The future will belong to corporations with hybrid collective intelligence

Human agents and AI agents operating together.

The race for artificial intelligence has turned into an investment race. In 2025, global spending on generative AI was estimated at US$644 billion, a 76.4% increase in just one year, according to Gartner. Better models, GPUs, infrastructure, platforms, and new applications are absorbing growing volumes of capital.

But investment doesn't necessarily mean impact.

BCG's 2025 global survey shows a relevant gap: 60% of companies still derive little or no material value from their AI investments. Only 5% are in the group considered most advanced, the so-called future-built companies. This small group manages to generate five times more revenue growth and three times more cost reduction from AI than other organizations.

At the same time, companies are committed to deploying AI agents in their operations.

But where to start, and with what purpose?

McKinsey clearly shows the size of this gap between adoption and transformation. In 2025, 88% of organizations were already regularly using AI in at least one function, but only about a third had begun to scale AI across the company. When we look specifically at agentic AI, 62% were already experimenting with or scaling agents, yet only 23% had begun to scale them in some part of the organization.

In other words: there is adoption. There is investment. There is experimentation.

What is still missing, in most organizations, is transformation.

The real competitive advantage should be built by using agentic AI. But to produce relevant impacts, its implementation needs to be thought of far beyond automating repetitive processes.

Our thesis is that every corporation, regardless of size or sector, should prioritize the application of agentic AI to enhance its collective capabilities to learn, decide, innovate, and continuously generate value.

These collective capabilities have always been sources of corporate power. They are knowledge, practices, relationships, and decision-making methods that are often tacit and intrinsic to organizational cultures. With the correct use of AI, these capabilities can be rediscovered, structured, and amplified.

Let me be direct.

It will not be simply access to the best AI that will determine the winning corporations of the next ten years.

It will be their collective intelligences.

AI will be the infrastructure capable of amplifying these capabilities.

Most companies have not yet fully realized this difference. Many continue to see AI primarily as a tool for automation and efficiency gains. They put a chatbot in a department, automate processes, reduce time or cost, and consider this transformation.

The numbers show the limit of this approach. In McKinsey's survey, 80% of organizations list efficiency among the objectives of their AI programs. But the companies that capture the most value are precisely those that combine efficiency with growth and innovation objectives and redesign their workflows.

The problem, therefore, is not automation.

It is stopping at automation.

When a company merely optimizes its current processes, it may gain productivity without necessarily building new capabilities for the business.

Transformation begins when, instead of simply removing people from repetitive work, we start asking how AI can enhance the organization's human and collective capacity.

This is the inversion we need to make.

What is collective intelligence really?

Collective intelligence is part of the culture. It is the ability of employees to understand and operate the business, producing responses, knowledge, and decisions that an individual alone would hardly be able to produce.

Caring for collective intelligence means allowing the organization to function more integrally, as a system that increases its own learning capacity as it operates.

Many organizations struggle to achieve this integration.

Knowledge remains trapped in silos. Each department has its answers, its experts, its systems, and its processes.

When someone new arrives, they often have to rebuild part of the learning.

When an experienced person leaves the company, an important part of their tacit knowledge leaves with them.

Experts operate individually, instead of integrating into a system that learns collectively.

Tacit knowledge is not captured.

Learning does not circulate.

This is precisely where AI can play a much larger role than simply executing tasks.

The illusion of automation

Process automation using Agents, by itself, does not transform an organization.

There is an apparently irresistible logic: if an AI tool can perform a certain task faster or cheaper than a human, why not automate it?

Because local efficiency gains do not guarantee an increase in the organization's intelligence.

And this distinction is already starting to appear in the data.

Deloitte points out that 66% of organizations already report productivity and efficiency gains with AI, but only 34% claim to be truly redesigning or reimagining the business with the technology.

This gap is fundamental.

A company can produce faster and still operate essentially the same way.

It can automate hundreds of tasks without creating a single new strategic capability.

The immediacy of automation also risks reproducing and accelerating problems present in current processes.

If we automate a poorly designed process, we will have a poorly designed process operating faster.

And there is another risk: when everything routine simply disappears from human experience, we also need to rethink how people learn, develop judgment, build context, and acquire business knowledge.

AI should not be used to eliminate knowledge.

It should be used to enhance the ability of groups to better understand the organization, combine perspectives, and produce better decisions.

The transformation from the current digital organization to a new agentic organization should start from this principle.

Hybrid collective intelligence

In the age of AI, collective intelligence can gain a new dimension.

It ceases to be solely human and begins to result from the interaction between people and AI Agents sharing organizational knowledge, technologies, models, data, and processes.

We call this hybrid collective intelligence.

A hybrid organization is composed of people working together with AI agents, creating new high-impact capabilities for the business.

Each part functions according to a new organizational design, delivering what it does best.

Human Agents defining purpose, shaping context, thinking critically, making decisions in ambiguous situations, innovating, and creating new value.

• Amplifying collective memory: helping to capture knowledge in a structured way, retrieve it when needed, and connect dots that would individually remain isolated;

• Improving collective attention: structuring priorities, connecting objectives, organizing information, and simplifying communication;

• Enhancing collective reasoning: bringing perspectives, simulating scenarios, questioning assumptions, expanding the space of ideas, and offering data and alternatives that enrich decision-making.

In none of these cases is the principle simply to hand over the entire decision to AI.

