AI Risks Come From Outdated Workflows, Not Tools: Clip

The AI integration through business workflows is redesigning human capital and their technical training. To overcome possible challenges, Victor Velázquez, Chief People and AI Enablement Officer, Clip, proposes a model centered on digital literacy, native security, and trust to mitigate operational risks and optimize corporate competitiveness for stakeholders across the financial technology sector.

The implementation of advanced technology remains ineffective if operational structures continue to follow analog models, said Victor Velázquez, Chief People and AI Enablement Officer, Clip, during Palo Alto Networks Ignite Tour 2026. He explained that the risk of failure does not reside in the capacities of the tool, but in the inertia of current operating methods, and presented a framework for AI integration that prioritizes organizational design and human behavior over technical infrastructure.

"AI is not what will make us fail. What will make us fail is the way we work and the way each individual adopts AI," says Velázquez. This premise justifies a profound transformation in governance to align business velocity with security protocols.

The contemporary business environment faces a disconnect between organizational design and the operational reality of AI. Velázquez notes that departments such as human resources and talent management have been historically perceived as administrative areas with low technical complexity. However, the management of human unpredictability in high-technology environments represents a sophisticated challenge.

The phenomenon known as Shadow AI emerges as a direct response to these organizational inefficiencies. When official corporate channels are slow or ambiguous, employees utilize external tools without compliance supervision. This behavior does not typically stem from malicious intent, but from a necessity to meet business objectives within timeframes that traditional frameworks do not permit.

In this context, the relevance of this development lies in a paradigm shift: cybersecurity must transition from focusing exclusively on technical perimeters, such as the cloud, firewalls, and networks, to the protection and design of human behavior within the workflow.

Technical Details and Organizational Framework 

The strategy implemented by Clip for AI enablement is based on a technical framework comprising three primary pillars: literacy, workflow redesign, and trust management. According to data reported by the corporation for 1Q26, these initiatives aim to bridge the gap between how organizations were designed to work and how AI actually operates.

Education serves as the first filter for institutional security. The organization has executed a comprehensive training program in which 470 collaborators, about half of its workforce, received specialized instruction in approved AI tools. “This process represents a significant investment of 12 hours of training per individual, totaling more than 5,600 hours dedicated to understanding usage protocols, data filtering, and prompt architecture,”  says Velázquez.

Furthermore, over 200 leaders participated in mindset-shift programs to facilitate the operational transition. This literacy goes beyond basic tool usage; it involves teaching staff to communicate and write for an AI-augmented environment. Velázquez says that just as employees are taught corporate communication standards, they must now master technical interaction with automated systems to prevent data leaks and optimize output quality.

Workflow Redesign and Native Security 

Instead of superimposing AI tools on top of obsolete processes, the corporation advocates for native redesign. In the department of fraud and risk, for example, a traditional model of four stages was restructured into a five-step flow. In this configuration, AI is not an additive component but an intrinsic part of the detection architecture. This allows security to be implemented by design rather than as an external layer.

A coordinated "tech stack" is fundamental to this pillar. Rather than allowing different departments to acquire fragmented solutions, the corporation centralizes technology. For example, the same conversational AI tool used by the department of customer service for answering calls can be programmed for the sales division to address inquiries. This centralized approach ensures observability and maintains strict security controls across all business units.

The transition from a Zero Trust model to one of institutional trustworthiness requires clear technical communication. To mitigate the risks associated with Shadow AI, the organization provides secure opportunities for experimentation. During 1H26, collaborators created nearly 500 AI agents within a controlled environment. Of those, 50 are currently fully functional and used at least once a week in regular business processes.

This approach allows talent to explore technological capabilities without compromising the integrity of corporate data. Velázquez says that when organizations treat employees as if they lack the intelligence to understand complex systems, they inadvertently encourage them to work outside the permitted perimeter. By providing "sandboxes" or safe environments, the corporation captures innovation while maintaining oversight.

The adoption of AI at Clip is formalized through key performance indicators. During the current period, 90% of the workforce has performance objectives directly linked to AI, ranging from theoretical understanding to the implementation of process improvements. 

Moving forward, Velázquez highlights that success of a corporation will depend on creating structures where technology and human factors operate in a symbiotic manner. Future leaders’ role will be in constructing ‘organizational infrastructures’ capable of operating at the speed of AI.