November 4, 2025 • 46 mins
The source provides an extensive overview of Nestlé’s massive workforce reduction announced in October 2025, detailing the elimination of 16,000 jobs—both white-collar and blue-collar—over two years. This large-scale layoff is presented as a crucial consequence of aggressive, company-wide AI implementation, which includes algorithmic pricing engines, generative AI for marketing, and advanced robotics on factory floors. The episode explores the immediate fallout, such as the company’s stock surge and widespread employee panic, while also examining the sophisticated AI technology stack responsible for replacing roles in finance, analytics, and manufacturing. Ultimately, the document frames Nestlé’s actions not as an isolated incident but as a global precursor to mass automation displacement across the consumer-goods sector and beyond, driven by investor demands and rising commodity costs.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Welcome to the deep dive. Today. We're opening the file
on well, a really pivotal corporate decision. It wasn't just
about efficiency. It felt more like a watershed moment, didn't
It really defining this battle between human labor and artificial intelligence.
Speaker 2 (00:16):
Absolutely, and the setting is key here. Yeah, we're not
talking Silicon Valley labs. No, we're talking Vevet, Switzerland. Yeah,
you know Quiet Scenic Nesle's global.
Speaker 1 (00:27):
Headquarters, right, the home of chocolate and coffee.
Speaker 2 (00:29):
Basically exactly. And it was October sixteenth, twenty twenty five. Nisle,
this giant stable company operating in what one hundred and
eighty eight countries, They dropped this announcement and it just
sent tremors really across every corporate sector on Earth. This
wasn't you know, a new coffee flavor or some temporary merger,
you know, this was fundamentally it's a fundamental, aggressive structural overhaul. Yeah.
(00:50):
And the driver purely the promise of well silicon efficiency.
Speaker 1 (00:54):
Let's jump straight to that central number, the one that
really made everyone set up. We are talking about sixteen thousand.
Speaker 2 (00:59):
Jobs thousand, yeah, cut over two years.
Speaker 1 (01:02):
And when you think about their total workforce, about two
hundred and seventy thousand people globally. That's nearly six percent.
Speaker 2 (01:09):
It's huge.
Speaker 1 (01:09):
That scale alone is massive, But the breakdown that's what
really hit hard, I think for investors and workers.
Speaker 2 (01:16):
Precisely because for decades, right when you thought mass layoffs,
you thought factory floor, Yeah, blue collar, blue collar. But
here twelve thousand of those rules were white collar finance specialists,
marketing directors, HR people, data analysts, the office job, the
office jobs, the traditional blue collar posts, the factory floors,
the warehouses. That was the other four thousand. So it
(01:39):
completely flipped the historical script.
Speaker 1 (01:40):
It really did, and it makes that statement from the CEO,
Philip Navtil sound particularly well chilling in retrospect.
Speaker 2 (01:48):
What did he say exactly?
Speaker 1 (01:49):
He said, the world is changing and Nesley needs to
change faster. And the key instrument he pointed to for that.
Speaker 2 (01:54):
Speed, let me guess AI.
Speaker 1 (01:56):
AI driven automation. Yeah, from you know, sophisticated predictive dashboards
all the way to these incredibly versatile generative tools. This
wasn't just cost cutting like we saw in the twenty tens.
This felt new.
Speaker 2 (02:07):
You're absolutely right, this isn't about chasing cheaper labor. Offshore anymore.
This is about eliminating the need for well human thinking
in that middle layer of the corporate structure, the operational hierarchy.
Speaker 1 (02:20):
So our mission today for this deep dive, it's not
just about Nesle's restructuring, is it.
Speaker 2 (02:26):
It's bigger, much bigger.
Speaker 1 (02:28):
Yeah.
Speaker 2 (02:28):
We're framing this, based on our sources, as the opening
salvo in a global war between carbon based workers and
silicon efficiency. That's the phrase they use.
Speaker 1 (02:38):
O wow, Okay, So we're going to unpack the specific
tech that Nesle had been kind of quietly weaponizing for
years ago. Yeah, and the financial pressures of the geopolitical
stuff that kind of forced their hand. And crucially, what
this blueprint means for you, the listener, whether you're in tech, finance, logistics,
because this playbook it's going to be copied, isn't it.
Speaker 2 (02:58):
Oh? Absolutely, it's already happen. And if you need proof
that the corporate world validated this human cost, just look at.
Speaker 1 (03:05):
The market reaction instant validation.
Speaker 2 (03:07):
Nesle stock jumped nine point three percent in a single day.
The efficiency janes, not the social impact, were immediately priced
in by global finance. That tells you everything.
Speaker 1 (03:17):
Okay, So let's track the actual chaos as it unfolded.
This deep dive really has to start with that forty
eight hour timeline, the corporate shockwave, because this wasn't some smooth,
perfectly planned announcement, was it. It started with a leak,
a massive leak, right.
Speaker 2 (03:31):
It wasn't clean at all. The initial phase was just well,
pure uncontrolled internal combustion.
Speaker 1 (03:36):
Really yeah.
Speaker 2 (03:37):
It kicked off late on Wednesday, October fifteenth, specifically eleven
point four to two pm Central European summertime, late at night,
very late, and these internal planning slides detailing something called
Project Accelerate Headcount Optimization catch you name, very corporate speak,
they somehow got leaked onto global internal slack channels.
Speaker 1 (03:57):
Global, so not just one office.
Speaker 2 (03:59):
No, no, across multiple time zones almost instantly. We're talking Salpolo, Singapore, London, everywhere.
And the key thing here is this information wasn't meant
for broad consumption yet, not at all.
Speaker 1 (04:11):
So it just flooded the internal comms exactly.
Speaker 2 (04:15):
It flooded these channels, creating immediate unofficial panic before management
could even get ahead of the story. Our sources say
that within minutes those slack channels were basically unusable, just
a mess of reactions, fear questions.
Speaker 1 (04:29):
Okay, so internal chaos first. Then came the attempt to
control the narrative, the public facing response on Thursday morning.
Speaker 2 (04:36):
Right the CEO Navartil. He moved fast but unconventionally, at
seven am, before any official pres elease hit the wires.
Speaker 1 (04:45):
What did he do?
Speaker 2 (04:46):
He posted directly to LinkedIn.
Speaker 1 (04:47):
LinkedIn okay, yeah.
Speaker 2 (04:49):
Talking about quote simplifying our organization and automating our processes.
It felt like a preemptive strike, positioning the cuts as
you know, futuristic optimization rather than just painful job losses.
Speaker 1 (05:00):
And the reaction on LinkedIn at that early.
Speaker 2 (05:03):
Hour it was huge, forty eight thousand reactions before most
people in Europe had even had their coffee.
Speaker 1 (05:08):
Wow.
Speaker 2 (05:09):
It really signals a shift in how corporations communicate, doesn't
it Using a professional social platform to immediately frame something
really controversial as innovation, speaking directly to the professional class
the investors first.
Speaker 1 (05:21):
But the actual hard numbers, that sixteen thousand figure that
was almost deliberately buried in the official releases later that morning,
is that fair? Well?
Speaker 2 (05:30):
Completely fair? At eight fifteen am, the investor relations team
put out the official Q three numbers, and the headline.
Speaker 1 (05:35):
Was good, right they beat expectations?
Speaker 2 (05:37):
Yeah, organic growth was strong plus four point three percent.
It beat the market consensus estimates by one hundred and
ten basis points.
Speaker 1 (05:43):
Okay, hold on one hundred and ten basis points for listeners,
maybe less familiar with market jargon. What does that really mean?
How good is that?
Speaker 2 (05:51):
It means they significantly outperformed what analysts expected. A basis
point is tiny one hundred of a percent, So beating
consensus by one point one percentage points on growth, that's
a big deal.
Speaker 1 (06:03):
And that good news provided cover.
Speaker 2 (06:05):
Perfect camouflage the sixteen thousand job cuts, the full time
equivalent reduction. It was, as we noted buried page nine
of the investor deck, a massive strategic shift hidden behind
really quite stellar quarterly performance.
Speaker 1 (06:17):
But the financial world they read the fine print, don't they.
Speaker 2 (06:20):
They saw it, oh instantly. By nine thirty am, barely
an hour after that Q three release, Zurich Kintonalbank upgraded
Nisley's stock to.
Speaker 1 (06:28):
Buy, and they gave a reason, a.
Speaker 2 (06:30):
Very specific reason. Their analysis targeted CHF one to eighteen
per share, and they explicitly cited quote margin upside from
AI led SGNA leverage is underappreciated.
Speaker 1 (06:42):
Okay, let's slow down on that phrase, because that really
is the financial translation of this whole story. AI led
SGNA leverage. What's SGNA leverage?
Speaker 2 (06:50):
Right? Sg and A is selling general and administrative expenses,
basically the white collar.
Speaker 1 (06:55):
Stuff, marketing it finance.
Speaker 2 (06:58):
Exactly all those corporate overheads. The twelve thousand rolls being
cut falls squarely into that bucket. Leverage in this context
means increasing your output, so more sales, better pricing, while
drastically reducing the costs required to support that output.
Speaker 1 (07:12):
So the bank was saying.
Speaker 2 (07:13):
The bank was saying, we see a huge, currently undervalued
opportunity for nesle to make more money using far fewer
human staff in these support roles. They saw the layoffs
not as a crisis but as pure immediate margin expansion potential,
cold hard numbers.
Speaker 1 (07:27):
So while the stock race is soaring in Zurich, the
actual physical impact is starting to land thousands of miles away.
Speaker 2 (07:33):
Yeah, the reality check hit hard in the traditional industrial
places that very same afternoon, by two point pm, Unite
the Union in York in the UK.
Speaker 1 (07:45):
The Kitcat factory, that's.
Speaker 2 (07:46):
One they had to schedule an emergency meeting. Why because
they were anticipating one hundred and eighty operator jobs disappearing
just from the Kitcat production lines.
Speaker 1 (07:55):
So it wasn't just the offices, it was hitting the
factory floor simultaneously.
Speaker 2 (07:59):
Absolutely, automation push was across the board, from the Corbett
HQ right down to the production lines.
Speaker 1 (08:05):
And then by Friday, this whole thing explodes into the
global public conversation, not really driven by the company anymore,
but by the people affected.
Speaker 2 (08:14):
It went completely viral. CNN had the headline the hashtag
hashtag Nestle layoffs was trending everywhere, over one point two
million posts. But the moment, that piece of content that
really defined the human side of this crisis. It came
from Manila a TikTok video, right yeah, a call center
agent's TikTok video.
Speaker 1 (08:31):
And it got massive views, for.
Speaker 2 (08:33):
Ten million views in less than twenty four hours. It
was incredible.
Speaker 1 (08:36):
What was the message? What made it resonate so much?
Speaker 2 (08:39):
It was heartbreakingly specific and just so relatable to anyone
in that kind of middle layer administrative job. Her role
was scheduling the Espresso promotional campaigns, you know, a routine,
rules based task, and her lament was simple. She basically said,
my job scheduling espresso promos. Now chat GPT does it
(08:59):
in an eleven seconds. That video just perfectly captured the
terrifying speed at which these very specific, one seemingly secure
jobs were suddenly becoming redundant everywhere.
Speaker 1 (09:10):
That TikTok story is so critical because it underlines that
this isn't some future threat. That tech is here, it's operational,
and it's actively taking jobs right now. So let's look
inside this automation arsenal. You mentioned Nesla was quietly weaponizing
AI for five years. This wasn't an overnight thing. Where
was this work happening.
Speaker 2 (09:26):
Well, the central hub, the engine room, if you like,
is the Nesle Digital Factory, the NDF. It's based in
los Anne, Switzerland, and it's a serious operation. Employs about
twelve hundred data scientists.
Speaker 1 (09:37):
Do one hundred. Wow.
Speaker 2 (09:38):
Yeah. And what makes their impact so well destructive to
human roles right now is the sheer speed of deployment.
Get this. They developed forty two new generative AI use
cases in twenty twenty.
Speaker 1 (09:52):
Four alone, forty two in one year.
Speaker 2 (09:54):
Yep. And these aren't just minor process tweaks. We're talking
and to in replacements for complex humans processes.
Speaker 1 (10:01):
And these are the tools directly responsible for many of
those twelve thousand white collar cuts we talked about. Can
you break down some specifics, like how did they target
core functions like marketing?
Speaker 2 (10:10):
Okay, so, one of the first major killer apps, as
our sources called it, was the promo Optimizer. This single
piece of software directly replaced three hundred and twenty trade
marketing planners globally.
