Billions of AI agents are entering the workforce, but most companies are still asking the wrong questions
The workforce as we know it is changing faster than most companies can keep up. And this time it is not an incremental update or another corporate buzzword. We are looking at a structural shift that will redefine how organizations operate, compete, and survive in the years ahead.
According to IDC projections, by 2029 more than 1 billion AI agents will be active in the market, executing around 217 billion actions per day. We are not talking about prototypes in a testing environment or experimental projects in a pilot phase. These agents are already analyzing data, making decisions, and carrying out real tasks inside real companies, in real time, with a direct impact on business results.
The problem is that a large number of organizations are still trying to figure out where this digital transformation even begins. And the biggest mistake is not the lack of technology or budget. It is that most are still asking the wrong questions. 🤔
While market leaders are already running thousands of AI agents in production, others are still debating which language model to pick or which vendor to hire. That gap, which seems small today, is about to become existential.
Here you will understand why companies still treating AI as a side project are falling behind, which questions actually matter right now, and what separates those who will lead this new era from those who will just watch from the sidelines. 🚀
What AI agents are and why they change everything
Before diving into the strategic side, it is worth getting on the same page about the concept. AI agents are not simple chatbots or virtual assistants that answer questions. They are autonomous systems capable of perceiving the environment around them, making decisions based on defined objectives, and executing actions independently, without needing a human to approve every step.
That autonomy is what makes the current landscape fundamentally different from any previous technology wave. When we talk about traditional automation, we are talking about scripts and fixed rules. When we talk about AI agents, we are talking about systems that learn, adapt, and evolve with every interaction.
In practice, an AI agent can monitor sales pipelines, identify missed opportunities, draft personalized follow-ups, update the CRM, and generate consolidated reports — all while the sales team focuses on high-value negotiations. In tech environments, agents are already reviewing code, detecting vulnerabilities, suggesting performance improvements, and documenting entire systems at a speed no human team could match. And this is happening now, not in 2030.
What makes this movement even more relevant for business strategy is the compounding effect. A well-configured agent does not just execute a task. It learns from each iteration, improves its performance over time, and can be replicated instantly. While hiring, training, and scaling a human team takes months and consumes significant resources, scaling a workforce based on AI agents can happen in days. That speed asymmetry is at the heart of the disruption underway.
The gap between ambition and execution
A recent EDB survey revealed a data point that perfectly illustrates the size of the problem: 95% of global organizations want to become their own AI and data platforms, but only 13% are actually managing to do that successfully today. That chasm between intention and reality is where most companies find themselves right now.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Kevin Dallas, CEO of EDB, puts the situation bluntly: the distance between leaders and laggards in AI is about to widen dramatically. When one organization is running thousands of agents and another is still running pilot projects, the gap stops being marginal and becomes existential.
And the most interesting part is that this growing distance may have less to do with the lack of a strategy than with the questions companies are asking in the first place. In many cases, organizations start by discussing which AI model to use, which vendor to hire, or where to position the pilot before even aligning the business priorities that AI should be serving. It is like choosing the color of the car before deciding where you want to go.
Why most companies are still behind
The most honest answer is that many organizations still see artificial intelligence as an additional layer on top of their existing processes, rather than a structural change in how they operate. This creates a vicious cycle: the tech team experiments with tools in isolated environments, produces feasibility reports, presents them to leaders who do not have enough context to make bold decisions, and the project ends up shelved or reduced to a marginal use case.
Charlene Li, a digital transformation expert who has advised several Fortune 100 companies, has a direct diagnosis on the matter. According to her, companies are still approaching AI like a hammer looking for a nail to hit. In Li’s view, shared in a recent episode of the AI & Data Horizons podcast, the approach is fundamentally backwards.
Another factor contributing to this delay is the confusion between one-off automation and real transformation. Installing a chatbot on your website or automating welcome emails is useful, but it is not transformation. Digital transformation with AI agents happens when these systems start operating in high-impact processes like dynamic pricing, supply chain management, real-time risk analysis, or personalization of experiences at scale. The difference between the two approaches is not technical — it is about mindset and strategic ambition.
There is also a third obstacle that rarely comes up in public discussions: the fear of failing at scale. When a human makes a mistake, the impact is localized. When an AI agent makes a mistake in a critical process, the impact can spread rapidly. This legitimate concern often paralyzes teams, who would rather not move forward than risk a visible failure. The problem is that this excessive caution carries a massive invisible cost: the cost of not evolving while the market around you transforms. Organizations that learn to balance speed and governance are the ones that manage to move forward without losing control.
The right questions for an AI agent strategy
The turning point for companies that want to break out of the experimentation cycle and move into real operations starts with a shift in perspective. In Charlene Li’s view, the first questions any organization should ask are: where can AI make a significant difference? Which strategic applications deserve focus at the enterprise level?
Li’s quote that sums it all up is simple and powerful: you do not need an AI strategy. You already have a strategy — it is called your business. So figure out how to use AI to support that strategy.
This approach marks the difference between organizations that will scale an agent-based workforce and those still debating tools. Instead of asking which language model is the most advanced or which platform has the most integrations, the questions that actually move the needle are different:
- Which processes in my business, if accelerated or automated, would generate the most impact on results?