The goal is to enhance humans' ability to understand, evaluate, and decide better.

This is the new standard.

Not AI or human.

AI and human.

Design integral: the method for developing the agentic organization

But here's another critical point: this doesn't happen automatically.

Hybrid collective intelligence requires a new method for designing and developing the organization, going beyond traditional software engineering.

The data reinforces this need.

In McKinsey's survey, organizations classified as AI high performers represent only about 6% of respondents. They are nearly three times more likely to fundamentally redesign their workflows and about three-quarters of them are already scaling AI, compared to about a third of other organizations.

The difference, therefore, is not solely in the technology used.

It lies in how the organization is redesigned to use it.

This new method needs to be applied both in rediscovering the current organization's knowledge and in designing the new agentic organization.

In the development of Squadra Agêntica, we have been applying the integral design method for two years.

• Rethinking the entire organization holistically: current capabilities, people, processes, knowledge, data, innovation, future capabilities, and technologies;

• Leaders understanding their new role: creating an environment that stimulates learning and the development of collective intelligence, not just controlling execution;

• A new organizational culture that values continuous learning, not just speed;

• Governance, security, and responsibility sufficiently distributed to allow for quick, cohesive, and value-generation-oriented decisions.

Agentic corporate capability and process automation

This is where most companies get lost.

While AI automation reduces costs, a new capability created with AI can generate value continuously.

While automation seeks to remove a specific activity from the process, an agentic capability can enhance what human talent can achieve within it.

Of course, there is room for automation.

Genuinely repetitive tasks, with low cognitive value and in which no relevant learning is involved, are natural candidates for automation.

But the ambition cannot stop there.

The very performance difference between companies suggests this. According to BCG, the most advanced organizations are not using AI solely for productivity: they have implemented capabilities necessary to apply AI also for innovation and reinvention. Only 5% of the evaluated organizations have reached this stage, while 60% still report minimal material gains despite investments.

Ambition needs to start from the legacy culture and ask how it can be amplified by the use of AI:

• How does AI allow my team to become even more specialized?

• How does it allow tacit knowledge to transform into explicit and shareable knowledge?

• How do AI and integral design allow us to rediscover the current culture and, above all, design new capabilities for the organization?

• How does AI connect people who need to be connected?

• How does it put the right knowledge in the right hands, at the right time?

• What capability will the business possess tomorrow that it did not possess before AI?

The last one is perhaps the most important question.

When we ask these questions, everything changes.

Decisions remain human.

But they become decisions supported by a system designed to enhance the organization's collective intelligence.

The new role of leadership

This requires a profound change in the role of leadership.

The traditional leader was the one who had the answers, controlled information, decisions, and processes, and guided people on what should be done.

In a world where AI expands access to information, analysis, and execution capabilities, this model loses strength.

The advantage is no longer solely in having one person who knows more than others.

It becomes in building a group capable of learning faster than its competitors.

This becomes even more evident when we observe the daily use of AI. In a BCG survey published in 2026, 74% of frontline workers were already using AI regularly. Among regular users, 42% reported saving about eight hours per week. But BCG itself points out that many organizations have not yet discovered how to transform this saved time into business value.

This is perhaps one of the biggest challenges for leadership now.

It is not enough to provide tools.

It is necessary to convert technological capability into organizational capability.

The leader's role becomes developing their group's ability to learn, think together, and decide better.

It becomes designing the processes, knowledge, relationships, and technology that allow collective intelligence to emerge.

It becomes questioning, exploring different perspectives, and creating space for curiosity, experimentation, and critical thinking.

This is harder.

But it is also much more powerful.

Why does all this matter now?

Because the gap between those who are simply adopting AI and those who are reorganizing around it has started to become measurable.

BCG found that only 5% of companies today belong to the most advanced group in AI maturity, while 35% are starting to scale and generate value, and 60% remain with little material impact. The companies in the first group already achieve five times more revenue growth and three times more cost reduction with AI than the others.

And agentic AI is beginning to participate directly in this difference.

According to the same survey, agents already accounted for approximately 17% of the total value generated by AI in 2025, a share projected to reach 29% by 2028. Among the most advanced companies, one-third already use agents, compared to only 12% of organizations that are still in the scaling stage and almost none of the companies lagging in adoption.

But it is important to realize what differentiates these organizations.

It is not simply access to better models.

The most advanced models are increasingly available to everyone.

The differentiator lies in how they redesign decisions. How they learn. How knowledge circulates. How people collaborate. How technologies amplify human capabilities. How experimentation and innovation become part of operations.

McKinsey's numbers reinforce this point: among organizations achieving significant AI impact, workflow redesign appears as one of the factors with the greatest contribution to generating real business impact.

Therefore, the bottleneck is ceasing to be exclusively technological.

It is also organizational.

It is about mindset.

It is about culture.

It is about design.

When we realize this, everything changes.

We stop looking only for tools and start designing capabilities.

We stop measuring transformation by the number of automated tasks and start asking what new capability the corporation has acquired.

We stop thinking only about replacing people and start thinking about how to enhance their skills.

Because organizational intelligence needs to be rediscovered, rebuilt, and developed.

It does not come ready-made within an LLM.

We build it by creating learning channels, capturing knowledge, amplifying information, connecting perspectives, aligning groups around purposes, and allowing people to think critically together.

AI amplifies all of this.

But the foundation remains organizational, not just technological.

Read also on LinkedIn.