Speaker 1 (10:22):
Trade marketing planners. That that sounds complex. What do they do?
Speaker 2 (10:25):
It's incredibly complex. It's deciding precisely when, where, and how
much to discount, say Maggie stock cubes and legos versus
escafe ands and coffee and Jakarta. You have to factor
in local holidays, what competitors are doing, inventory levels. It
requires a lot of local knowledge, right.
Speaker 1 (10:40):
It sounds like something needing human intuition, local feel. How
did software replace that?
Speaker 2 (10:45):
Well, the AI basically removes the need for human intuition
by just crunching vastly more data. It analyzes millions, literally
millions of data points, consumer behavior, supply chain costs, target
profit virgins. It runs constant ad tests on promotions at
a scale no human team could ever manage.
Speaker 1 (11:05):
So it doesn't have instinct.
Speaker 2 (11:06):
It has probability, exactly optimal probability. It figures out the
most profitable discount, depth, timing, and channel better and faster
than any human planner ever could, And just like that,
three hundred and twenty sparalized roles become obsolete.
Speaker 1 (11:20):
Okay, that's marketing planning. What was the second big application
you mentioned? It targeted revenue management?
Speaker 2 (11:25):
Yeah, that was the dynamic pricing engine. This software is
directly responsible for eliminating one hundred and eighty revenue management
analysts just in Zone Americas alone.
Speaker 1 (11:34):
Dynamic pricing so constantly adjusting prices continuously.
Speaker 2 (11:38):
This engine adjusts the shelf price or the recommended price
for products based on real time factors things like competitor
stock levels. It scrapes online, local digital ad spend, even
tiny fluctuations in raw material costs for coco or coffee.
It's all aimed at maximizing revenue capture minute by.
Speaker 1 (11:54):
Minute, that speed, that scale of adjustment, It's just impossible
for a human team, isn't it, No matter how smart
they are.
Speaker 2 (12:01):
Exactly a human analyst team might review pricing strategy quarterly,
maybe monthly. If they're really on it. This engine it
can review and adjust every five minutes if needed. The
efficiency gap is just too vast to bridge with people.
Speaker 1 (12:15):
Okay, so that's the front end the marketing and pricing.
Moving to the back office, you mentioned huge consolidation and
shared services. This is where big tech meets well corporate bureaucracy.
Speaker 2 (12:27):
That's a good way to put it, Nesley. Like many
multinationals had already centralized huge chunks of their financial and
admin processes into these big shared service centers, mainly in
places like Crackow, Poland and Quala Loompoor, Malaysia, lower cost locations.
The big technological leap here was combining their new SAPs
four ANA system that's their core enterprise resource planning software,
(12:49):
with robotic process automation or RPA. Specifically, the used uipathbots
quite heavily.
Speaker 1 (12:55):
Okay, RPA for our listeners, What are these uipathbots actually doing.
They're not physical robots, right.
Speaker 2 (13:00):
No, not physical robots walking around. RPA is software designed
to mimic routine human actions on a computer. Think of
them as digital employees. These UiPath bots log into different systems,
they read data from incoming invoices, They check that data
against purchase orders stored in the SAP system, They handle
basic exceptions, they route approvals, and they trigger payments, all
(13:23):
without a human touching a keyboard.
Speaker 1 (13:24):
And the result combining that centralized data in SAP with
these specialized.
Speaker 2 (13:30):
Bots, pure automated efficiency, the number they quoted was staggering.
Ninety two percent of all their invoices worldwide are now
processed end to end without.
Speaker 1 (13:40):
Any human touch, ninety two percent.
Speaker 2 (13:42):
Ninety two percent, And that huge efficiency gain allowed NESL
to consolidate or eliminate twenty eight hundred finance and accounting
roles globally.
Speaker 1 (13:50):
Wow, that's almost a quarter of the total white collar
cuts right there, just from that one tech combination, the
massive chump. And then we have to talk about the
most pervasive tool, the one that kind of infiltrated every
day generative AI copilots.
Speaker 2 (14:02):
Ah Yes, NESLEI GBT. This wasn't just off the shelf
chat GPT. They rolled out their own sophisticated, fine tuned
LAMA three model. And that fine tuned part is crucial.
Speaker 1 (14:14):
Why what does fine tuning LAMA three mean in this context?
Speaker 2 (14:18):
It means they took a powerful general purpose large language
model LAMA three developed by Meta, and they trained it
specifically on Nesle's own proprietary data. We're talking internal brand guidelines,
decades of legal documents and precedents, supply chain specifics, approved
corporate language, marketing campaign histories.
Speaker 1 (14:37):
So it learns to speak fluent Nesle exactly.
Speaker 2 (14:40):
It makes the AI far more effective and accurate for
internal corporate tasks than a generic model like chat GPT
would be. It knows the company inside.
Speaker 1 (14:48):
Out, and how widespread was its use, how many employees
were using this.
Speaker 2 (14:51):
Eighteen thousand employees were already using it daily as their
copilot eighteen thousand. Yeah, and the range of tasks it
could handle was astonishingly broad. Our sources mentioned things like
drafting customized sheld talkers those little signs you see on
supermarket shells for specific grocery chains, automatically generating first drafts
of complex legal contracts, and maybe the most eyebrow raising example,
(15:14):
oh on, even writing haiku poetry for Maggie Noodle advertising campaign.
Speaker 1 (15:18):
I cou seriously, seriously, if.
Speaker 2 (15:20):
A task involves synthesizing the information and generating text, images
or even code, it's vulnerable.
Speaker 1 (15:26):
And that sheer versatility of gena AI. It means it's
targeting jobs we might not have even thought of as
traditional automation targets before, which brings us to that highly
ironic internal memo that got leaked.
Speaker 2 (15:38):
Yes, the infamous memo. It was later verified. It listed
thirty eight specific job families that NESLE deemed at high
automation risk by the end of Q four twenty twenty six,
so very near term.
Speaker 1 (15:48):
And it included the obvious ones I assume.
Speaker 2 (15:50):
Oh yeah, demand planners, category analysts, master data stewards roles
heavily focused on analyzing, organizing, and managing information. Pretty predictable targets.
Speaker 1 (15:58):
But the really striking inclines collusion, the one that caused
a lot of internal chatter, was a role that was
supposed to be part of the solution to automation.
Speaker 2 (16:06):
Exactly also on that high risk list automation coordinators. You're kidding, Nope,
the irony is just profound. The human managers who were
initially brought in to oversee the deployment of these automation
projects to manage the transition their own roles are now
being automated by increasingly autonomous, self managing AI systems. The
(16:29):
solution is optimizing its own management out of existence.
Speaker 1 (16:32):
Okay, so the technology was clearly ready, maybe even overdue,
But the question is still why the sudden, really aggressive
move in twenty twenty five. Automation tools have been around
for years, improving steadily. The speed of these NESLAY cuts
suggests they hit some kind of critical point. Maybe were
survival or at least satisfying stakeholders demanded radical action.
Speaker 2 (16:52):
That's absolutely the right way to think about it. This
wasn't just about shiny new tech. It happened against a
backdrop of intense pressure hitting Nessley from well at least
three directions at once.
Speaker 1 (17:02):
Let's break those down. What was the first major pressure?
Speaker 2 (17:05):
First and foremost the financial imperative. This was driven hard
by activist shareholders. You know, despite Neslay being incredibly profitable already,
their operating margins were hovering around seventeen point two percent,
which is pretty healthy.
Speaker 1 (17:17):
Yeah, sounds good.
Speaker 2 (17:18):
Influential investors like Third point Artisan Partners, they weren't satisfied.
They wanted more.
Speaker 1 (17:23):
More than seventeen point two percent margins. What were they
pushing for?
Speaker 2 (17:26):
They were demanding fifteen percent margins by twenty twenty seven.
Now that sounds lower, but the seventeen point two percent
was the overall margin. Their demand for fifteen percent was
implicitly a push for massive efficiency gains and better asset utilization.
It was basically a demand for ruthless cost control, especially
in those SDNA categories.
Speaker 1 (17:44):
We talked about the white collar overheads exactly.
Speaker 2 (17:47):
They wanted to ensure that, even if markets fluctuated, Nesley's
profitability floor remained incredibly high, and that pressure landed squarely
on the CEO of Vertil. He had to find ways
to slash operational spend big ways.
Speaker 1 (18:01):
And this shareholder demand forced him to up the ante
on cost savings.
Speaker 2 (18:05):
Yes, he publicly increased the company's overall cost saving target.
It went from an already ambitious CHF two point five
billion to a truly staggering CHF three billion by twenty twenty.
Speaker 1 (18:18):
Seven, three billion Swiss francs.
Speaker 2 (18:20):
Yeah. And if you do the rough math you factor
in the sixteen thousand layoffs, the implied saving per laid
off employee comes out to around CHF one hundred and
eighty seven thy five hundred annually. Wow. So this wasn't
some philosophical choice about the future of work. It was
a cold, hard, financially necessary move to meet shareholder expectations.
Automation provided the means.
Speaker 1 (18:41):
Okay, so intense investor pressure. What was the second major
factor you mentioned? External shocks?
Speaker 2 (18:46):
Yeah, on top of that internal pressure came a massive
external shock, the commodity price tsunami. This wasn't just normal
inflation for a food company like an Esla. This was
bordering on.
Speaker 1 (18:54):
A crisis which commodities were hit hardest.
Speaker 2 (18:57):
Two main ingredients that are absolutely core to many their
biggest brands, so I historic price spikes Coco and a Rabica.
Speaker 1 (19:04):
Coffee the core of chocolate and coffee exactly.
Speaker 2 (19:06):
Coco futures prices were up an unbelievable one hundred and
eighty percent since twenty twenty three.
Speaker 1 (19:10):
One hundred and eighty percent.
Speaker 2 (19:11):
Yeah, and a Rabbitica coffee was up seventy two percent.
Speaker 1 (19:14):
So what does one hundred and eighty percent spike and
Coco actually mean for you know, the cost of making
a kit cat or a crunch bar.
Speaker 2 (19:21):
It means every single product containing chocolate or coffee suddenly
becomes dramatically more expensive to produce. It instantly eats into
your profit margins across Nesle's entire portfolio. The total increase
in these input costs hit CHF two point one billion
in twenty twenty four.
Speaker 1 (19:39):
Alone, two point one billion just from raw materials going on, just.
Speaker 2 (19:42):
From raw materials. Now, they couldn't just pass all of
that cost onto consumers through higher prices right now without
seeing huge drops in sales volume. People would just stop
buying or switch to cheaper alternatives. So if you can't
raise prices enough and your input costs are exploding, you
only have one other major lever.
Speaker 1 (19:57):
To pull cut your internal costs.
Speaker 2 (19:58):
Brutally control your into operational costs. Automation suddenly went from
being a nice to have efficiency tool too, basically the
only way to offset this commodity price tsunami and protect margins.
Speaker 1 (20:09):
Okay, investor pressure, commodity crisis. What was a third major
pressure point? Geopolitical headwinds?
Speaker 2 (20:15):
Yeah, the increasing instability and friction in global trade that
was hitting both sales and logistics hard. The China freeze out,
as some analysts called it, was particularly painful.
Speaker 1 (20:25):
What was happening there.
Speaker 2 (20:26):
Sales in their zone AOA, that's Asia, Oceania and Africa
were down eight point four percent year to date in
twenty twenty five. A big part of that was consumers
in emerging markets who maybe a few years ago were
trading up to premium Western brands like nes Cafe Gold,
were now rapidly switching back to cheaper, local, often digitally
savvy alternatives. The prime example was Luck and Coffee in China,
(20:49):
often selling at half the price. This kind of volume
contraction put even more strain on regional teams and profitability.
Speaker 1 (20:56):
And it wasn't just sales. The regulatory environment was getting
tougher twoffs exactly.
Speaker 2 (21:01):
We saw what our sources called the terrat tsunami. After
October first, twenty twenty five, the US slapped a huge
thirty nine percent duty on a range of Swiss imports,
not everything, but.
Speaker 1 (21:10):
Enough to hurt thirty nine percent.
Speaker 2 (21:12):
Ouch yeah, and then Mexico immediately retaliated with its own
twenty five percent tariff on imported powdered milk from certain regions.
These tariffs just add layers of complexity and cost to
cross border supply chains. They make the old often manually
managed logistics and distribution models much less economically viable. Almost overnight.
It just accelerated the need the mandate for streamlined automated
(21:35):
logistics solutions.