- Where does my team spend the most time on repetitive tasks that could be delegated to an agent?
- Which routine decisions could be made based on data, freeing people up for more complex decisions?
- What is the minimum clarity and context the system needs to deliver useful work?
These questions put business strategy at the center of the discussion, not technology.
Who is in the room when decisions are made?
Another critical point raised by Li is about who is deciding how the agent-based workforce will be deployed. Most companies still follow the traditional IT procurement process, where CIOs, CDOs, and CTOs hold the primary authority. Less than a third of Chief AI Officers currently report to business leadership.
Li’s challenge is spot-on: where are the business people in this conversation? The business leaders — those who would have clarity of purpose and direction — are absent. When billions of agents enter the workforce, decisions about what they do and what authority they carry will be critical. Treating AI as just a technology challenge, rather than a business execution strategy, is a recipe for falling behind.
Two outdated reflexes that stall progress
Charlene Li identifies two outdated thinking patterns that still dominate many organizations and need to be dropped for the transition to AI agents to actually work.
Waiting for perfect data before acting
The first is the idea that you need perfect data before putting AI to work. According to Li, data is always going to be messy because it is always being created. Leading companies are not waiting for immaculate data. They are asking what is the minimum clarity and context needed for the system to deliver useful work.
This is why infrastructure matters so much. As Dallas puts it: most organizations are trying to run autonomous agents on foundations designed for manual queries. That is like dispatching self-driving trucks on a road system that still uses hand-painted signs.
Automating broken processes
The second outdated reflex is trying to automate what already exists without questioning whether it even makes sense. Li warns very clearly: do not automate a process that is fundamentally broken. Instead, think about how to rethink that process now that you have AI at your disposal. Blindly automating an existing process does not necessarily help — it can actually make things worse.
True innovation does not come from automation layered on top of the old system. It comes from rethinking whether that system still makes sense in the first place. This mindset shift is what separates AI projects that generate real results from projects that simply digitize existing red tape.
Real-world case: how Konecta transformed its operation with AI
A concrete example of this approach comes from Charlene Li’s book, Winning with AI. Konecta is a call center company with 130,000 employees, and its starting point is revealing. According to Li, the company’s problem was not having too many people. The problem was that it did not have people with the skill level it needed.
Konecta began deploying AI to augment its workforce capacity and automate back-office processes. The results in the first year were impressive:
- 25 to 30% reduction in complaint processing time
- 85 to 90% drop in error rates
- 40% decrease in training time for new employees
On top of that, the company created new revenue streams, such as an AI-powered service that analyzes 100% of defaulted loan documentation for law firms. And here is the data point that breaks the narrative that AI only exists to cut jobs: instead of reducing headcount, Konecta plans to grow its workforce by 5% by 2028.
Better-equipped people doing higher-value work, with AI handling the repetitive tasks nobody should have been doing in the first place. That is the essence of digital transformation done right.
The future of the relationship between humans and AI agents
Li makes an interesting prediction: by the end of 2026, the prevailing sentiment will be something like — I cannot believe I ever thought AI was going to take my job. I cannot imagine doing my work without AI.
But that reality will only exist inside organizations that committed early enough to stop treating AI as a side project, that asked the right questions from the start, and that aligned people, processes, and culture around this shift.
For everyone else, the agent-based workforce will not feel like empowerment. It will feel like watching a competitor operating in a different century. And that is a feeling no company wants to experience. 😬
What separates those who lead from those who watch
Companies at the forefront of AI agent adoption share some characteristics that go beyond available budget or the size of their tech team.
The first is speed of experimentation with discipline. They test fast, measure rigorously, and scale what works without hesitation. There is no twelve-layer approval committee for every new initiative. There is an agile process with clear success criteria and the freedom to iterate. This creates an organizational learning curve that becomes increasingly difficult to replicate over time.
The second characteristic is integrating AI into executive-level strategy. In leading companies, the topic is not confined to the technology department. It shows up in board meetings, in growth targets, in capital allocation decisions. This happens because the leaders of these organizations understood that AI agents are not an operational efficiency tool. They are a vector for competitive advantage. And competitive advantage is always a senior leadership conversation.
The third characteristic is building a data culture. AI agents are only as good as the data that feeds their operation. Companies that invested years in data quality, systems integration, and an analytical culture are now reaping the benefits of that investment in ways others simply cannot replicate quickly. This reinforces the idea that real digital transformation does not start with adopting a new technology. It starts much earlier, in how the organization collects, organizes, and uses its information to make decisions.
As Dallas emphasizes: the pilot phase is over. In the high-stakes shift to agent-based AI, a wait-and-see attitude stops being cautious and becomes catastrophic. The 13% that lead are winning because they recognized early that sovereignty over data and AI is the critical infrastructure of the modern enterprise. If you do not control your data and your AI, you do not control your business. 🎯
The picture taking shape is clear: the distance between companies that adopt AI agents strategically and those that stay on the sidelines will grow at an accelerating pace over the coming years. Not because the technologies are too complex to adopt, but because the window of time to build real competitive advantage with them is closing. Billions of agents are coming. The only question left is whether they will be working for you or against you. 💡
