as a trap is doingwhat is easy versus what's
valuable. So we've seen clients who've
started to do this work on theirown, and they'll enable
something that's like, oh, well,you know, this would be easy to
do, but there's actually not a very clear ROI on actually doing
that work. And the success metrics may not

(23:45):
be very clear either. And so then, you know, it's very
hard to see like, well, did I get value out of this?
Should I keep doing these things, you know, and so that's
one trap. The other one is 1 I kind of
touched upon earlier. It's like just starting with the
use case that's so small that it's hard to really have
meaningful impact. And so that's why we

(24:06):
recommending not just looking atjust a pricing use case, but
maybe a constellation of use cases that delivers an overall
business outcome. And ideally you'd actually have
it tied to a common set of usersfrom a change management
perspective. So you're kind of making the
change holistically and enablinglike a constellation of use

(24:29):
cases makes it easier to create and measure step change success.
So now we're not no longer talking about, oh, I, you know,
improved prices for this many transactions.
It's more like actually this warhelped us change our win rate
from 10% to 11%, which is more of a step change success, which

(24:50):
is actually incredibly importantwhen you're building early
momentum for this kind of work in an organization.
And so if you know where the value is, you know where your
starting point is and you've started to prioritize and you
have a sense for like, OK, this is these are the things, this is
what the road map's going to look like.
The last thing you need to do isactually prepare your

(25:11):
organization for change. And so that means several
different things. One, you need to have clear
rules and responsibilities for the team that's going to support
the program. And that means they're clear on
they're just who like who's going to make The Who has
decision rights, who's accountable for execution.
The second is, you know, you're going to have to think about

(25:31):
your org structure that supportsa change to make sure that it
creates like sustained, long lasting changes.
And so it's like, you know, how centralized should some of these
capabilities be? Should they be be be you LED,
you know, like some of those decisions have to be have to be
thought through. You need a good change pro
management program and culture in place that identifies the

(25:53):
right sponsors, change agents, activities and communications
that really foster a data culture and improves Gen.
AI and AI literacy across the organization.
Like people have to understand what are these technologies?
Why are we using them? I mean, because I think at the
core of it, you know, people areworried about change.

(26:15):
And especially when you talk about Gen.
AI and AI, people are worried about, does this mean you're
replacing me? You know?
And so having that dialogue around, you know, what we're
trying to do in terms of improving outcomes, how these
technologies can be used, and each person's role in that
journey is going to be very important.
I mentioned earlier that having the right talent is going to be

(26:39):
really important. You're going to need talent that
understands, you know, AI skillsand you're going to have to
probably hire for some of these these talent gaps because most
people don't have necessarily, you know, these kinds of skill
sets inside of their organization today.
And beyond just hiring, you actually need to set up programs

(27:00):
to retain and continuously develop this talent.
So they want to stick around. And lastly, you'll need a
governance structure that actually sets up policies on how
to govern the data. Yeah, how to govern the
investments who track results ofpilots deployed and, you know,
really work on fully embedding like responsible AI in the

(27:22):
target governance framework. So there's quite a few things
you need to do to really prepareyour organization for change and
also bring them along in the change journey.
That's good. That's really good.
I'm also glad you said originally that this is a tool
and this is not something that the companies need to fully rely

(27:42):
on. We still have to put in the work
in the effort to strategize and they come up with a plan since
around our pricing goals and everything in that nature.
And so it's I'm also grateful that you mentioned that.
And I even think about this, youknow, a lot of companies don't
have an individual or personnel in their organization that is

(28:02):
familiar with AI or a Gen. AI.
And and even if they do, what are they doing to continue that
individual or those group of people to retain those, those
those people? And so that's that's mind
boggling. But I think it's time continues
to progress programs, things of the things of that nature will
continue to kind of come to surface and give companies more

(28:27):
reason to invest in those to be able to retain such individuals.
That's a good point. They mentioned mentioned as
well. But where do you see this
headed? You know, because we've already
come such a long way, but it feels like, it feels like it's
at the starting point, to be honest.
Yeah, it's interesting because you know, AIML has been around
and so, you know, I think my talk was like, is it high?

(28:50):
You know, is it, you know, high stakes game changer or not?
And, and the reality is, is thatthere are many use, you have
many examples of traditional AI being very successful and
definitely tried and true. Some organizations do it better
than others. And as I mentioned earlier in in
the talk, you know, Jenny, I, I think is actually only going to

(29:10):
continue to help actually improve the outcomes.
If I'm being perfectly honest, Ithink that Jenny, I right now
people are experimenting with things, but it's probably a
little bit still more hyped, youknow, than in reality.
But this space is moving so quickly.
It's one of those things where it's like, oh, I can ignore it,

(29:32):
you know, for the next like whatever couple of years.
Because the reality is, is that the capabilities with Gen.
AI has propelled the topic of AIin general to the forefront of
the business world. I mean, in my work with clients,
I have so many, you know, we hear so many board members, PE
owners, private equity owners that now want to know from the

(29:55):
management teams like, hey, how are you planning on leveraging
these capabilities just to create a sustained advantage in
the marketplace? And as I shared, like, you know,
using both the traditional and the Gen.
AI capabilities, that's only going to continue to grow.
I don't care what sector you're in.
You know, even if it's slow, it's going to continue to grow.
And in some places it's actuallygoing to move pretty rapidly.

(30:18):
And we know that market leaders already experimenting with new
ways to unlock value for their organizations through these new
capabilities. So my advice is like, don't get
caught flat footed. You know, want to start
experimenting, investing and thinking about these

(30:38):
capabilities because it's going to take time for you to probably
build all the things that you need to internally and get
things, you know, moving in the right direction.
And so I think, you know, now isthe time, if you haven't
already, to be seriously thinking about how these
technologies can be used to really up your game in your
business. That's good.

(30:58):
That's good. OK, awesome Gen.
AI pricing hype or high stakes game changer.
Miss Stephanie, thank you so much for your time today.
We are super excited to have youas one of our keynote speakers
for this upcoming fall conference.
I mean, you carry such a a tremendous amount of insight in
this particular topic and you have a lot, a lot of experience

(31:20):
behind you. And so we're super grateful and
excited to have you. One more question for the
listeners. Where can they go to learn more
about you, the company you work for?
Any resources you might want to kind of promote?
Where can they go to learn more about those things?
Yeah. So obviously people can find me
on LinkedIn, on bain.com. We we have information about our

(31:44):
pricing practice and all of the good work we we do there.
And so those are ways in which you're more than welcome to
reach out and we can continue the conversation.
All right. Well, thank you again for your
time today and listeners. And so next time we'll see you
guys later. Have a good one.
Bye bye.