Speaker 1 (21:37):
So sixteen thousand people caught in the crosshairs of this
perfect storm really technology readiness, intense shareholder demands, exploding costs,
global instability. When we shift focus now to the human collateral,
the stories of the people displaced really illustrate the pain.
It's not just numbers on a page, it's the loss
of skills, experience, and deeply human contributions.
Speaker 2 (21:58):
Absolutely, let's maybe focus on one specific example. Maria Delgado.
She was forty one, a senior brand manager based in
Mexico City. Her story, I think, perfectly illustrates this paradox
of human creativity being recognized and then dismissed by the algorithm.
Speaker 1 (22:11):
What did she do? What was her success?
Speaker 2 (22:13):
She was the driving force behind a massively successful product launch,
the limited edition KitKat Turo Flavor.
Speaker 1 (22:19):
Oh I think I remember that.
Speaker 2 (22:20):
It was huge, especially in Latin America, sold something like
three million bars in the first seventy two hours. Yeah,
I mean that's textbook marketing success, right, Identifying a local trend,
creating a product, launching it brilliantly.
Speaker 1 (22:33):
Exactly a home run. So why on earth was her
role deemed non core in the restructuring Well, her.
Speaker 2 (22:40):
Success was localized, and ironically, the very processes she used
trend analysis, flavor concept and launch planning could now be
replicated and arguably exceeded in scale and speed by the
new AI tools, the promo Optimizer and the dynamic pricing
engine we talked about. They could run say, ten thousand
flavor simulations and regional AIB tests overnight, something Maria brilliant
(23:02):
tis she was just couldn't physically do so.
Speaker 1 (23:04):
Her complex synthesis of consumer trends logistics pricing.
Speaker 2 (23:09):
It was digitized, digitized, and scaled up massively. Her role
managing that process was deemed noncore because the function itself
had been largely automated, and she had a.
Speaker 1 (23:17):
Quote, didn't she about the nature of her replacement. It
really captures the essence of this white collar shift she did.
Speaker 2 (23:23):
It was widely reported she said, the algorithm that ab
tests flavors doesn't need maternity leave.
Speaker 1 (23:29):
Wow, that's blunt.
Speaker 2 (23:31):
It really encapsulates the cold, hard reality for many corporate
workers now, the human costs of employment, benefits, time off,
parental leave, sick days, the need for flexibility, even basic
things like empathy. These are increasingly seen as liabilities when
compared to an AI subscription fee. The algorithm works two
hundred and forty seven, demands no benefits, takes no leave,
(23:52):
carries zero hr risk.
Speaker 1 (23:54):
Okay, that's the white collar creative side. What about on
the factory floor. We mentioned Liam O'Connor as technician in
the York factory. His concerns were different, but maybe just
as fundamental.
Speaker 2 (24:04):
Yeah, Liam represents the displacement happening even among highly skilled
blue collar workers. His job wasn't necessarily eliminated immediately, but
the context changed entirely. His concern was about the new
automated systems rapidly filling the factory floor worried him. He
pointed out that the robots and automated lines quote don't unionize,
they don't develop propetitive strain injury, and maybe most poignantly,
(24:28):
they don't care that my daughter's school play is on Thursday.
His point highlights the loss of well flexibility and human
dignity in the workplace. Automation often means relentless optimized scheduling
that doesn't easily accommodate the messi realities of human life
outside the factory gates.
Speaker 1 (24:46):
Now, Nestle, like any big company facing this, they offered
the standard corporate response right severance packages help finding new jobs.
Speaker 2 (24:53):
Yes, the official line was generous. Severance reports suggested it
was typically twelve to eighteen month salary in Europe, but
interesting significantly last maybe three to six months in Latin
America and parts of Asia. They also promised investment in
outplacement services to help displaced workers transition.
Speaker 1 (25:08):
Right the standard package. But there was a deep, painful
irony in those outplacement services, wasn't there Almost satirical?
Speaker 2 (25:15):
Oh, the irony was thick enough to cut with a knife.
The very outplacement services offered the AI powered resume builders.
The automated career matching platforms designed to help people find
new jobs were themselves powered by the exact same kind
of generative AI technology that had just eliminated their old jobs.
Speaker 1 (25:34):
So the tech that fired you is now helping you
look for work.
Speaker 2 (25:37):
Pretty much. Critics rightly jumped on this, pointing out that
NESLI was essentially adding algorithmic insult to injury, using the
technology that caused the crisis to provide a largely ineffective
automated solution to it.
Speaker 1 (25:50):
Let's dig a bit deeper into this trend. We called
the white collar genocide earlier. That might be strong phrasing,
but it captures the surprise, why were these office rules.
These information workers hit harder in this way than the
traditional factory jobs. That seems counterintuitive to many people.
Speaker 2 (26:06):
It does seem counterintuitive, but it makes sense when you
look at what the current generation of AI, especially generitive AI,
is actually good at. Blue collar tasks often involve complex
physical dexterity, handling unpredictable materials, working in messy or unstructured environments,
things that are still quite challenging for general purpose robots.
Speaker 1 (26:26):
Right, robots are good at repetitive, predictable physical tasks exactly.
Speaker 2 (26:31):
But JENAI excels at tasks involving information, reading, summarizing, analyzing data,
generating reports, writing code, creating marketing copy, even making complex forecasts.
These are the core functions of middle management, finance departments,
marketing teams, analysts, the heartland of white color.
Speaker 1 (26:49):
Work, and the data backs this up. The risk analysis.
Speaker 2 (26:52):
Yeah, estimates from places like Mackenzie are pretty stark, and
we're circulating widely. Around this time, they suggested that something
like forty five percent of core finance and accout tasks
were already highly automatable with current tech. For marketing analytics,
the figure was around thirty eight percent, but the highest
vulnerability basic administrative and data entry roles. The invoice processing
data validation were estimated at sixty two percent automatable.
Speaker 1 (27:13):
Sixty two percent.
Speaker 2 (27:15):
These are the jobs most directly exposed to tools like those UiPath,
RPA bots and the internal and SLAGPT.
Speaker 1 (27:21):
We can see this very clearly in the fay of
one specific team you mentioned, the Revenue Growth Management team.
This wasn't data entry. This was a team full of
highly educated, highly paid experts.
Speaker 2 (27:32):
Absolutely, this team was the epitome of high value human expertise.
It was one staff by around eleven hundred people globally,
many with PhDs advanced degrees in economics statistics. Top tier
analysts focus purely on complex pricing and promotion strategies to
maximize revenue.
Speaker 1 (27:50):
And after the AI implementation.
Speaker 2 (27:51):
They were ruthlessly cut down, reduced to just four hundred and.
Speaker 1 (27:54):
Twelve humans from eleven hundred to four hundred and twelve. Wow.
Speaker 2 (27:57):
Yeah, And crucially, those remaining four hundred and twelve p
M are now primarily focused on overseeing the AI. They're
supplemented by a massive AWS sage Maker cluster. That's Amazon's
cloud AI platform which runs, according to internal documents, forty
million different pricing and promotion simulations every single night.
Speaker 1 (28:14):
Okay, let's unpack that for listeners. An AWS sage Maker
cluster running forty million simulations nightly? What is that actually
doing and why does it outperform nearly seven hundred PhDs?
Speaker 2 (28:23):
Right? Sage Maker is Amazon's cloud platform that makes it
relatively easy to build, train, and deploy machine learning models
at enormous scale. This cluster uses parallel processing, basically lots
of computers working together to simultaneously test forty million different
scenarios every night. Think different price points in different regions,
bundled offers, competitor reactions, impact on volume versus margin, endless variations. Okay,
(28:46):
Now compare that to a human PhD analyst. Maybe they
could meticulously analyze what twenty different scenarios in a whole week.
The machine analyzes forty million complex scenarios in just a
few hours while everyone's asleep. The human capacity for analysis
at that time scale and speed simply cannot compete. The
human role inevitably shifts from doing the analysis to defining
(29:06):
the parameters. For the AI and interpreting its output. It's
a fundamental change in the nature of the job.
Speaker 1 (29:12):
So the ness lay story, as shocking as it is,
it's clearly not an isolated incident. You called it the blueprint,
and twenty twenty five really does seem to be crystallizing
as the year automation eight corporate Earth. Let's zoom out
now and place this nest lay blueprint within that wider context.
We're seeing parallel displacements driven by AI across major global
(29:33):
industries right absolutely.
Speaker 2 (29:34):
NESLEI was maybe the most high profile example in the
consumer goods sector, but it's just one domino falling in
a very long line. That same core philosophy using advanced
AI and robotics to drive radical efficiency gains, primarily through
headcat reduction, became standard corporate strategy across the board in
twenty twenty five. We have whole lists from resources of
parallel mass layoffs explicitly tied to adopting this kind of
(29:57):
advanced automation.
Speaker 1 (29:58):
Okay, let's run through some of those key industry starting
with logistics, maybe a sector heavily reliant on moving physical things.
Speaker 2 (30:04):
Right, look at UPS. They announced cuts of forty eight
thousand package sorterers and warehouse handling staff.
Speaker 1 (30:09):
Globally forty eight thousand.
Speaker 2 (30:11):
Yeah, the direct replacement advanced robotics, heavily leaning on systems
similar to those used by Amazon Robotics, plus their own
proprietary automated sorting systems. The need for human eyes and
hands for basic sorting in those massive hubs just plummeted.
Speaker 1 (30:26):
And Amazon itself they weren't immune.
Speaker 2 (30:28):
Far from it. Amazon cut thirty thousand middle managers. Now,
these weren't primarily warehouse workers. These were supervisors, team leads,
program managers. Their roles were essentially made redundant by increasingly
sophisticated AI managing the systems. Think about the just walkout
technology in their retail stores, eliminated and checkout staff and supervisors,
(30:49):
or the AI optimizing supply chain flows which removed layers
of human oversight.
Speaker 1 (30:53):
Even deep tech. The companies making the chips and software,
they started cutting the people who build the product.
Speaker 2 (30:58):
Yes, Intel announced cuts of twenty four thousand fab technicians.
Speaker 1 (31:03):
People working in the chip factories.
Speaker 2 (31:05):
Exactly highly skilled technicians. This displacement was driven largely by
the accelerating integration of things like extreme ultraviolet or EUV
lithography machines. These are incredibly complex, expensive machines, but they're
also increasingly automated and self monitoring. They simply require far
fewer human technicians on the floor to maintain and operate
(31:27):
the semiconductor fabrication lines.
Speaker 1 (31:29):
And the cuts even hit the software engineers, the coders themselves,
the people writing the AI.
Speaker 2 (31:35):
That was maybe the most ironic twist. Microsoft, a company
at the absolute forefront of AI development, cut nine thousand coders.
Why well, largely because tools like their own GitHub Copilot,
which is basically generative AI specifically trained to write and
debug computer code, started automating significant chunks of the routine
boilerplate coding work.
Speaker 1 (31:53):
So the AI is writing its own code now to.
Speaker 2 (31:55):
A large extent, yes, or at least assisting human coders
so much the nique viewer of them. We saw a
similar trend in IT services and consulting. Big players like
Tata Consultant News Services TCS cut twelve thousand software testers.
Their jobs were directly replaced by new AI driven automated
testing suites that can run continuous quality assurance checks much
(32:16):
faster and more comprehensively than manual testing teams, and the
overall numbers tracked by industry analysts. They reflect this acceleration dramatically.
Challenger Gray and Christmas, who track layoff announcements, reported a
really startling data point for twenty twenty five, eighty seven
thousand job cuts in the US were explicitly attributed by
the companies themselves to AI displacement. Eighty seven thousand just
(32:39):
in the US just explicitly linked to AI. And that
number was triple the tally they recorded for AI linked
cuts in twenty twenty four. So the acceleration isn't just steady,
it's exponential. It really confirms twenty twenty five as the
tipping point, the year corporate adoption of these powerful AI
tools translated into mass workforce displacement.
Speaker 1 (32:57):
Okay, but if we circle back to the factory floor
firm moment, the blue collar jobs that do remain, they're
changing drastically too, aren't they. They're becoming more specialized, often
focused on managing the robots themselves. Ness Lee had already
seen big efficiency gains there even before the twenty twenty
five cuts.
Speaker 2 (33:14):
Oh, definitely, Automation on the factory floor wasn't new, but
the sophistication ramped up look at their Kaler chocolate plant
in Brock, Switzerland. It's a historic site. Back in twenty eighteen,
it operated with about four hundred and twenty human workers. Okay,
by twenty twenty five was running with only two hundred
and eighty humans, but alongside them were one hundred and
fourteen Abbu Mei dual arm cobots. These are collaborative robots
(33:38):
designed to work safely alongside people.
