McKinsey & Company has quietly deployed 25,000 AI agents alongside its 40,000 consultants — and this changes far more than productivity.
*https://www.youtube.com/watch?v=UxVRrorZ4RM
**
таймкоды
00:00:00 Введение в цифровую армию McKinsey
- McKinsey использует 25 000 агентов искусственного интеллекта.
- Общая операционная мощность компании составляет 35 000 специалистов-людей и цифровых агентов.
- Структура консалтингового бизнеса меняется, разрушая старые модели.
00:00:52 Старая модель консалтинга
- Консалтинговый бизнес основывался на пирамиде: старшие партнёры, руководители, младшие аналитики.
- Младшие аналитики выполняли рутинную работу, создавая бесконечные слайды в PowerPoint.
- Модель была медленной и дорогостоящей, но требовала больших человеческих ресурсов.
00:01:50 Необходимость изменений
- Корпоративные клиенты требовали более быстрых результатов.
- Старый способ работы становился помехой в мире, где ценятся скорость и исполнение.
00:02:48 Внедрение искусственного интеллекта
- McKinsey внедрила искусственный интеллект и цифровую рабочую силу.
- Цель — иметь одного агента с ИИ в паре с каждым сотрудником-человеком.
- Цифровая армия построена на двух платформах: Lilly AI и Quantum Black.
00:03:44 Платформа Lilly AI
- Lilly AI предоставляет мгновенный доступ к базе знаний фирмы.
- Инструмент экономит время на исследованиях и анализе.
- За год инструменты сэкономили компании полтора миллиона часов работы.
00:04:36 Платформа Quantum Black
- Quantum Black создаёт и внедряет системы ИИ промышленного уровня для клиентов.
- Команда из более чем 1700 специалистов работает над проектами.
- McKinsey занимает около 40% рынка работы с клиентами по ИИ.
00:05:24 Принципы успеха
- Первый принцип: хранение и очистка данных для безопасного повторного использования интеллектуальной собственности клиентов.
- Второй принцип: оценка качества по объективным критериям, а не по «показателям тщеславия».
- Третий принцип: управление ИИ как основным продуктом, а не как ИТ-проектом.
00:06:23 Оценка эффективности ИИ
- McKinsey проводит слепые сравнения результатов с помощью Lilly и без неё.
- Критерии оценки: точность, полнота и уникальность.
- ИИ оценивается по тому, насколько объективно он улучшил конечный результат.
00:07:22 Искусственный интеллект как инфраструктура
- Искусственный интеллект становится инфраструктурой, что позволяет ему перестать быть обязательным элементом.
- McKinsey использует ИИ для автоматизации задач, сокращая затраты и высвобождая ресурсы для стратегического консультирования.
- Автоматизация позволяет сосредоточиться на том, что ценят клиенты: стратегическом консультировании, управлении заинтересованными сторонами и суждениях руководителей.
00:08:22 Переход от оплаты времени к оплате результатов
- Консалтинговая индустрия традиционно платила за человеческое время, но ИИ разрушает эту модель.
- McKinsey переходит от продажи времени к продаже результатов, предлагая клиентам конкретные измеримые бизнес-результаты.
- Вознаграждение McKinsey зависит от достижения этих результатов, что кардинально меняет правила игры.
00:09:18 Преимущества новой модели
- Клиенты предпочитают новую модель, так как платят за ощутимую ценность, а не за усилия.
- McKinsey предпочитает новую модель из-за предсказуемости и масштабируемости результатов.
- Модель McKinsey становится прообразом будущего для всех профессий, основанных на знаниях.
00:09:59 Изменения в структуре компаний
- Традиционная пирамида с младшими аналитиками превращается в ромб из-за автоматизации рутинной работы.
- Ценность приобретают специалисты, умеющие формулировать вопросы для ИИ, проводить стресс-тестирование результатов и разрабатывать стратегии.
- McKinsey ищет универсальных специалистов, одинаково хорошо разбирающихся в технологиях, данных и стратегическом планировании.
00:10:51 Влияние на другие отрасли
- Модель McKinsey, сочетающая ИИ и элитные команды, становится образцом для других отраслей.
- Будущее за компаниями, которые смогут создать и внедрить собственную цифровую рабочую силу для достижения измеримых результатов.
- Эпоха продажи человеко-часов подходит к концу, уступая место новой модели.
