Revisionist History - L’Oréal and IBM: AI-Powered Beauty from Smart Talks with IBM
Episode Date: August 28, 2025This week we're sharing an episode from Malcolm's other podcast, Smart Talks with IBM. The episode takes you behind the scenes at L'Oréal’s research center in New Jersey. Malcolm Gladwell... delves into the complexities of cosmetic formulation and the AI partnership with IBM. Learn how AI is poised to revolutionize the creation of beauty products, to make them even more sustainable and innovative. This is a paid advertisement from IBM. The conversations on this podcast don't necessarily represent IBM's positions, strategies or opinions. Visit us at https://www.ibm.com/think/podcasts/smart-talksSee omnystudio.com/listener for privacy information.
Transcript
Discussion (0)
This is an I-Heart podcast.
Barbie's gone blockbuster.
OnlyFans is rewriting the rules for creators,
and ESPN is leading the game in sports media.
I'm Chris Grimes, the FTs, LA Bureau Chief,
and on September 17th and 18th,
I'll be hosting the Financial Times Business of Entertainment Summit
in West Hollywood,
where CEOs from Mattel, OnlyFans, ESPN, and many more
will tackle the future of entertainment.
Head to FTT.com slash entertainment
to unlock your exclusive
F.T. Discount. Use code FT. Podcast to save 20%.
To understand why the Cosmetics
Super Giant L'Oreal Group is teaming up with IBM,
you must first take a closer look at its products.
Take lipstick, for example. It's one of those things
that seem straightforward. A waxy cylinder that you rub on your lips to
turn of a different color. Easy, right?
Well, maybe not. As my colleague Lucy Sullivan found
out when I sent her an assignment to L'Oreal's North America Research and Innovation Center.
All right. I'm reporting live from the L'Oreal visitor's parking lot. Malcolm told me that
he would be sending me to Paris, France for this L'Oreal excursion, but instead I am in Clark,
New Jersey, passed a lot of strip malls on the way here. But to be fair to Clark, New Jersey and
L'Oreal, this is a beautiful compound. It kind of looks like
a spa. Lucy went into the center and was
blown away. The facility houses about 600 scientists and
experts across skin care, makeup, fragrance, hair care, innovative
packaging and tech. It is one of the largest formulation lab
spaces in the industry. It's the size of six basketball courts.
The reason L'Oreal's facility is so big and has so many people is that everything
L'Oreal does to bring a product to market happens here, from molecule discovery and product
development to consumer testing. The center even has its own manufacturing. My conception of
lipstick, that it's just a waxy stick, was plain wrong. Lipstick is a high-performance product
born from years of research, consumer insights, and precision science. Lipstick isn't simple.
It's incredibly complex. And one of the main reasons it's so complex is just the nature of
fashion trends. The kind of lipstick consumers want is constantly changing. A lot of our consumer
insights with L'Oreal is like, where are consumers going in the future? This is Nadine Gomez. She's
vice president for L'Oreal's research and innovation development team. Our chemists are working
on five, six years down the line. We predicted that consumers wanted more of a softer look on
their lips as well. How do you predict something like that? We see slow signals from fashion houses and
social media and things like that.
We kind of see that trend evolving a little bit
and then we know five, six years
it's going to become big.
Lucy talked with her about the origins
of one of their products,
Mabeline, matte ink, liquid lipstick.
Our competitors had two steps.
The first step is a base code.
It's super opaque.
You get the color and you get the mattity.
But it's very, very dry in your lips.
You cannot wear that honestly more than 10 minutes.
It feels like your lips are like aching at one point.
So we had to develop a top.
and you'll see many of our competitors did the same thing.
It's like a bomb. You put it on top, it's super comfortable.
But we also notice that consumers kind of get tired of reapplying a bomb.
So we're like, what can we do to create this two-step into one step?
So L'Oreal had a challenge.
How do you make a comfortable liquid matte lipstick that doesn't require consumers to reapply a top layer of bomb?
Solving this type of problem takes a lot of resources and a lot of expertise.
And crucially, it takes time.
time. Remember, Nadine said that working on a breakthrough product such as Matt Inc. can take
years before it comes out. But can this process be accelerated, taken further, be even more
sustainable? That's what IBM and L'Oreal are hoping to find out. My name is Malcolm Globwell.