Speaker 1 (33:40):
Fewer people, more robots. What did that do to output
in safety?
Speaker 2 (33:43):
The results were pretty dramatic. Output from the plant increased
by twenty eight percent, and maybe even more importantly, workplace
accidents dropped by a massive seventy one percent. Robots don't
get tired or make mistakes in the same way humans do, and.
Speaker 1 (33:55):
Those remaining two hundred and eighty humans their jobs must
be completely different.
Speaker 2 (33:59):
Now hardly different. The source material notes quite starkly that
their primary job now is to calibrate the robots that
replace their friends. Their work is about maintenance, programming the robots,
troubleshooting issues, handling exceptions the automation can't deal with. It's
not direct production work anymore. It requires a higher level technical.
Speaker 1 (34:18):
Skill set, and we saw similar stories in other regions
like Asia.
Speaker 2 (34:22):
Yeah. Another clear example was in their factory in Batongas
in the Philippines. One specific task there was manually inspecting
pouches of nest quick powder for defects, tears, bad seals,
printing errors.
Speaker 1 (34:34):
Used to be done by I.
Speaker 2 (34:35):
Used to be done by a team of human inspectors
visually checking pouch after pouch. Now a single sophisticated AI
powered vision system does the inspection. It scans eight hundred
units per minute, far faster than any human could, and
it replaced a team of twelve human inspectors on that
line alone. The AI's accuracy is higher, its speed is incomparable.
(34:56):
It just makes the human visual inspection role economically redundant
in that context.
Speaker 1 (35:00):
So as the sheer scope of this AI driven displacement
becomes clearer, hitting both white collar and blue collar, across industry,
is across the globe. The fundamental question for governments, for
society really is how do we absorb this shock, this
potential for mass unemployment for widespread skills obsolescens. It's huge.
We've seen some policy ideas floated, haven't we, But the
(35:22):
results haven't been particularly encouraging so far.
Speaker 2 (35:25):
No the policy response has really struggled to keep pace
with the technology. The debate around universal basic income or UBI,
flared up again intensely around twenty twenty five. It was
positioned by proponents as a necessary safety net for potentially
millions displaced by automation.
Speaker 1 (35:39):
And there's been trials, right, pilots.
Speaker 2 (35:42):
Yeah, pilots have run in various forms in places like Finland, Canada,
parts of the US like Stockton, California, even Kenya, but
the results have been really mixed and often quite controversial.
UBI seems to provide some basic income stability, which is good,
but critics argue, often quite forcefully, that it doesn't do
enough to encourage the necessary reskilling or active re engagement
(36:04):
with a rapidly changing job market. It risks creating dependency
without opportunity.
Speaker 1 (36:09):
Okay, So UBIS debated what about the idea of taxing
the automation itself, the robot tax concept, basically treating a
productive robot or algorithm as if it were a taxable
employee to help fund social programs or retraining.
Speaker 2 (36:22):
That idea gained some traction too, particularly in Europe. France
actually tried to implement a version of it, but the
outcome really illustrated the fundamental difficulty of regulating, let alone taxing,
highly mobile technology and capital.
Speaker 1 (36:34):
What happened in France well, Almost immediately.
Speaker 2 (36:36):
After the policy was seriously proposed, Boston Dynamics, a major
player in advanced robotics though not directly involved in Nesle's case,
announced they were shifting a significant portion of their R
and D operations out of Europe and into Singapore. Why Singapore,
because Singapore offered a much more favorable, essentially tax free,
regulatory environment for AI and robotics development. The message was clear,
(37:01):
you can't easily tax an algorithm, a piece of software,
or the intellectual property behind a robot because it can
be relocated digitally across borders almost instantly. Attempting to tax
it heavily in one jurisdiction just risks driving the innovation,
investment and jobs elsewhere. The French effort kind of fizzled out,
and even.
Speaker 1 (37:19):
When companies like Nesli make public pledges to help their
own displaced workforce, the actual budget numbers often tell a
different story, don't they a story of priorities?
Speaker 2 (37:28):
They really do. Nesle publicly played chf fifty million towards
employee retraining and resclling programs in the wake of the layoffs,
which sounds like a decent chunk of money.
Speaker 1 (37:38):
Fifty millions was Franks.
Speaker 2 (37:39):
Yeah, but critics were incredibly quick to put that number
in perspective. They juxtaposted against Nesle's technology investments. They pointed
out that roughly the same amount around CHF fifty million,
was what Nesle had spent purchasing about one hundred and
eighty large industrial robots just the previous year. Ah so
the investment in the automated replacement fundamentally dwarfed the investment
(38:02):
in human recovery and retraining. The priorities seemed pretty clear
despite the pr pledges.
Speaker 1 (38:07):
Okay, so policy is struggling corporate retraining budgets or maybe
less than meets the eye. What about the court of
public opinion consumer behavior. We saw a huge, very vocal
backlash online against the ethics of these cuts.
Speaker 2 (38:20):
Oh yeah, the ethical outrage was immediate and loud, especially
on social media. Hashtag boycott Nessele trended globally. You had
TikTokers doing these dramatic videos of themselves smashing kit cat
bars and protest.
Speaker 1 (38:32):
Right performative maybe, but visible very visible.
Speaker 2 (38:35):
And there were some tangible impacts. Initially. Sales of Nesli
Pure Life bottled water, for example, reportedly dropped by about
four percent on US college campuses in the week immediately
following the announcement. Campuses are often hotspots for this kind
of ethical consumerism, so there was definitely a strong negative
reaction in the digital sphere and amongst certain demographics.
Speaker 1 (38:55):
Did that outrage translate into a lasting hit to the
bottom line? Did consumers actually vote with their wallets in
large numbers?
Speaker 2 (39:03):
That's the crucial disconnect, isn't it. Despite all the online fury,
the hard sales data told a different story. Nesle's overall
Q three results, the ones released right alongside the layoff news,
still showed global confectionery growth was up six point eight percent,
so strong growth, still strong growth. It seems that while
many people might protest the ethics online, when they're actually
(39:25):
standing in the supermarket aisle, factors like price, convenience, brand, loyalty,
or maybe just the simple craving for a familiar chocolate bar,
those tend to win out. The data suggests consumers continued
to quote vote with their wallets and apparently still crave
crunch bars. The business imperative to maintain profitability through automation
(39:45):
clearly outweigh the short term pr risk from the ethical backlash.
Speaker 1 (39:50):
So with the financial markets validating the move, policy responses lagging,
and consumer boycotts proving somewhat temporary, what's the path forward
for Nesley itself? What does see EO Navertill's longer term
vision look like now through to say twenty twenty eight.
Speaker 2 (40:05):
The internal roadmap, parts of which were leaked, shows a
plan that's even more aggressive, really pushing the boundaries of
autonomous corporate operations. For twenty twenty six, the goal was
the full global rollout of something called zero touch planning.
Speaker 1 (40:17):
Zero touch planning what's that?
Speaker 2 (40:19):
This is where AI takes over almost the entire demand
forecasting and supply chain management process. The I predicts demand,
manages inventory levels, automatically generates production orders, and even optimizes
logistics down to the level of stocking and individual storeeshelf,
all with minimal or ideally zero human intervention.
Speaker 1 (40:38):
Okay, so automating the flow of goods. What about the
creative side product development? Is AI taking that over to increasingly?
Speaker 2 (40:45):
Yes? By twenty twenty seven, the internal target was for
forty percent of their entire global portfolio of skewed stock
keeping units, basically individual products to be managed by generative design.
Speaker 1 (40:56):
Generative design, so the AI is inventing products essentially.
Speaker 2 (40:59):
Yes, it means I won't just be optimizing existing products
or promotions. It'll be actively generating entirely new flavor profiles,
suggesting novel packaging designs, even formulating new product concepts based
on its analysis of market trends, competitor moves, and predictive
modeling of consumer desires. AI as inventor.
Speaker 1 (41:16):
And the ultimate end goal, the pinnacle of this automation drive, the.
Speaker 2 (41:19):
Twenty twenty eight target, the holy grail for this kind
of efficiency push was the commissioning of the first fully
autonomous Nesliy factory, a completely lights out human free facility
planned for somewhere in Poland.
Speaker 1 (41:32):
According to the documents, lights out, human free.
Speaker 2 (41:35):
That's the end point of this journey, where humans are
entirely removed from the physical production cycle. There was even
an internal slogan reported to be circulating within the digital
transformation teams that sums up this whole strategic shift from
two thousand brands to two thousand algorithms WOW.
Speaker 1 (41:50):
From brands to algorithms. That really says it all.
Speaker 2 (41:53):
Okay, So wrapping this deep dive up, looking back at
Nesle's silicon layoffs, it's crystal clear this wasn't just an
ice decision by one food giant. It really serves as
a stark blueprint for global corporate strategy in the age
of AI. This model integrating massive scale, generative AI with
specialized automation tools to fundamentally streamline or eliminate core white
(42:15):
collar functions like finance, marketing, supply chain planning, even aspects
of R and D. This is the playbook. It's the
playbook every major conglomerate is now studying, if not actively implementing.
Absolutely What really stands out from the Nestley case is
the sheer, stunning speed at which traditional, often highly valued
corporate expertise seem to just erode when confronted with targeted,
(42:36):
scalable automation, the pressure cooker environment, the need to hit
aggressive shareholder demands, compounded by those external shocks like commodity prices.
It basically compressed maybe a decade's worth of gradual technological
change into well forty eight hours of crisis driven decision making.
Speaker 1 (42:53):
We started with the shock of the announcement, we went
through the specific tech stack they use, the automation arsenal,
and we saw the painful human impact and the societal
friction it created. If there's one crucial takeaway from all
this for you, the listener, it has to be about
the immediate, urgent need to reassess essential career skills. The
ground has fundamentally shifted beneath our feet.
Speaker 2 (43:13):
Definitely think about that letter to the class of twenty
twenty six. It was apparently circulated widely, partly inspired by
the NESLE events. The core message was blunt. If you're
graduating now or soon with traditional degrees, even prestigious ones
like MBAs or specialized masters in things like supply chain management,
(43:34):
those qualifications are potentially becoming obsolete pert least significantly devalued
by the very AI systems they might teach you about.
The education system is almost inevitably training people for the
jobs of the last decade, not the next one.
Speaker 1 (43:47):
Which raises the unavoidable and frankly quite daunting question. If
the analysis, the administration, the planning, even some of the
creative work is being automated, what skills remain essential? What
skills are defeat sensible. In this new landscape.
Speaker 2 (44:01):
The skills that seem most valuable now are highly specific
and often quite technically focused. You almost certainly need to
become adept at prompt engineering, which sounds simple, but it
is actually the complex art of communicating effectively with these
powerful AI engines to get the optimal desired output. It's
about knowing how to ask the right questions in the
right way.
Speaker 1 (44:19):
Okay, prompt engineering, what else you need?
Speaker 2 (44:21):
People who understand robot maintenance, not just the physical upkeep
of machines on a factory floor, but also the software maintenance,
the continuous monitoring and fine tuning of the algorithms themselves,
keeping the automated systems running smoothly.
Speaker 1 (44:36):
But maybe the most important skill, the only truly secure
role long term, is perhaps broader than just the technology itself.
Speaker 2 (44:43):
I think that's right, And maybe this is the most
provocative but also hopeful thought to end on the conclusion
many are reaching. Is this the only jobs AI can't automate?
At least not yet and maybe not ever. Are the
ones involved in inventing and defining the rules that govern
them AI itself.
Speaker 1 (45:00):
How we're talking about.
Speaker 2 (45:01):
Policy development, ethical framework design, crafting regulations, defining standards for
human machine collaboration, shaping the strategy for how AI integrates
into society and the economy. These metal level roles are
potentially safe because they involve setting the boundaries, defining the goals,
and managing the societal impact rather than just executing tasks
within the system the AI now runs.
Speaker 1 (45:22):
Nesley certainly showed the world what radical AI driven disruption
looks like in practice. Sixteen thousand families suddenly rewriting their
resumes while servers hum quietly in data centers in Switzerland
and elsewhere, delivering unprecedented efficiency gains directly to shareholders. The future,
that sci fi future we used to talk about, it's
(45:44):
not distant anymore. Yeah, it's here, it arrives, it's here,
and it's demanding new skills, new thinking, and maybe entirely
new rules starting right now.
