Rebuilding Organizations Around AI Agents Creates Sustainable Competitive Advantage

In a world of rapid technological advancement, particularly in AI, the conventional wisdom of simply augmenting human capabilities with new tools is proving insufficient. This conversation with Carlos García Ottati, founder and CEO of Kavak, reveals a more profound truth: true transformation requires not just adopting AI, but fundamentally rebuilding organizational structures and processes around it. The hidden consequence? Companies that embrace this radical shift, even through painful transitions, can forge deep, sustainable competitive advantages. This analysis is crucial for leaders in emerging markets and beyond, offering a blueprint for navigating technological disruption and building truly resilient, future-proof businesses.

The narrative of technological adoption often follows a predictable path: introduce a new tool, expect immediate efficiency gains, and watch productivity soar. However, Carlos García Ottati, founder and CEO of Kavak, presents a starkly different reality, particularly when it comes to AI. His experience building Kavak, Latin America's largest online used car marketplace, underscores a critical insight: simply layering "copilot" tools onto existing structures is a recipe for failure. The real opportunity, and the source of lasting competitive advantage, lies in a more radical approach -- replacing human-led processes with AI agents, a transition that, while fraught with difficulty, unlocks unprecedented scalability and efficiency.

Kavak's journey is a masterclass in consequence mapping, demonstrating how initial decisions ripple through an organization and its market. From its inception, Kavak was built not just as an e-commerce platform, but as a complex ecosystem addressing fundamental market failures in Latin America. This meant building four distinct businesses under one roof: e-commerce, reconditioning and warranty, financing, and logistics. This foundational complexity, born from necessity in emerging markets where infrastructure is lacking, set the stage for Kavak's later AI pivot. As Ottati explains, the sheer volume of edge cases--unique customer situations requiring intricate problem-solving--made human-led processes unsustainable at scale.

"When you're building AI, the first thing you need to build is the brakes of the system. It's the understanding of where you're going to deploy it and how you're going to deploy it."

This "brakes" analogy is crucial. Before deploying AI, Kavak focused on understanding its users and their interactions -- building a comprehensive ontology of its operations. This data-rich foundation allowed them to identify critical junctures where AI could have the most impact. The initial attempt to provide AI copilot tools to employees, a common strategy, met with dismal adoption rates. This failure was a pivotal moment, revealing that the problem wasn't a lack of tools, but a fundamental mismatch between human workflows and AI capabilities. The subsequent pivot to deploying AI agents directly in front of critical business functions, rather than as assistive tools, was a painful but necessary step.

The transition was far from smooth. Ottati candidly admits it was "very painful," leading to a year of flat growth as the company restructured. This period highlights the "discomfort now, advantage later" principle. While conventional wisdom might dictate a retreat or a slower, more incremental approach, Kavak leaned into the difficulty. They prioritized solving the hardest problems first -- underwriting loans and handling car breakdowns -- rather than the more obvious customer service issues. This focus on core, complex functions, where AI could demonstrate a significant leap in capability, laid the groundwork for future success. The result? Agents began performing at parity, and eventually, 1.5 times better than their best human counterparts in specific funnels.

"We made the typical mistake that everybody's making that is, you know, we build these co-pilot tools to give them to our teams so they could just use them and just provide a better customer experience or just have the information to provide, you know, a better solution to our customers. And we did just that, you know, and in late 2022, early 2023, we spent a lot of time building these tools for everybody in our organization. And we gave them to them, and we realized very quickly they did not adopt them."

This emphasis on AI agents, rather than assistive tools, is where Kavak builds its moat. By replacing human processes with AI, they not only achieve scalability but also create a system that learns and improves continuously. Ottati’s philosophy of building for "ChatGPT 7" rather than the current iteration underscores a forward-looking approach. This isn't about optimizing for today's technology; it's about building an architecture that can harness future AI advancements. This proactive stance, anticipating the evolution of AI models, positions Kavak to capitalize on breakthroughs before competitors even recognize their potential.

The broader implications of this strategy are immense. In emerging markets, where building foundational infrastructure is a constant challenge, AI agents can leapfrog traditional development hurdles. Instead of spending years building payment rails or logistics networks, AI can be trained to navigate these complexities, offering a more efficient path to scale. Furthermore, the transition to AI agents forces a rigorous re-evaluation of organizational structure and human roles. As Ottati describes his annual "firing himself" exercise, it becomes clear that leadership itself must evolve. The focus shifts from managing people to orchestrating AI systems, requiring a different set of skills and a willingness to let go of outdated paradigms.

"We had a year. So we were growing like the business was growing 300%. Like 2022 was growing 100%, and 2023 we were flat. You know, so the reason behind that was we were just restructuring everything below the line."

The year of flat growth, while daunting, was a necessary period of rebuilding. It allowed Kavak to emerge not just leaner, but fundamentally more capable. By embracing the discomfort of this transition, they have positioned themselves for sustained, accelerated growth. The market implosion of 2022, while devastating for many, provided a forced catalyst for this transformation, making the "streaming" transition inevitable. Companies that fail to make this shift, Ottati warns, face a "world of slow death and pain." The advantage lies with those who, like Kavak, recognize that true innovation requires not just adopting new tools, but fundamentally reimagining the business around them.

Key Action Items:

• Immediate Actions (Next 1-3 Months):

  • Map Critical Customer Journeys: Identify the 2-3 most complex or friction-filled customer touchpoints within your business.
  • Assess Current AI Adoption: Evaluate the effectiveness of existing AI tools. Are they being adopted, or are they merely "copilots" that employees ignore?
  • Investigate Data Ontology: Begin building a comprehensive understanding of your customer data and operational processes. What is the "highway of information" within your organization?
  • Pilot AI Agent Replacements: Select one critical, non-customer-facing process (e.g., internal data analysis, report generation) and pilot an AI agent to perform the task, rather than just assist.

• Medium-Term Investments (Next 6-18 Months):

  • Develop AI Agent Skills: Focus on training AI agents to handle specific, complex tasks that currently strain human capacity, prioritizing areas with high edge-case volume.
  • Build an Orchestration Layer: Design systems that allow AI agents to communicate and collaborate, moving beyond siloed AI tools.
  • Rethink Organizational Structure: Begin planning for how human roles will evolve as AI agents take on more core functions. This may involve reskilling or redefining responsibilities.
  • Embrace the "Year of Flat Growth": Budget for and communicate a potential period of flat growth or even temporary decline as core processes are re-engineered around AI. This requires strong leadership conviction.

• Longer-Term Investments (18+ Months):

  • Develop for Future AI Capabilities: Design AI systems with an eye toward anticipating future model advancements, rather than optimizing for current capabilities.
  • Establish Continuous Learning Loops: Ensure your AI systems are designed to learn from failures and successes, creating a compounding advantage over time.
  • Regularly Re-evaluate Leadership Role: Implement a process for self-assessment, questioning whether current leadership is best suited for the next phase of AI-driven transformation.