You're listening to the latest episode of Smart Talks with IBM, where we offer our listeners
a glimpse behind the curtain of the world of technology. In our last episode,
we talked about how an AI assistant created with IBM Watson X
helps future teachers practice responsive teaching
by simulating classroom interactions with students.
In this episode, we take you on an even more unexpected journey
into the world of cosmetics, hair care, skin care, fragrance, makeup,
and how a custom AI model could help L'Oreal's researchers
shape the future of what we put on our faces every morning.
I want to stay on lipstick a moment longer
to help illustrate what goes into L'Oreal's product development.
And let's focus on matte ink lipstick.
L'Oreel wanted to create something that was comfortable
and could be applied in one step.
So to go from two-step to one step,
we had to look cross-functionally
and try to figure out what can we bring into the product
to make it more comfortable.
And luckily, we have many days.
different types of products at L'Oreal.
That's Alex Good, a senior chemist
who leads the lip products team
in North America. She says
the trick to making Matt Ink work
was finding an elastomer,
a substance they were already using
in foundation.
We have this elastomer that can give you
more comfortable and make it
feel like there's something on your lips, like a cushion.
She handed Lucy two jars.
The first jar contained the former version of the product
that was used in Superstay, 20.
By the way, this is exactly why I sent Lucy to the lab in my place, the samples.
And I actually have something for you to try here.
So you can try this is what was the initial product.
Okay, so this is like, sort of looks like, okay, it is clay.
It looks like Vaseline that has like a more of a color.
It's kind of a beige.
Looks like some skin.
Okay.
So this was from the two steps.
This would go on after.
I would try it on your hand.
Oh, okay.
All right.
I feel it. Okay.
So it feels like very wet.
Yeah.
As you can see, it's kind of, it's going to absorb into your skin and leave,
and then you're going to feel the dryness of the product once it's gone.
Okay.
So then we're going to move from the clay product that you have on your hand now to the elastomer.
I'll let you try that one.
This jar held the elastomer that L'Oreal had spent years developing in the lab.
This one is it clear?
Looks like aquifold.
Oh, much clearer.
And you can feel like a physical layer that you're putting on your hands.
Yeah, so that's much thicker.
It kind of like clumps together.
Yeah.
It's more of a cloudy.
It's less shimmery though.
Mm-hmm.
That's intended.
Yes.
So this is a, like a powder that's dispersed in dimethicone,
and it creates like a comfort on your lips.
It feels like there's something there for a barrier to keep the film form on.
Mm-hmm.
And that's like the key ingredient that came from foundation that we transferred into lipstick
to give us this innovative product ahead of the market.
Yeah.
This is what gives it comfort.
So the difference between Superstay 24 and matink is really the comfort.
They both last a long time.
But this matink, you don't have to apply the bomb over and over again.
So you can apply matink once for the day and you're good.
Alex Good is underselling it here.
Once for the day and you're good, that's a liquid lipstick revolution.
Literally, millions of L'Oreal consumers around the world have worn Matt Inc.
It's a blockbuster.
It's also a marvel of science.
The world's first liquid lipstick was developed in the 1930s, and it was actually just a stain
for your lips.
It barely counts as lipstick.
Then came another wave of liquid lipstick when they were able to make it mad.
That was a two-step version.
It felt heavy on your lips.
You had to keep reapplying the top coat.
it was inconvenient.
L'Oreal tackled that challenge in the lab
with chemists like Alex and Nadine leading the charge.
Their breakthrough, Matt Inc.
But creating Matt Inc. took a long time,
trial and error, the hard work of scientific experimentation.
As Nadine told Lucy,
the lipstick team had to put the new product through extensive tests.
We do a very robust stability system here.
You know, we have color, odor appearance.
We monitor this in extreme.
conditions. We simulate in 45 degrees Celsius, and that can be something like a three-year shelf life,
I'm saying. We simulate your real-life product. Like, if you leave your lip gloss in a car in
Arizona, it's 112 degrees for three days, is it still going to perform? Is it going to smell? Is it
going to look granted? Is it going to change colors? We do all that. See what I mean.
Lipstick is complex. Most people would never consider it a piece of technology, but one lip product
has millions of data points.
So much science behind.
And you can see here
how many scientists we have.
Some of them have PhDs,
some of them have master's degrees,
chemistry, biology, psychology also.