Speaker 2 (45:52):
It really is a profound wake up call for everyone
in the corporate world and beyond.
Speaker 1 (45:56):
Thanks for diving deep with us on this critical story.
We'll see next time on the Deep Dive
Welcome to the deep dive. Today. We're opening the file
on well, a really pivotal corporate decision. It wasn't just
about efficiency. It felt more like a watershed moment, didn't
It really defining this battle between human labor and artificial intelligence.
Speaker 2 (00:16):
Absolutely, and the setting is key here. Yeah, we're not
talking Silicon Valley labs. No, we're talking Vevet, Switzerland. Yeah,
you know Quiet Scenic Nesle's global.
Speaker 1 (00:27):
Headquarters, right, the home of chocolate and coffee.
Speaker 2 (00:29):
Basically exactly. And it was October sixteenth, twenty twenty five. Nisle,
this giant stable company operating in what one hundred and
eighty eight countries, They dropped this announcement and it just
sent tremors really across every corporate sector on Earth. This
wasn't you know, a new coffee flavor or some temporary merger,
you know, this was fundamentally it's a fundamental, aggressive structural overhaul. Yeah.
(00:50):
And the driver purely the promise of well silicon efficiency.
Speaker 1 (00:54):
Let's jump straight to that central number, the one that
really made everyone set up. We are talking about sixteen thousand.
Speaker 2 (00:59):
Jobs thousand, yeah, cut over two years.
Speaker 1 (01:02):
And when you think about their total workforce, about two
hundred and seventy thousand people globally. That's nearly six percent.
Speaker 2 (01:09):
It's huge.
Speaker 1 (01:09):
That scale alone is massive, But the breakdown that's what
really hit hard, I think for investors and workers.
Speaker 2 (01:16):
Precisely because for decades, right when you thought mass layoffs,
you thought factory floor, Yeah, blue collar, blue collar. But
here twelve thousand of those rules were white collar finance specialists,
marketing directors, HR people, data analysts, the office job, the
office jobs, the traditional blue collar posts, the factory floors,
the warehouses. That was the other four thousand. So it
(01:39):
completely flipped the historical script.
Speaker 1 (01:40):
It really did, and it makes that statement from the CEO,
Philip Navtil sound particularly well chilling in retrospect.
Speaker 2 (01:48):
What did he say exactly?
Speaker 1 (01:49):
He said, the world is changing and Nesley needs to
change faster. And the key instrument he pointed to for that.
Speaker 2 (01:54):
Speed, let me guess AI.
Speaker 1 (01:56):
AI driven automation. Yeah, from you know, sophisticated predictive dashboards
all the way to these incredibly versatile generative tools. This
wasn't just cost cutting like we saw in the twenty tens.
This felt new.
Speaker 2 (02:07):
You're absolutely right, this isn't about chasing cheaper labor. Offshore anymore.
This is about eliminating the need for well human thinking
in that middle layer of the corporate structure, the operational hierarchy.
Speaker 1 (02:20):
So our mission today for this deep dive, it's not
just about Nesle's restructuring, is it.
Speaker 2 (02:26):
It's bigger, much bigger.
Speaker 1 (02:28):
Yeah.
Speaker 2 (02:28):
We're framing this, based on our sources, as the opening
salvo in a global war between carbon based workers and
silicon efficiency. That's the phrase they use.
Speaker 1 (02:38):
O wow, Okay, So we're going to unpack the specific
tech that Nesle had been kind of quietly weaponizing for
years ago. Yeah, and the financial pressures of the geopolitical
stuff that kind of forced their hand. And crucially, what
this blueprint means for you, the listener, whether you're in tech, finance, logistics,
because this playbook it's going to be copied, isn't it.
Speaker 2 (02:58):
Oh? Absolutely, it's already happen. And if you need proof
that the corporate world validated this human cost, just look at.
Speaker 1 (03:05):
The market reaction instant validation.
Speaker 2 (03:07):
Nesle stock jumped nine point three percent in a single day.
The efficiency janes, not the social impact, were immediately priced
in by global finance. That tells you everything.
Speaker 1 (03:17):
Okay, So let's track the actual chaos as it unfolded.
This deep dive really has to start with that forty
eight hour timeline, the corporate shockwave, because this wasn't some smooth,
perfectly planned announcement, was it. It started with a leak,
a massive leak, right.
Speaker 2 (03:31):
It wasn't clean at all. The initial phase was just well,
pure uncontrolled internal combustion.
Speaker 1 (03:36):
Really yeah.
Speaker 2 (03:37):
It kicked off late on Wednesday, October fifteenth, specifically eleven
point four to two pm Central European summertime, late at night,
very late, and these internal planning slides detailing something called
Project Accelerate Headcount Optimization catch you name, very corporate speak,
they somehow got leaked onto global internal slack channels.
Speaker 1 (03:57):
Global, so not just one office.
Speaker 2 (03:59):
No, no, across multiple time zones almost instantly. We're talking Salpolo, Singapore, London, everywhere.
And the key thing here is this information wasn't meant
for broad consumption yet, not at all.
Speaker 1 (04:11):
So it just flooded the internal comms exactly.
Speaker 2 (04:15):
It flooded these channels, creating immediate unofficial panic before management
could even get ahead of the story. Our sources say
that within minutes those slack channels were basically unusable, just
a mess of reactions, fear questions.
Speaker 1 (04:29):
Okay, so internal chaos first. Then came the attempt to
control the narrative, the public facing response on Thursday morning.
Speaker 2 (04:36):
Right the CEO Navartil. He moved fast but unconventionally, at
seven am, before any official pres elease hit the wires.
Speaker 1 (04:45):
What did he do?
Speaker 2 (04:46):
He posted directly to LinkedIn.
Speaker 1 (04:47):
LinkedIn okay, yeah.
Speaker 2 (04:49):
Talking about quote simplifying our organization and automating our processes.
It felt like a preemptive strike, positioning the cuts as
you know, futuristic optimization rather than just painful job losses.
Speaker 1 (05:00):
And the reaction on LinkedIn at that early.
Speaker 2 (05:03):
Hour it was huge, forty eight thousand reactions before most
people in Europe had even had their coffee.
Speaker 1 (05:08):
Wow.
Speaker 2 (05:09):
It really signals a shift in how corporations communicate, doesn't
it Using a professional social platform to immediately frame something
really controversial as innovation, speaking directly to the professional class
the investors first.
Speaker 1 (05:21):
But the actual hard numbers, that sixteen thousand figure that
was almost deliberately buried in the official releases later that morning,
is that fair? Well?
Speaker 2 (05:30):
Completely fair? At eight fifteen am, the investor relations team
put out the official Q three numbers, and the headline.
Speaker 1 (05:35):
Was good, right they beat expectations?
Speaker 2 (05:37):
Yeah, organic growth was strong plus four point three percent.
It beat the market consensus estimates by one hundred and
ten basis points.
Speaker 1 (05:43):
Okay, hold on one hundred and ten basis points for listeners,
maybe less familiar with market jargon. What does that really mean?
How good is that?
Speaker 2 (05:51):
It means they significantly outperformed what analysts expected. A basis
point is tiny one hundred of a percent, So beating
consensus by one point one percentage points on growth, that's
a big deal.
Speaker 1 (06:03):
And that good news provided cover.
Speaker 2 (06:05):
Perfect camouflage the sixteen thousand job cuts, the full time
equivalent reduction. It was, as we noted buried page nine
of the investor deck, a massive strategic shift hidden behind
really quite stellar quarterly performance.
Speaker 1 (06:17):
But the financial world they read the fine print, don't they.
Speaker 2 (06:20):
They saw it, oh instantly. By nine thirty am, barely
an hour after that Q three release, Zurich Kintonalbank upgraded
Nisley's stock to.
Speaker 1 (06:28):
Buy, and they gave a reason, a.
Speaker 2 (06:30):
Very specific reason. Their analysis targeted CHF one to eighteen
per share, and they explicitly cited quote margin upside from
AI led SGNA leverage is underappreciated.
Speaker 1 (06:42):
Okay, let's slow down on that phrase, because that really
is the financial translation of this whole story. AI led
SGNA leverage. What's SGNA leverage?
Speaker 2 (06:50):
Right? Sg and A is selling general and administrative expenses,
basically the white collar.
Speaker 1 (06:55):
Stuff, marketing it finance.
Speaker 2 (06:58):
Exactly all those corporate overheads. The twelve thousand rolls being
cut falls squarely into that bucket. Leverage in this context
means increasing your output, so more sales, better pricing, while
drastically reducing the costs required to support that output.
Speaker 1 (07:12):
So the bank was saying.
Speaker 2 (07:13):
The bank was saying, we see a huge, currently undervalued
opportunity for nesle to make more money using far fewer
human staff in these support roles. They saw the layoffs
not as a crisis but as pure immediate margin expansion potential,
cold hard numbers.
Speaker 1 (07:27):
So while the stock race is soaring in Zurich, the
actual physical impact is starting to land thousands of miles away.
Speaker 2 (07:33):
Yeah, the reality check hit hard in the traditional industrial
places that very same afternoon, by two point pm, Unite
the Union in York in the UK.
Speaker 1 (07:45):
The Kitcat factory, that's.
Speaker 2 (07:46):
One they had to schedule an emergency meeting. Why because
they were anticipating one hundred and eighty operator jobs disappearing
just from the Kitcat production lines.
Speaker 1 (07:55):
So it wasn't just the offices, it was hitting the
factory floor simultaneously.
Speaker 2 (07:59):
Absolutely, automation push was across the board, from the Corbett
HQ right down to the production lines.
Speaker 1 (08:05):
And then by Friday, this whole thing explodes into the
global public conversation, not really driven by the company anymore,
but by the people affected.
Speaker 2 (08:14):
It went completely viral. CNN had the headline the hashtag
hashtag Nestle layoffs was trending everywhere, over one point two
million posts. But the moment, that piece of content that
really defined the human side of this crisis. It came
from Manila a TikTok video, right yeah, a call center
agent's TikTok video.
Speaker 1 (08:31):
And it got massive views, for.
Speaker 2 (08:33):
Ten million views in less than twenty four hours. It
was incredible.
Speaker 1 (08:36):
What was the message? What made it resonate so much?
Speaker 2 (08:39):
It was heartbreakingly specific and just so relatable to anyone
in that kind of middle layer administrative job. Her role
was scheduling the Espresso promotional campaigns, you know, a routine,
rules based task, and her lament was simple. She basically said,
my job scheduling espresso promos. Now chat GPT does it
(08:59):
in an eleven seconds. That video just perfectly captured the
terrifying speed at which these very specific, one seemingly secure
jobs were suddenly becoming redundant everywhere.
Speaker 1 (09:10):
That TikTok story is so critical because it underlines that
this isn't some future threat. That tech is here, it's operational,
and it's actively taking jobs right now. So let's look
inside this automation arsenal. You mentioned Nesla was quietly weaponizing
AI for five years. This wasn't an overnight thing. Where
was this work happening.
Speaker 2 (09:26):
Well, the central hub, the engine room, if you like,
is the Nesle Digital Factory, the NDF. It's based in
los Anne, Switzerland, and it's a serious operation. Employs about
twelve hundred data scientists.
Speaker 1 (09:37):
Do one hundred. Wow.
Speaker 2 (09:38):
Yeah. And what makes their impact so well destructive to
human roles right now is the sheer speed of deployment.
Get this. They developed forty two new generative AI use
cases in twenty twenty.
Speaker 1 (09:52):
Four alone, forty two in one year.
Speaker 2 (09:54):
Yep. And these aren't just minor process tweaks. We're talking
and to in replacements for complex humans processes.
Speaker 1 (10:01):
And these are the tools directly responsible for many of
those twelve thousand white collar cuts we talked about. Can
you break down some specifics, like how did they target
core functions like marketing?
Speaker 2 (10:10):
Okay, so, one of the first major killer apps, as
our sources called it, was the promo Optimizer. This single
piece of software directly replaced three hundred and twenty trade
marketing planners globally.
Speaker 1 (10:22):
Trade marketing planners. That that sounds complex. What do they do?
Speaker 2 (10:25):
It's incredibly complex. It's deciding precisely when, where, and how
much to discount, say Maggie stock cubes and legos versus
escafe ands and coffee and Jakarta. You have to factor
in local holidays, what competitors are doing, inventory levels. It
requires a lot of local knowledge, right.
Speaker 1 (10:40):
It sounds like something needing human intuition, local feel. How
did software replace that?