00:12:43 Будущее профессиональной деятельности
- ИИ освобождает специалистов-людей для сосредоточения на стратегии и взаимоотношениях с клиентами.
- Переход от продажи времени к продаже гарантированных результатов превращает ИИ в мощный экономический двигатель.
- Самые влиятельные операторы следующего десятилетия будут те, кто сможет эффективно объединить элитные человеческие таланты с интеллектуальной цифровой рабочей силой.
00:13:29 Заключение
- Самыми влиятельными будут те, кто сможет наиболее эффективно объединить элитные человеческие таланты с интеллектуальной автономной цифровой рабочей силой.
In this video
Chapters
Transcript
Why McKinsey’s 25,000 AI Agents Change Consulting Forever
0:00
The world’s most prestigious consulting
0:02
firm has a workforce it doesn’t talk
0:05
about. Not a few chatbot, but an army of
0:08
25,000 AI agents. In a recent reveal,
0:12
CEO Bob Stern confirmed McKenzie’s total
0:15
[music] operating capacity is its 35,000
0:19
human experts plus this massive digital
0:22
force. For any ambitious professional, a
0:24
job at McKenzie has always been seen as
0:27
ultimate career leverage. But now the
0:30
structure that created that leverage is
0:33
being systematically
0:36
re-engineered from within. Stay with me
0:38
and I’ll explain the mechanics of this
0:41
digital army. How it’s creating a new
0:44
economic engine and what it signals for
0:48
anyone still playing by the old rules of
0:51
elite knowledge work. Hey, if you’re new
0:53
here, this is Ramen and I’ve spent the
0:55
last decade inside consulting, including
0:58
top tier consulting firms, and I’m still
1:01
in the system today. Let’s dive in.
How the Consulting Pyramid Actually Works
1:05
To understand how profound this shift
1:08
is, you have to understand the old
1:10
consulting model. For decades, the
1:13
business was very simple, immensely
1:16
profitable and built on a human
1:20
intensive leverage structure. It was a
1:23
pyramid. [music] At a top, a handful of
1:26
senior partners who own the client
1:29
relationship. Below them, layers of
1:32
principles and managers tasked to
1:34
deliver those engagement. And at the
1:37
bottom of this pyramid, [music] a
1:39
massive base of junior analysts hire out
1:43
of the [music] world’s top universities.
1:45
The life of a junior analyst was a right
1:48
of passage defined by brutal hours.
1:50
Their function was to be a human search
1:53
engine and a presentation [music]
1:55
factory. They spent allnighter grinding
1:57
through data rooms, firsthand
1:59
interviews, and market reports. The
2:01
output was always the same. Endless
2:05
PowerPoint slides. The core currency of
2:08
consulting work. The entire system was
2:11
built on one financial principle,
2:15
leverage. A firm could build a client
2:18
for thousands of hours of this junior
2:21
[music] level work forming the economic
2:23
foundation of every project. It was
2:26
relatively slow, expensive, and it was
2:30
simply how the game was played. But that
2:33
model’s foundation were cracking.
2:35
Corporate clients began demanding faster
2:38
tangible result, not thicker slide deck.
2:41
The old way of working was becoming a
2:44
liability in a world that values speed
2:47
and execution. McKenzie knew that to
2:49
maintain its position in the premium
2:52
consulting space, it couldn’t [music]
2:54
just optimize the old model. it had to
2:57
fundamentally [music]
2:59
rewire its own DNA. So, how did they
3:03
rewire?