When I first heard about this collaboration
between L'Oreal and IBM,
I was surprised.
I thought, these are two very different companies.
What do they really have in common?
Pleasure to meet you guys?
Pleasure.
To find out, I went to the IBM Research Center
outside New York City,
which I have to say
is one of the coolest buildings I've ever been.
it, a semi-circular modernist masterpiece with a long curving wall of windows, looks like something
out of a Stanley Kubrick movie. I was there to talk with two experts from research and innovation
at L'Oreal, Mathieu Cassier and Gabriel Bertoli. Mathieu is VP for Digital and Transformation.
Gabriel is the chief digital transformation officer for formulation. These are the people
whose jobs are to oversee big changes within the company. And Matthew told
me to try on some lipstick.
I'm going to make you try this one.
Okay. This is Superstay.
Vinil ink.
vinyl ink.
Yeah, so that's a glossy finish.
I've never in my life put on lipstick. I have no idea what I'm doing.
You don't have to put it. You can try it virtually.
Oh.
This may not be news to people who buy makeup, but it was news to me.
You can try on L'Oreal products virtually.
They call it augmented beauty.
Oh my goodness.
That is the strangest thing I've ever seen.
I look quite fetching.
fetching.
I think it's quite amazing.
Yeah.
And I can just hit.
You can choose your color, absolutely.
So I'm on a little app.
It's looking at me and it's just showing me exactly how I would look with different shades of lipstick.
So the old idea of going into a store and trying on each one, you can do that from home if you're not even at the store.
Yeah, absolutely.
That's all purpose.
If you want to match a trend, I would go for something more like peach.
You think I'm a peach person?
I don't know.
Yeah.
that looks kind of natural. It just is enhanced. It's giving me a boyish air I would not otherwise
have. This is why Loreal says it creates beauty products and beauty experiences. Loreal is a beauty
tech company. Over the last decade, Loreal has seized the power of AI and more recently
generative AI. Technology has become a driving force alongside science and creativity. And while
some of this digital technology is relatively new, Mathieu helped me see that IBM and L'Oreal
have always had a lot in common.
So the original creator of L'Gene Schuller was a chemist in 1909, so 116 years ago,
and he created this new air-color type for the market in France.
And then little by little, it has been always a very scientific company.
So if you look a little bit at key facts, we invented sun filters in the 1930s.
There was a very, very big milestone where we also invented not only product, but a reconstructed skin.
So if you look at 1979, we've been decorated this reconstructed skin that helped us to go out of animal testing very fast.
And by the way, before the law even asked it to cosmetic companies.
And then more recently, because it's a history of innovation, we launched some new molecules like one that you can find in La Roche Pose, Melabee,
which is really helping people to find again some, you know, spots
that they could have on their skin.
It's all about, like, pigmentation, how to regulate it.
L'OLE and IBM were both started in the early 20th century.
L'OLE in 1909 and IBM in 11.
Both companies have long-standing histories of innovation,
of using trial and error to improve everything they do.
The two companies have been doing that in parallel for more than a century,
until recently.
When does it start? When do L'Oreal and IBM start working together?
So we started in 2023 at the end of the year, but, you know, really the discussion...
Oh, this is really recent.
Absolutely, absolutely. It's really recent. In reality, you know, I would say the first really interaction happened at the beginning of 2024.
This is Gabriel Batole, who I spoke to alongside Mathieu.
What really played a key role here is we wanted to bring, from a logic perspective, to
R&D together, which normally, you know, companies like us, you just go to a provider, you
know, it's a customer and a supplier and you work, they deliver to you. Here the concept was
totally different. Mitu said that the collaboration began with simple conversations. So if you
look at the way IBM entered into L'Oreal labs, it started by interviewing people. What would
help you to do your job? What is your business need? So it was, by the way, the two months of
a long series of interviews.
And from all the people around the world we have in research,
in Brazil, in India, in China, Japan, US, France, of course.
So we really want to make sure that at the end of the day,
this new model, this new tool that will give to people,
is really people-centric in the way that it serves their daily need.
More to the point, L'Oreal has leveraged technology for decades
and accumulated a mountain of scientific knowledge.
everything from consumer aspirations and market trends
to the results of all the experiments conducted during product development
to which formulations melt in a hot car.
It's hard to get your head around.
L'Oreal isn't just a cosmetics company,
it's a beauty data powerhouse.