Speaker 2 (10:45):
Well, the AI basically removes the need for human intuition
by just crunching vastly more data. It analyzes millions, literally
millions of data points, consumer behavior, supply chain costs, target
profit virgins. It runs constant ad tests on promotions at
a scale no human team could ever manage.
Speaker 1 (11:05):
So it doesn't have instinct.
Speaker 2 (11:06):
It has probability, exactly optimal probability. It figures out the
most profitable discount, depth, timing, and channel better and faster
than any human planner ever could, And just like that,
three hundred and twenty sparalized roles become obsolete.
Speaker 1 (11:20):
Okay, that's marketing planning. What was the second big application
you mentioned? It targeted revenue management?
Speaker 2 (11:25):
Yeah, that was the dynamic pricing engine. This software is
directly responsible for eliminating one hundred and eighty revenue management
analysts just in Zone Americas alone.
Speaker 1 (11:34):
Dynamic pricing so constantly adjusting prices continuously.
Speaker 2 (11:38):
This engine adjusts the shelf price or the recommended price
for products based on real time factors things like competitor
stock levels. It scrapes online, local digital ad spend, even
tiny fluctuations in raw material costs for coco or coffee.
It's all aimed at maximizing revenue capture minute by.
Speaker 1 (11:54):
Minute, that speed, that scale of adjustment, It's just impossible
for a human team, isn't it, No matter how smart
they are.
Speaker 2 (12:01):
Exactly a human analyst team might review pricing strategy quarterly,
maybe monthly. If they're really on it. This engine it
can review and adjust every five minutes if needed. The
efficiency gap is just too vast to bridge with people.
Speaker 1 (12:15):
Okay, so that's the front end the marketing and pricing.
Moving to the back office, you mentioned huge consolidation and
shared services. This is where big tech meets well corporate bureaucracy.
Speaker 2 (12:27):
That's a good way to put it, Nesley. Like many
multinationals had already centralized huge chunks of their financial and
admin processes into these big shared service centers, mainly in
places like Crackow, Poland and Quala Loompoor, Malaysia, lower cost locations.
The big technological leap here was combining their new SAPs
four ANA system that's their core enterprise resource planning software,
(12:49):
with robotic process automation or RPA. Specifically, the used uipathbots
quite heavily.
Speaker 1 (12:55):
Okay, RPA for our listeners, What are these uipathbots actually doing.
They're not physical robots, right.
Speaker 2 (13:00):
No, not physical robots walking around. RPA is software designed
to mimic routine human actions on a computer. Think of
them as digital employees. These UiPath bots log into different systems,
they read data from incoming invoices, They check that data
against purchase orders stored in the SAP system, They handle
basic exceptions, they route approvals, and they trigger payments, all
(13:23):
without a human touching a keyboard.
Speaker 1 (13:24):
And the result combining that centralized data in SAP with
these specialized.
Speaker 2 (13:30):
Bots, pure automated efficiency, the number they quoted was staggering.
Ninety two percent of all their invoices worldwide are now
processed end to end without.
Speaker 1 (13:40):
Any human touch, ninety two percent.
Speaker 2 (13:42):
Ninety two percent, And that huge efficiency gain allowed NESL
to consolidate or eliminate twenty eight hundred finance and accounting
roles globally.
Speaker 1 (13:50):
Wow, that's almost a quarter of the total white collar
cuts right there, just from that one tech combination, the
massive chump. And then we have to talk about the
most pervasive tool, the one that kind of infiltrated every
day generative AI copilots.
Speaker 2 (14:02):
Ah Yes, NESLEI GBT. This wasn't just off the shelf
chat GPT. They rolled out their own sophisticated, fine tuned
LAMA three model. And that fine tuned part is crucial.
Speaker 1 (14:14):
Why what does fine tuning LAMA three mean in this context?
Speaker 2 (14:18):
It means they took a powerful general purpose large language
model LAMA three developed by Meta, and they trained it
specifically on Nesle's own proprietary data. We're talking internal brand guidelines,
decades of legal documents and precedents, supply chain specifics, approved
corporate language, marketing campaign histories.
Speaker 1 (14:37):
So it learns to speak fluent Nesle exactly.
Speaker 2 (14:40):
It makes the AI far more effective and accurate for
internal corporate tasks than a generic model like chat GPT
would be. It knows the company inside.
Speaker 1 (14:48):
Out, and how widespread was its use, how many employees
were using this.
Speaker 2 (14:51):
Eighteen thousand employees were already using it daily as their
copilot eighteen thousand. Yeah, and the range of tasks it
could handle was astonishingly broad. Our sources mentioned things like
drafting customized sheld talkers those little signs you see on
supermarket shells for specific grocery chains, automatically generating first drafts
of complex legal contracts, and maybe the most eyebrow raising example,
(15:14):
oh on, even writing haiku poetry for Maggie Noodle advertising campaign.
Speaker 1 (15:18):
I cou seriously, seriously, if.
Speaker 2 (15:20):
A task involves synthesizing the information and generating text, images
or even code, it's vulnerable.
Speaker 1 (15:26):
And that sheer versatility of gena AI. It means it's
targeting jobs we might not have even thought of as
traditional automation targets before, which brings us to that highly
ironic internal memo that got leaked.
Speaker 2 (15:38):
Yes, the infamous memo. It was later verified. It listed
thirty eight specific job families that NESLE deemed at high
automation risk by the end of Q four twenty twenty six,
so very near term.
Speaker 1 (15:48):
And it included the obvious ones I assume.
Speaker 2 (15:50):
Oh yeah, demand planners, category analysts, master data stewards roles
heavily focused on analyzing, organizing, and managing information. Pretty predictable targets.
Speaker 1 (15:58):
But the really striking inclines collusion, the one that caused
a lot of internal chatter, was a role that was
supposed to be part of the solution to automation.
Speaker 2 (16:06):
Exactly also on that high risk list automation coordinators. You're kidding, Nope,
the irony is just profound. The human managers who were
initially brought in to oversee the deployment of these automation
projects to manage the transition their own roles are now
being automated by increasingly autonomous, self managing AI systems. The
(16:29):
solution is optimizing its own management out of existence.
Speaker 1 (16:32):
Okay, so the technology was clearly ready, maybe even overdue,
But the question is still why the sudden, really aggressive
move in twenty twenty five. Automation tools have been around
for years, improving steadily. The speed of these NESLAY cuts
suggests they hit some kind of critical point. Maybe were
survival or at least satisfying stakeholders demanded radical action.
Speaker 2 (16:52):
That's absolutely the right way to think about it. This
wasn't just about shiny new tech. It happened against a
backdrop of intense pressure hitting Nessley from well at least
three directions at once.
Speaker 1 (17:02):
Let's break those down. What was the first major pressure?
Speaker 2 (17:05):
First and foremost the financial imperative. This was driven hard
by activist shareholders. You know, despite Neslay being incredibly profitable already,
their operating margins were hovering around seventeen point two percent,
which is pretty healthy.
Speaker 1 (17:17):
Yeah, sounds good.
Speaker 2 (17:18):
Influential investors like Third point Artisan Partners, they weren't satisfied.
They wanted more.
Speaker 1 (17:23):
More than seventeen point two percent margins. What were they
pushing for?
Speaker 2 (17:26):
They were demanding fifteen percent margins by twenty twenty seven.
Now that sounds lower, but the seventeen point two percent
was the overall margin. Their demand for fifteen percent was
implicitly a push for massive efficiency gains and better asset utilization.
It was basically a demand for ruthless cost control, especially
in those SDNA categories.
Speaker 1 (17:44):
We talked about the white collar overheads exactly.
Speaker 2 (17:47):
They wanted to ensure that, even if markets fluctuated, Nesley's
profitability floor remained incredibly high, and that pressure landed squarely
on the CEO of Vertil. He had to find ways
to slash operational spend big ways.
Speaker 1 (18:01):
And this shareholder demand forced him to up the ante
on cost savings.
Speaker 2 (18:05):
Yes, he publicly increased the company's overall cost saving target.
It went from an already ambitious CHF two point five
billion to a truly staggering CHF three billion by twenty twenty.
Speaker 1 (18:18):
Seven, three billion Swiss francs.
Speaker 2 (18:20):
Yeah. And if you do the rough math you factor
in the sixteen thousand layoffs, the implied saving per laid
off employee comes out to around CHF one hundred and
eighty seven thy five hundred annually. Wow. So this wasn't
some philosophical choice about the future of work. It was
a cold, hard, financially necessary move to meet shareholder expectations.
Automation provided the means.
Speaker 1 (18:41):
Okay, so intense investor pressure. What was the second major
factor you mentioned? External shocks?
Speaker 2 (18:46):
Yeah, on top of that internal pressure came a massive
external shock, the commodity price tsunami. This wasn't just normal
inflation for a food company like an Esla. This was
bordering on.
Speaker 1 (18:54):
A crisis which commodities were hit hardest.
Speaker 2 (18:57):
Two main ingredients that are absolutely core to many their
biggest brands, so I historic price spikes Coco and a Rabica.
Speaker 1 (19:04):
Coffee the core of chocolate and coffee exactly.
Speaker 2 (19:06):
Coco futures prices were up an unbelievable one hundred and
eighty percent since twenty twenty three.
Speaker 1 (19:10):
One hundred and eighty percent.
Speaker 2 (19:11):
Yeah, and a Rabbitica coffee was up seventy two percent.
Speaker 1 (19:14):
So what does one hundred and eighty percent spike and
Coco actually mean for you know, the cost of making
a kit cat or a crunch bar.
Speaker 2 (19:21):
It means every single product containing chocolate or coffee suddenly
becomes dramatically more expensive to produce. It instantly eats into
your profit margins across Nesle's entire portfolio. The total increase
in these input costs hit CHF two point one billion
in twenty twenty four.
Speaker 1 (19:39):
Alone, two point one billion just from raw materials going on, just.
Speaker 2 (19:42):
From raw materials. Now, they couldn't just pass all of
that cost onto consumers through higher prices right now without
seeing huge drops in sales volume. People would just stop
buying or switch to cheaper alternatives. So if you can't
raise prices enough and your input costs are exploding, you
only have one other major lever.
Speaker 1 (19:57):
To pull cut your internal costs.
Speaker 2 (19:58):
Brutally control your into operational costs. Automation suddenly went from
being a nice to have efficiency tool too, basically the
only way to offset this commodity price tsunami and protect margins.
Speaker 1 (20:09):
Okay, investor pressure, commodity crisis. What was a third major
pressure point? Geopolitical headwinds?
Speaker 2 (20:15):
Yeah, the increasing instability and friction in global trade that
was hitting both sales and logistics hard. The China freeze out,
as some analysts called it, was particularly painful.
Speaker 1 (20:25):
What was happening there.
Speaker 2 (20:26):
Sales in their zone AOA, that's Asia, Oceania and Africa
were down eight point four percent year to date in
twenty twenty five. A big part of that was consumers
in emerging markets who maybe a few years ago were
trading up to premium Western brands like nes Cafe Gold,
were now rapidly switching back to cheaper, local, often digitally
savvy alternatives. The prime example was Luck and Coffee in China,
(20:49):
often selling at half the price. This kind of volume
contraction put even more strain on regional teams and profitability.
Speaker 1 (20:56):
And it wasn't just sales. The regulatory environment was getting
tougher twoffs exactly.
Speaker 2 (21:01):
We saw what our sources called the terrat tsunami. After
October first, twenty twenty five, the US slapped a huge
thirty nine percent duty on a range of Swiss imports,
not everything, but.
Speaker 1 (21:10):
Enough to hurt thirty nine percent.
Speaker 2 (21:12):
Ouch yeah, and then Mexico immediately retaliated with its own
twenty five percent tariff on imported powdered milk from certain regions.
These tariffs just add layers of complexity and cost to
cross border supply chains. They make the old often manually
managed logistics and distribution models much less economically viable. Almost overnight.
It just accelerated the need the mandate for streamlined automated
(21:35):
logistics solutions.
Speaker 1 (21:37):
So sixteen thousand people caught in the crosshairs of this
perfect storm really technology readiness, intense shareholder demands, exploding costs,
global instability. When we shift focus now to the human collateral,
the stories of the people displaced really illustrate the pain.
It's not just numbers on a page, it's the loss
of skills, experience, and deeply human contributions.
Speaker 2 (21:58):
Absolutely, let's maybe focus on one specific example. Maria Delgado.
She was forty one, a senior brand manager based in
Mexico City. Her story, I think, perfectly illustrates this paradox
of human creativity being recognized and then dismissed by the algorithm.