How AI Is Rewriting McKinsey’s Consulting Economics
3:06
McKenzie didn’t just adopt AI, they
3:08
deploy a digital [music] workforce. In
3:11
early 2026, CEO Bob Sternfills announced
3:15
that McKenzie’s operating capacity was
3:19
not just its 35,000 employee, but also
3:23
25,000
3:24
AI agent. The firm’s stated goal is to
3:29
have at least one AI agent paired with
3:32
every human employees. [music]
3:35
This digital army is built on two viable
3:39
platform. The first is a proprietary
3:41
generative AI platform named Lily. Lily
3:45
AI platform acts [music] as McKenzie
3:47
institutional memory. It gives every
3:50
consultant instant access to the firm’s
3:54
entire knowledge base, every sanitized
3:58
case, [music] report and expert
4:00
transcript. When a consultant poses a
4:02
question, Lily doesn’t just return
4:04
documents. It synthesize critical
4:07
information, draft initial summaries,
4:10
and identifies the [music] most relevant
4:13
internal expert. The firm estimates its
4:17
AI tool are saving vast amount of time
4:20
on research and analysis in a single
4:23
year. Apparently, these tools saved the
4:26
company 1 and a half million hours of
4:30
work. The second initiative is Quantum
4:32
Black, McKenzie’s AI delivery arm. Think
4:36
of Quantum Black as the firm’s [music]
4:38
AI factory. With a team of over 1,700
4:42
data [music] scientists, engineers, and
4:45
product managers, Quantum Black doesn’t
4:47
just advise on [music] AI strategy. It
4:50
builds and deploy production AI system
4:54
for client. And according to its
4:56
leadership, it’s now drives about 40% of
5:00
McKenzie’s client work. At this point,
5:03
[music] you might think McKenzie just
5:05
licensed the powerful LLM and
5:07
integrated, but that’s a surface level
5:10
view. And it’s why most companies trying
5:13
to replicate this will fail. Lily and
5:16
Quan Black only function because of
5:19
three pillars that support them. The
5:22
first is data storage. McKenzie already
5:26
had hundreds of people globally whose
5:28
full-time job wasn’t client [music]
5:31
strategy. It was content hygiene. These
5:34
data towards curate the firm’s internal
5:37
knowledge, cleaning it, verifying it,
5:40
and [music] stripping out sensitive
5:42
client IP to ensure it can be safely
5:46
reused. This is the only reason a system
5:49
like Lily is viable. Without this layer,
5:53
an internal AI doesn’t create
5:55
intelligence, it creates risk. A single
5:58
breach of confidentiality destroys
6:01
client trust and without trust there is
6:04
no consulting business. This is the step
6:07
most companies skip because it isn’t
6:09
innovative but it’s everything. The
6:12
second pillar out of these three hidden
6:15
pillars is measuring quality not
6:18
adoption. McKenzie didn’t measure
6:21
success with vanity metric. They didn’t
6:24
just ask are people using the tool. They
6:26
ran blind comparisons, scoring [music]
6:30
outputs created with Lily against those
6:34
created without it. The criteria were
6:37
accuracy, richness, and [music]
6:39
distinctiveness. In other words, the AI
6:42
wasn’t judged on whether people liked
6:44
it. [music] It was judged on whether it
6:46
made the final work product objectively
6:50
better. That distinction is critical
6:52
because it reframes who and what is
6:56
truly valuable. Finally, the third
6:59
pillar out of these three hidden pillar
7:02
is treating AI like a core product.
7:05
McKenzie managed Lily not as an IT
7:08
project but as a software company
7:10
managers a flagship product. Today,
7:13
apparently roughly 70% of the firm
7:16
actively uses it, generating hundreds of
7:19
thousands of prompt months. That’s not
7:21
an experiment. That’s infrastructure.
7:24
And once AI [music] becomes
7:26
infrastructure, it’s no longer optional.
7:28
When you combine these elements, curated
7:31
7:31
data, rigorous quality control, and
7:34
product level discipline, McKenzie isn’t
7:37
just using AI. It has operationalized
7:41
it. And that’s what unlocks the new
7:43
[music] economic engine. First, there is
7:47
the obvious win. Massive cost reduction
7:49
at the base [music] of the pyramid by
7:52
automating task that once consumed
7:54
thousands of junior consulting hours.
7:57
This automation frees up expensive human
8:01
consulting to focus on what clients
8:04
actually value. high level strategic
8:07
council, stakeholder management and
8:10
executive judgment. But the real
8:13
strategic insight is how this efficiency
8:17
destroyed the old business model to
8:20
create a more lucrative one. The
8:23
consulting industry has always run on
8:26
billable hour, a fee for a service model
8:29
where clients pay for human time. AI
8:32
[music] shatters that model. When an AI
8:35
can perform 30 hours of analysis in just
8:39
30 minutes, it can no longer credibly
8:42
bill for time. So McKenzie leaned into
8:46
the [music] disruption. They began
8:48
moving aggressively away from selling
8:50
time and towards selling results. This
8:53
is the outcomebased [music]
8:55
pricing, right? Instead of paying for a
8:57
team’s time, a client agrees on a
8:59
specific [music]
9:00
measurable business outcome like a 5%
9:04
increase in profit margin or [music] 10%
9:07
reduction in customer churn. McKenzie’s
9:09
fee is then tied directly to its ability
9:12
to deliver that result. And this model
9:16
is a gamecher. Client prefer it because
9:19
they’re paying for tangible value, not
9:22
just effort. and McKenzie prefer it
9:25
because [music] its AI powered delivery
9:28
model makes those outcome more [music]
9:30
predictable and scalable with platforms
9:33
like Lily and repeatable solutions from
9:35
quantum blacks. What McKenzie is
9:37
executing here is more than a strategy.