If we have 16,000 terabyte of data,
coming from consumer insights,
coming from market research,
coming from sales.
Well, with the new technology,
maybe by aligning those two
and using best-in-class technology,
you can solve that problem.
So you say you have 16 terabytes of data.
Put that in perspective.
How much data is that?
Give me...
This is 100 year of L'Oreal data
based on the last, you know,
40 years of data in the systems.
So this is really, I mean,
we're talking about 100 years,
of data that only L'Oreal have.
Let's take the example of the lipsticks.
I mean, you know, if lipsticks can be
between 20 and 30 row material,
each raw material will have, I would say,
10 or 15 way of doing things.
Gabriel is talking about how things used to be done.
Researchers at L'Oreal
needed roughly 25 ingredients
for a new lipstick formulation,
but they have to choose from a pool of hundreds,
if not thousands of raw materials.
And even after they settle on the ones they want,
they have to figure out how much of each ingredient they need.
And in what form?
What molecular weight?
What combination?
It's not just a math problem.
It's a problem that requires balancing multiple perspectives.
Safety, performance, quality, compliance standards,
sustainability, and more.
It can take years.
But what if you could simulate hundreds of cars?
parked in a sweltering heat?
What if you could do all those trials and errors virtually
over and over and over again?
What if, instead of mixing materials together by hand,
you could ask AI to predict what combinations might work best
and then try those out first?
This is 10 on the power of 25.
Yeah, yeah.
This is 100 billion of years for a human to do a change in the formula
or the possibility they have,
you can only do this by using technology,
power of technology, and data that you have.
This, Mathieu says, is where IBM can come in to help take things further.
Using artificial intelligence, IBM can help L'Oreal create a custom AI model
that helps to crunch those numbers,
to be a companion to the researchers to give them superpowers.
We don't want to replace the intuition of the scientists.
We just want to make sure that,
This intuition is really augmented by some calculation power that, as Gabriel said,
can do those 10 at the power of 25 solutions and say,
hmm, probably try this one, this one, this one, this one.
It looks like a better solution.
And then ultimately, that's really the decision of the chemistry to make it happen.
Well, to make a predictive AI model that can give L'Oreal researchers those superpowers,
you'd need that mountain of data, years' worth of laboratory testing,
and all L'Oreal's data digitized and AI-ready.
You'd need to train artificial intelligence on everything the company has already done
in order for it to predict what it could do.
L'Oreal has 100 years of data, 50 years of digitized data.
This is Mariam Ashuri, Senior Director of Product Management for IBM Watson X.
L'Oréal has the data, and part of IBM's job is to help put that data to work,
which involves ensuring data quality.
Merriam talked about the concept of AI-ready data.
The sole purpose of this data engineering pipeline
is to clean the data,
and we call them AI-ready data,
makes them ready to be consumed by AI.
So basically looking into biases in the data
to fix the distribution,
looking into darkness that we are putting into place
in terms of removing personal information.
And Miriam then explained that a custom model, like the one IBM is creating with L'Oreal,
can be more efficient and targeted than the larger general-purpose AI models.
You've heard about large language models.
The reason that they call them large language model is they are exposed into really large amount of data.
So the larger the model, the more capable, the models are, but also the larger computed requires.
that translates to an increase.
Carbon footprint and energy consumption that translates to an increase.
Latency, that's your response time, that translates to an increase cost.
So we started seeing that enterprises started grabbing a much smaller model,
customize it on their proprietary data,
that's their domain specific data or the data about their users,
to create something differentiated that is applicable to a real-world use case,
but also delivers the performance that they needed for a fraction of the cost.
And that's why there's been a lot of push around using custom models
versus very large and all-purpose models.
So how is a custom model created?
Merriam says you start with a base model.
Imagine you're buying a car.
You could get a minivan or a sedan or a sports car.
And then you get to customize it.
You could add a sunroof, leather seats, or a rear-view camera.
Turns out you can do the same thing with your AI model.
You pick a base and then you customize it.
You tune it on the data unique to your organization.
We do believe that one model doesn't fit all use cases.
You want to truly have access to any model anywhere.
And by any model anywhere, I really mean any model anywhere.
Open source, proprietary, local at your machine.
Wherever the model is, you want to host it yourself.
because then you would be able to take advantage of the best of the technology at any point
and pick the right model for the target use case.