Speaker 1 (22:11):
What did she do? What was her success?
Speaker 2 (22:13):
She was the driving force behind a massively successful product launch,
the limited edition KitKat Turo Flavor.
Speaker 1 (22:19):
Oh I think I remember that.
Speaker 2 (22:20):
It was huge, especially in Latin America, sold something like
three million bars in the first seventy two hours. Yeah,
I mean that's textbook marketing success, right, Identifying a local trend,
creating a product, launching it brilliantly.
Speaker 1 (22:33):
Exactly a home run. So why on earth was her
role deemed non core in the restructuring Well, her.
Speaker 2 (22:40):
Success was localized, and ironically, the very processes she used
trend analysis, flavor concept and launch planning could now be
replicated and arguably exceeded in scale and speed by the
new AI tools, the promo Optimizer and the dynamic pricing
engine we talked about. They could run say, ten thousand
flavor simulations and regional AIB tests overnight, something Maria brilliant
(23:02):
tis she was just couldn't physically do so.
Speaker 1 (23:04):
Her complex synthesis of consumer trends logistics pricing.
Speaker 2 (23:09):
It was digitized, digitized, and scaled up massively. Her role
managing that process was deemed noncore because the function itself
had been largely automated, and she had a.
Speaker 1 (23:17):
Quote, didn't she about the nature of her replacement. It
really captures the essence of this white collar shift she did.
Speaker 2 (23:23):
It was widely reported she said, the algorithm that ab
tests flavors doesn't need maternity leave.
Speaker 1 (23:29):
Wow, that's blunt.
Speaker 2 (23:31):
It really encapsulates the cold, hard reality for many corporate
workers now, the human costs of employment, benefits, time off,
parental leave, sick days, the need for flexibility, even basic
things like empathy. These are increasingly seen as liabilities when
compared to an AI subscription fee. The algorithm works two
hundred and forty seven, demands no benefits, takes no leave,
(23:52):
carries zero hr risk.
Speaker 1 (23:54):
Okay, that's the white collar creative side. What about on
the factory floor. We mentioned Liam O'Connor as technician in
the York factory. His concerns were different, but maybe just
as fundamental.
Speaker 2 (24:04):
Yeah, Liam represents the displacement happening even among highly skilled
blue collar workers. His job wasn't necessarily eliminated immediately, but
the context changed entirely. His concern was about the new
automated systems rapidly filling the factory floor worried him. He
pointed out that the robots and automated lines quote don't unionize,
they don't develop propetitive strain injury, and maybe most poignantly,
(24:28):
they don't care that my daughter's school play is on Thursday.
His point highlights the loss of well flexibility and human
dignity in the workplace. Automation often means relentless optimized scheduling
that doesn't easily accommodate the messi realities of human life
outside the factory gates.
Speaker 1 (24:46):
Now, Nestle, like any big company facing this, they offered
the standard corporate response right severance packages help finding new jobs.
Speaker 2 (24:53):
Yes, the official line was generous. Severance reports suggested it
was typically twelve to eighteen month salary in Europe, but
interesting significantly last maybe three to six months in Latin
America and parts of Asia. They also promised investment in
outplacement services to help displaced workers transition.
Speaker 1 (25:08):
Right the standard package. But there was a deep, painful
irony in those outplacement services, wasn't there Almost satirical?
Speaker 2 (25:15):
Oh, the irony was thick enough to cut with a knife.
The very outplacement services offered the AI powered resume builders.
The automated career matching platforms designed to help people find
new jobs were themselves powered by the exact same kind
of generative AI technology that had just eliminated their old jobs.
Speaker 1 (25:34):
So the tech that fired you is now helping you
look for work.
Speaker 2 (25:37):
Pretty much. Critics rightly jumped on this, pointing out that
NESLI was essentially adding algorithmic insult to injury, using the
technology that caused the crisis to provide a largely ineffective
automated solution to it.
Speaker 1 (25:50):
Let's dig a bit deeper into this trend. We called
the white collar genocide earlier. That might be strong phrasing,
but it captures the surprise, why were these office rules.
These information workers hit harder in this way than the
traditional factory jobs. That seems counterintuitive to many people.
Speaker 2 (26:06):
It does seem counterintuitive, but it makes sense when you
look at what the current generation of AI, especially generitive AI,
is actually good at. Blue collar tasks often involve complex
physical dexterity, handling unpredictable materials, working in messy or unstructured environments,
things that are still quite challenging for general purpose robots.
Speaker 1 (26:26):
Right, robots are good at repetitive, predictable physical tasks exactly.
Speaker 2 (26:31):
But JENAI excels at tasks involving information, reading, summarizing, analyzing data,
generating reports, writing code, creating marketing copy, even making complex forecasts.
These are the core functions of middle management, finance departments,
marketing teams, analysts, the heartland of white color.
Speaker 1 (26:49):
Work, and the data backs this up. The risk analysis.
Speaker 2 (26:52):
Yeah, estimates from places like Mackenzie are pretty stark, and
we're circulating widely. Around this time, they suggested that something
like forty five percent of core finance and accout tasks
were already highly automatable with current tech. For marketing analytics,
the figure was around thirty eight percent, but the highest
vulnerability basic administrative and data entry roles. The invoice processing
data validation were estimated at sixty two percent automatable.
Speaker 1 (27:13):
Sixty two percent.
Speaker 2 (27:15):
These are the jobs most directly exposed to tools like those UiPath,
RPA bots and the internal and SLAGPT.
Speaker 1 (27:21):
We can see this very clearly in the fay of
one specific team you mentioned, the Revenue Growth Management team.
This wasn't data entry. This was a team full of
highly educated, highly paid experts.
Speaker 2 (27:32):
Absolutely, this team was the epitome of high value human expertise.
It was one staff by around eleven hundred people globally,
many with PhDs advanced degrees in economics statistics. Top tier
analysts focus purely on complex pricing and promotion strategies to
maximize revenue.
Speaker 1 (27:50):
And after the AI implementation.
Speaker 2 (27:51):
They were ruthlessly cut down, reduced to just four hundred and.
Speaker 1 (27:54):
Twelve humans from eleven hundred to four hundred and twelve. Wow.
Speaker 2 (27:57):
Yeah, And crucially, those remaining four hundred and twelve p
M are now primarily focused on overseeing the AI. They're
supplemented by a massive AWS sage Maker cluster. That's Amazon's
cloud AI platform which runs, according to internal documents, forty
million different pricing and promotion simulations every single night.
Speaker 1 (28:14):
Okay, let's unpack that for listeners. An AWS sage Maker
cluster running forty million simulations nightly? What is that actually
doing and why does it outperform nearly seven hundred PhDs?
Speaker 2 (28:23):
Right? Sage Maker is Amazon's cloud platform that makes it
relatively easy to build, train, and deploy machine learning models
at enormous scale. This cluster uses parallel processing, basically lots
of computers working together to simultaneously test forty million different
scenarios every night. Think different price points in different regions,
bundled offers, competitor reactions, impact on volume versus margin, endless variations. Okay,
(28:46):
Now compare that to a human PhD analyst. Maybe they
could meticulously analyze what twenty different scenarios in a whole week.
The machine analyzes forty million complex scenarios in just a
few hours while everyone's asleep. The human capacity for analysis
at that time scale and speed simply cannot compete. The
human role inevitably shifts from doing the analysis to defining
(29:06):
the parameters. For the AI and interpreting its output. It's
a fundamental change in the nature of the job.
Speaker 1 (29:12):
So the ness lay story, as shocking as it is,
it's clearly not an isolated incident. You called it the blueprint,
and twenty twenty five really does seem to be crystallizing
as the year automation eight corporate Earth. Let's zoom out
now and place this nest lay blueprint within that wider context.
We're seeing parallel displacements driven by AI across major global
(29:33):
industries right absolutely.
Speaker 2 (29:34):
NESLEI was maybe the most high profile example in the
consumer goods sector, but it's just one domino falling in
a very long line. That same core philosophy using advanced
AI and robotics to drive radical efficiency gains, primarily through
headcat reduction, became standard corporate strategy across the board in
twenty twenty five. We have whole lists from resources of
parallel mass layoffs explicitly tied to adopting this kind of
(29:57):
advanced automation.
Speaker 1 (29:58):
Okay, let's run through some of those key industry starting
with logistics, maybe a sector heavily reliant on moving physical things.
Speaker 2 (30:04):
Right, look at UPS. They announced cuts of forty eight
thousand package sorterers and warehouse handling staff.
Speaker 1 (30:09):
Globally forty eight thousand.
Speaker 2 (30:11):
Yeah, the direct replacement advanced robotics, heavily leaning on systems
similar to those used by Amazon Robotics, plus their own
proprietary automated sorting systems. The need for human eyes and
hands for basic sorting in those massive hubs just plummeted.
Speaker 1 (30:26):
And Amazon itself they weren't immune.
Speaker 2 (30:28):
Far from it. Amazon cut thirty thousand middle managers. Now,
these weren't primarily warehouse workers. These were supervisors, team leads,
program managers. Their roles were essentially made redundant by increasingly
sophisticated AI managing the systems. Think about the just walkout
technology in their retail stores, eliminated and checkout staff and supervisors,
(30:49):
or the AI optimizing supply chain flows which removed layers
of human oversight.
Speaker 1 (30:53):
Even deep tech. The companies making the chips and software,
they started cutting the people who build the product.
Speaker 2 (30:58):
Yes, Intel announced cuts of twenty four thousand fab technicians.
Speaker 1 (31:03):
People working in the chip factories.
Speaker 2 (31:05):
Exactly highly skilled technicians. This displacement was driven largely by
the accelerating integration of things like extreme ultraviolet or EUV
lithography machines. These are incredibly complex, expensive machines, but they're
also increasingly automated and self monitoring. They simply require far
fewer human technicians on the floor to maintain and operate
(31:27):
the semiconductor fabrication lines.
Speaker 1 (31:29):
And the cuts even hit the software engineers, the coders themselves,
the people writing the AI.
Speaker 2 (31:35):
That was maybe the most ironic twist. Microsoft, a company
at the absolute forefront of AI development, cut nine thousand coders.
Why well, largely because tools like their own GitHub Copilot,
which is basically generative AI specifically trained to write and
debug computer code, started automating significant chunks of the routine
boilerplate coding work.
Speaker 1 (31:53):
So the AI is writing its own code now to.
Speaker 2 (31:55):
A large extent, yes, or at least assisting human coders
so much the nique viewer of them. We saw a
similar trend in IT services and consulting. Big players like
Tata Consultant News Services TCS cut twelve thousand software testers.
Their jobs were directly replaced by new AI driven automated
testing suites that can run continuous quality assurance checks much
(32:16):
faster and more comprehensively than manual testing teams, and the
overall numbers tracked by industry analysts. They reflect this acceleration dramatically.
Challenger Gray and Christmas, who track layoff announcements, reported a
really startling data point for twenty twenty five, eighty seven
thousand job cuts in the US were explicitly attributed by
the companies themselves to AI displacement. Eighty seven thousand just
(32:39):
in the US just explicitly linked to AI. And that
number was triple the tally they recorded for AI linked
cuts in twenty twenty four. So the acceleration isn't just steady,
it's exponential. It really confirms twenty twenty five as the
tipping point, the year corporate adoption of these powerful AI
tools translated into mass workforce displacement.
Speaker 1 (32:57):
Okay, but if we circle back to the factory floor
firm moment, the blue collar jobs that do remain, they're
changing drastically too, aren't they. They're becoming more specialized, often
focused on managing the robots themselves. Ness Lee had already
seen big efficiency gains there even before the twenty twenty
five cuts.
Speaker 2 (33:14):
Oh, definitely, Automation on the factory floor wasn't new, but
the sophistication ramped up look at their Kaler chocolate plant
in Brock, Switzerland. It's a historic site. Back in twenty eighteen,
it operated with about four hundred and twenty human workers. Okay,
by twenty twenty five was running with only two hundred
and eighty humans, but alongside them were one hundred and
fourteen Abbu Mei dual arm cobots. These are collaborative robots
(33:38):
designed to work safely alongside people.
Speaker 1 (33:40):
Fewer people, more robots. What did that do to output
in safety?
Speaker 2 (33:43):
The results were pretty dramatic. Output from the plant increased
by twenty eight percent, and maybe even more importantly, workplace
accidents dropped by a massive seventy one percent. Robots don't
get tired or make mistakes in the same way humans do, and.
Speaker 1 (33:55):
Those remaining two hundred and eighty humans their jobs must
be completely different.