9:41
It’s a preview of future for all
9:43
knowledgebased professions and the
9:46
implications are structural and this is
9:49
the part you have to be aware if you are
9:51
planning to get into consulting
9:53
especially McKenzie and other top
What This Means for Consulting Careers and Entry-Level Roles
9:56
[music] tier firm.
9:59
First the shape of elite firms are
10:01
changing. The traditional pyramid with
10:04
its wide base of junior analysts is
10:07
inverting into a diamond. The
10:09
commoditized work at the bottom is
10:11
[music] being automated. This doesn’t
10:13
just mean fewer entrylevel jobs. It
10:16
means the job itself is changing. The
10:19
valuable skill is no longer about doing
10:22
the research. It’s about framing the
10:24
right question for the AI stress testing
10:26
its output [music] and weaving those
10:29
insight into a coherent strategy for the
10:32
client. The premium is shifting from
10:34
analysis to judgment. Second, this
10:38
changes who gets hired and [music]
10:40
promoted. McKenzie is now actively
10:42
hunting for what I call hybrid talent.
10:46
People [music] who are equally fluent in
10:48
technology, data, and highlevel
10:51
strategy. This focus is shifting from
10:53
the prestige of your alma matter to the
10:56
tangible crossfunctional skill you
10:59
possess. Finally, this model combining
11:03
specialized AI agents with elite human
11:05
teams and tying compensation to
11:08
performance is a blueprint that will be
11:11
replicated across industry. Law firms,
11:14
accounting firms, marketing agency and
11:16
software development and so on so forth
11:19
are all built on a similar leverage
11:22
[music] model that is now vulnerable.
11:25
The era of selling human hours is
11:28
closing. The future in my view belongs
11:31
to those who can build and deploy their
11:35
own digital workforce to deliver
11:39
measurable [music] results. So
11:41
McKenzie’s AR army show that this
11:44
technology isn’t a tool to makes us
11:47
slightly more efficient at our old jobs.
11:50
It’s a force that is fundamentally
11:53
rewriting the rules of how professional
11:56
value is created and captured. This
11:59
raises a critical question not for your
12:01
company but for your career. What is the
12:05
core leverage model you’re currently
12:08
part of? Is it based on billing hours,
12:11
managing junior talents, or owning a
12:14
niche [music] process? And is that
12:15
leverage model about to be broken by a
12:18
system like this? Think about it. Share
12:21
your thoughts with me in the comment
12:22
section below. I love to hear your
12:24
thoughts and perspective. The answer
12:27
will probably detai [music] your next
12:30
move. So what is the real takeaway?
12:34
McKenzie’s 25,000 strong AI army isn’t
12:38
about replacing human. It’s about
12:41
releveraging
12:42
them by automating low-level analytical
12:45
work. They have freed their human
12:48
experts to [music] focus exclusively on
12:51
high value strategy, problem solving and
12:55
client relationships. More [music]
12:57
importantly, McKenzie have used that
12:59
efficiency to flip their entire business
13:03
model, moving from selling time to
13:06
selling guaranteed [music] outcomes.
13:08
This shift is turning the AI
13:10
infrastructure into a powerful economic
13:13
engine and creating a defensible
13:17
competitive mode. And remember, this
13:20
isn’t just the story of one consulting
13:22
[music] firm. It’s a playbook for the
13:25
future of professional work. And the
13:28
message is clear. The most powerful
13:30
operators of the next decade won’t be
13:33
the ones with the largest teams, but the
13:36
ones who can most effectively integrate
13:39
[music] elite human talent with an
13:41
intelligent autonomous digital [music]
13:44
workforce. Want to know how to land your
13:47
dream job in business and finance in the
13:49
AI era? Click the next video here where
13:52
I share [music] with you how I did it