So a custom model tuned on L'Oreel's data would be more targeted and efficient than a general purpose model.
It would understand a researcher's world and provide transparency into its workings.
That's part of the magic.
And what could a custom AI foundation model do for a company like L'Oreal?
The goal of this model is to tame the complexity of the formulation.
That's Guillaume Loha Malin, an IBM distinguished engineer and one of the people working on the AI model.
And to help, I would say, the formulator to go not only past, but also, I would say, be able to include more complexity also in their formulation,
more personalization, more sustainability, better select the ingredient.
So it's really a tool to help them and to also help them also to unleash the creativity.
Guillaume is saying that with its custom AI model,
L'OREL could improve every step of its product development pipeline,
make the process faster and more sustainable.
But he's also saying that the model could help L'Oreal create something that's never been done.
before. What could that product be?
So I'm warning you that so of my questions
are going to be really dumb.
Okay. No, please. By all means.
You're right. All right. Right. To find out what people at L'Oreal are dreaming of,
I spoke with Tricia Ayigari, Global General Manager at L'Oreal's
Mabelene brand. And I asked her about her own dreams
and how technology and science could help bring those dreams into the world.
Do you have a secret wish list of things you think that this partnership could produce?
Like, is there a product out there that's been technically too difficult that you think could be a worthy target?
Oh, yes.
There is one that I think could be really amazing.
What's that?
So shine products in general are harder to create.
And we're unable to create a shiny, long wearing eye shadow.
So basically, like a shadow that could stay on your eyelids, that won't settle into creases, that won't move all over your.
your face that has a glossy effect.
It's like the Holy Grail.
That's the Holy Grail.
Yeah.
Yeah.
You may have seen that look in fashion shows.
But that look isn't real.
Not for people like me and Lucy anyway.
If you're walking down a runway, you see a lot of makeup artists doing techniques where they put some
eye shadow on.
They layer Vaseline over it.
Or like slather Vaseline on somebody's eyes to create this very like glossy look.
But, you know, within five minutes after they walk down the runway, I'm sure it's all
over their face or being washed off.
So the look is kind of more of like a fashion look that we've been unable to create in real.
Real consumers can't wear it because it would get it everywhere.
Trisha had another thing on her wish list too.
The other that we would really like is semi-permanent makeup.
So we've talked a lot about really, really comfortable, thin film makeup that you could wear all over your face and that you can sleep in.
and that it will last a couple of days, basically.
So whether it be on your face, on your lashes, on your brows,
so anything that's like more of a semi-permanent,
meaning lasting for three days or more, would be amazing.
Yeah, yeah.
You say those two things, how long have they been on the wish list of L'Oreal?
Oh, my gosh.
I have been trying to develop this shiny eye shadow since I started,
what year did I start?
Like 2010, and I'm sure many people had asked before me,
and we tried so many iterations of it,
and nobody's been able to achieve it.
It's clear that L'Oreal's experts like Tricia
have a lot of ideas.
I once did what I called a magic wand project,
where I called up scientists and technologists
in as many different fields as possible,
and asked them what they could create
if they could just wave a magic wand
and make it real.
And everyone had some.
something they'd want to create. Everyone. That's not the issue. The issue is that there are a million
different impediments to make the ideas on the wish list real. Lack of resources, lack of time. Some
crucial bit of know-how is lacking. There's a gap between what we want and what we can actually
have. And one of the simplest ways to think of the promise of AI is that it can narrow that gap.
Not close it, of course, but do enough that people with dreams realize there are more things within their
grasp than they could ever have imagined.
Smart Talks with IBM is produced by Matt Romano, Amy Gaines McQuaid, Lucy
Sullivan, and Jake Harper.
We're edited by Lacey Roberts. Engineering by Nina Bird Lawrence.
Mastering by Sarah Brugger, music by Grammiscope.
Special thanks to Tatiana Lieberman and Cassidy Meyer.
Smart Talks with IBM is a production of Pushkin Industries and Ruby Studio at IHeartMedia.
To find more Pushkin podcasts, listen on the IHeart Radio app, Apple Podcasts, or wherever you get your podcasts.
I'm Malcolm Glaba.
This is a paid advertisement from IBM.
The conversations on this podcast don't necessarily represent IBM's positions.
strategies, or opinions.