Speaker 2 (33:59):
Now hardly different. The source material notes quite starkly that
their primary job now is to calibrate the robots that
replace their friends. Their work is about maintenance, programming the robots,
troubleshooting issues, handling exceptions the automation can't deal with. It's
not direct production work anymore. It requires a higher level technical.
Speaker 1 (34:18):
Skill set, and we saw similar stories in other regions
like Asia.
Speaker 2 (34:22):
Yeah. Another clear example was in their factory in Batongas
in the Philippines. One specific task there was manually inspecting
pouches of nest quick powder for defects, tears, bad seals,
printing errors.
Speaker 1 (34:34):
Used to be done by I.
Speaker 2 (34:35):
Used to be done by a team of human inspectors
visually checking pouch after pouch. Now a single sophisticated AI
powered vision system does the inspection. It scans eight hundred
units per minute, far faster than any human could, and
it replaced a team of twelve human inspectors on that
line alone. The AI's accuracy is higher, its speed is incomparable.
(34:56):
It just makes the human visual inspection role economically redundant
in that context.
Speaker 1 (35:00):
So as the sheer scope of this AI driven displacement
becomes clearer, hitting both white collar and blue collar, across industry,
is across the globe. The fundamental question for governments, for
society really is how do we absorb this shock, this
potential for mass unemployment for widespread skills obsolescens. It's huge.
We've seen some policy ideas floated, haven't we, But the
(35:22):
results haven't been particularly encouraging so far.
Speaker 2 (35:25):
No the policy response has really struggled to keep pace
with the technology. The debate around universal basic income or UBI,
flared up again intensely around twenty twenty five. It was
positioned by proponents as a necessary safety net for potentially
millions displaced by automation.
Speaker 1 (35:39):
And there's been trials, right, pilots.
Speaker 2 (35:42):
Yeah, pilots have run in various forms in places like Finland, Canada,
parts of the US like Stockton, California, even Kenya, but
the results have been really mixed and often quite controversial.
UBI seems to provide some basic income stability, which is good,
but critics argue, often quite forcefully, that it doesn't do
enough to encourage the necessary reskilling or active re engagement
(36:04):
with a rapidly changing job market. It risks creating dependency
without opportunity.
Speaker 1 (36:09):
Okay, So UBIS debated what about the idea of taxing
the automation itself, the robot tax concept, basically treating a
productive robot or algorithm as if it were a taxable
employee to help fund social programs or retraining.
Speaker 2 (36:22):
That idea gained some traction too, particularly in Europe. France
actually tried to implement a version of it, but the
outcome really illustrated the fundamental difficulty of regulating, let alone taxing,
highly mobile technology and capital.
Speaker 1 (36:34):
What happened in France well, Almost immediately.
Speaker 2 (36:36):
After the policy was seriously proposed, Boston Dynamics, a major
player in advanced robotics though not directly involved in Nesle's case,
announced they were shifting a significant portion of their R
and D operations out of Europe and into Singapore. Why Singapore,
because Singapore offered a much more favorable, essentially tax free,
regulatory environment for AI and robotics development. The message was clear,
(37:01):
you can't easily tax an algorithm, a piece of software,
or the intellectual property behind a robot because it can
be relocated digitally across borders almost instantly. Attempting to tax
it heavily in one jurisdiction just risks driving the innovation,
investment and jobs elsewhere. The French effort kind of fizzled out,
and even.
Speaker 1 (37:19):
When companies like Nesli make public pledges to help their
own displaced workforce, the actual budget numbers often tell a
different story, don't they a story of priorities?
Speaker 2 (37:28):
They really do. Nesle publicly played chf fifty million towards
employee retraining and resclling programs in the wake of the layoffs,
which sounds like a decent chunk of money.
Speaker 1 (37:38):
Fifty millions was Franks.
Speaker 2 (37:39):
Yeah, but critics were incredibly quick to put that number
in perspective. They juxtaposted against Nesle's technology investments. They pointed
out that roughly the same amount around CHF fifty million,
was what Nesle had spent purchasing about one hundred and
eighty large industrial robots just the previous year. Ah so
the investment in the automated replacement fundamentally dwarfed the investment
(38:02):
in human recovery and retraining. The priorities seemed pretty clear
despite the pr pledges.
Speaker 1 (38:07):
Okay, so policy is struggling corporate retraining budgets or maybe
less than meets the eye. What about the court of
public opinion consumer behavior. We saw a huge, very vocal
backlash online against the ethics of these cuts.
Speaker 2 (38:20):
Oh yeah, the ethical outrage was immediate and loud, especially
on social media. Hashtag boycott Nessele trended globally. You had
TikTokers doing these dramatic videos of themselves smashing kit cat
bars and protest.
Speaker 1 (38:32):
Right performative maybe, but visible very visible.
Speaker 2 (38:35):
And there were some tangible impacts. Initially. Sales of Nesli
Pure Life bottled water, for example, reportedly dropped by about
four percent on US college campuses in the week immediately
following the announcement. Campuses are often hotspots for this kind
of ethical consumerism, so there was definitely a strong negative
reaction in the digital sphere and amongst certain demographics.
Speaker 1 (38:55):
Did that outrage translate into a lasting hit to the
bottom line? Did consumers actually vote with their wallets in
large numbers?
Speaker 2 (39:03):
That's the crucial disconnect, isn't it. Despite all the online fury,
the hard sales data told a different story. Nesle's overall
Q three results, the ones released right alongside the layoff news,
still showed global confectionery growth was up six point eight percent,
so strong growth, still strong growth. It seems that while
many people might protest the ethics online, when they're actually
(39:25):
standing in the supermarket aisle, factors like price, convenience, brand, loyalty,
or maybe just the simple craving for a familiar chocolate bar,
those tend to win out. The data suggests consumers continued
to quote vote with their wallets and apparently still crave
crunch bars. The business imperative to maintain profitability through automation
(39:45):
clearly outweigh the short term pr risk from the ethical backlash.
Speaker 1 (39:50):
So with the financial markets validating the move, policy responses lagging,
and consumer boycotts proving somewhat temporary, what's the path forward
for Nesley itself? What does see EO Navertill's longer term
vision look like now through to say twenty twenty eight.
Speaker 2 (40:05):
The internal roadmap, parts of which were leaked, shows a
plan that's even more aggressive, really pushing the boundaries of
autonomous corporate operations. For twenty twenty six, the goal was
the full global rollout of something called zero touch planning.
Speaker 1 (40:17):
Zero touch planning what's that?
Speaker 2 (40:19):
This is where AI takes over almost the entire demand
forecasting and supply chain management process. The I predicts demand,
manages inventory levels, automatically generates production orders, and even optimizes
logistics down to the level of stocking and individual storeeshelf,
all with minimal or ideally zero human intervention.
Speaker 1 (40:38):
Okay, so automating the flow of goods. What about the
creative side product development? Is AI taking that over to increasingly?
Speaker 2 (40:45):
Yes? By twenty twenty seven, the internal target was for
forty percent of their entire global portfolio of skewed stock
keeping units, basically individual products to be managed by generative design.
Speaker 1 (40:56):
Generative design, so the AI is inventing products essentially.
Speaker 2 (40:59):
Yes, it means I won't just be optimizing existing products
or promotions. It'll be actively generating entirely new flavor profiles,
suggesting novel packaging designs, even formulating new product concepts based
on its analysis of market trends, competitor moves, and predictive
modeling of consumer desires. AI as inventor.
Speaker 1 (41:16):
And the ultimate end goal, the pinnacle of this automation drive, the.
Speaker 2 (41:19):
Twenty twenty eight target, the holy grail for this kind
of efficiency push was the commissioning of the first fully
autonomous Nesliy factory, a completely lights out human free facility
planned for somewhere in Poland.
Speaker 1 (41:32):
According to the documents, lights out, human free.
Speaker 2 (41:35):
That's the end point of this journey, where humans are
entirely removed from the physical production cycle. There was even
an internal slogan reported to be circulating within the digital
transformation teams that sums up this whole strategic shift from
two thousand brands to two thousand algorithms WOW.
Speaker 1 (41:50):
From brands to algorithms. That really says it all.
Speaker 2 (41:53):
Okay, So wrapping this deep dive up, looking back at
Nesle's silicon layoffs, it's crystal clear this wasn't just an
ice decision by one food giant. It really serves as
a stark blueprint for global corporate strategy in the age
of AI. This model integrating massive scale, generative AI with
specialized automation tools to fundamentally streamline or eliminate core white
(42:15):
collar functions like finance, marketing, supply chain planning, even aspects
of R and D. This is the playbook. It's the
playbook every major conglomerate is now studying, if not actively implementing.
Absolutely What really stands out from the Nestley case is
the sheer, stunning speed at which traditional, often highly valued
corporate expertise seem to just erode when confronted with targeted,
(42:36):
scalable automation, the pressure cooker environment, the need to hit
aggressive shareholder demands, compounded by those external shocks like commodity prices.
It basically compressed maybe a decade's worth of gradual technological
change into well forty eight hours of crisis driven decision making.
Speaker 1 (42:53):
We started with the shock of the announcement, we went
through the specific tech stack they use, the automation arsenal,
and we saw the painful human impact and the societal
friction it created. If there's one crucial takeaway from all
this for you, the listener, it has to be about
the immediate, urgent need to reassess essential career skills. The
ground has fundamentally shifted beneath our feet.
Speaker 2 (43:13):
Definitely think about that letter to the class of twenty
twenty six. It was apparently circulated widely, partly inspired by
the NESLE events. The core message was blunt. If you're
graduating now or soon with traditional degrees, even prestigious ones
like MBAs or specialized masters in things like supply chain management,
(43:34):
those qualifications are potentially becoming obsolete pert least significantly devalued
by the very AI systems they might teach you about.
The education system is almost inevitably training people for the
jobs of the last decade, not the next one.
Speaker 1 (43:47):
Which raises the unavoidable and frankly quite daunting question. If
the analysis, the administration, the planning, even some of the
creative work is being automated, what skills remain essential? What
skills are defeat sensible. In this new landscape.
Speaker 2 (44:01):
The skills that seem most valuable now are highly specific
and often quite technically focused. You almost certainly need to
become adept at prompt engineering, which sounds simple, but it
is actually the complex art of communicating effectively with these
powerful AI engines to get the optimal desired output. It's
about knowing how to ask the right questions in the
right way.
Speaker 1 (44:19):
Okay, prompt engineering, what else you need?
Speaker 2 (44:21):
People who understand robot maintenance, not just the physical upkeep
of machines on a factory floor, but also the software maintenance,
the continuous monitoring and fine tuning of the algorithms themselves,
keeping the automated systems running smoothly.
Speaker 1 (44:36):
But maybe the most important skill, the only truly secure
role long term, is perhaps broader than just the technology itself.
Speaker 2 (44:43):
I think that's right, And maybe this is the most
provocative but also hopeful thought to end on the conclusion
many are reaching. Is this the only jobs AI can't automate?
At least not yet and maybe not ever. Are the
ones involved in inventing and defining the rules that govern
them AI itself.
Speaker 1 (45:00):
How we're talking about.
Speaker 2 (45:01):
Policy development, ethical framework design, crafting regulations, defining standards for
human machine collaboration, shaping the strategy for how AI integrates
into society and the economy. These metal level roles are
potentially safe because they involve setting the boundaries, defining the goals,
and managing the societal impact rather than just executing tasks
within the system the AI now runs.
Speaker 1 (45:22):
Nesley certainly showed the world what radical AI driven disruption
looks like in practice. Sixteen thousand families suddenly rewriting their
resumes while servers hum quietly in data centers in Switzerland
and elsewhere, delivering unprecedented efficiency gains directly to shareholders. The future,
that sci fi future we used to talk about, it's
(45:44):
not distant anymore. Yeah, it's here, it arrives, it's here,
and it's demanding new skills, new thinking, and maybe entirely
new rules starting right now.
Speaker 2 (45:52):
It really is a profound wake up call for everyone
in the corporate world and beyond.
Speaker 1 (45:56):
Thanks for diving deep with us on this critical story.
We'll see next time on the Deep Dive
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.
Las Culturistas with Matt Rogers and Bowen Yang
Ding dong! Join your culture consultants, Matt Rogers and Bowen Yang, on an unforgettable journey into the beating heart of CULTURE. Alongside sizzling special guests, they GET INTO the hottest pop-culture moments of the day and the formative cultural experiences that turned them into Culturistas. Produced by the Big Money Players Network and iHeartRadio.
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