As generative AI becomes more widespread, many employees are significantly boosting their personal productivity with AI tools. However, the overall value and output of companies have not grown in tandem. AI data analytics company Hebbia founder George Sivulka recently published a lengthy article titled “High Personal Output Doesn’t Equate to a High-Performing Company,” pointing out that the problem isn’t the technology itself but that organizations haven’t redesigned themselves to incorporate AI. He introduces the concept of “Institutional AI,” believing that future corporate competitiveness will depend on how deeply AI is embedded into processes, decision-making, and governance structures—not just as a tool for individual efficiency.
AI Productivity Paradox: Personal Efficiency Rises, Company Value Remains Stagnant
In recent years, generative AI has rapidly gained popularity, with tools like ChatGPT and Claude widely used by employees for writing, coding, and data analysis. Media and creators emphasize that AI can multiply individual work efficiency by several times or even ten times.
However, Sivulka points out that the overall output and value of companies have not increased proportionally: “In other words, the productivity gains brought by AI haven’t truly translated into business value at the organizational level.”
He believes the reason is that most companies currently allow employees to use AI tools independently, without changing organizational structures, processes, or decision-making mechanisms. Therefore, “high individual efficiency does not equal high organizational efficiency.”
(Does AI-assisted coding cause problems? Amazon experiences system failures four times in a week, prompting urgent review meetings)
Looking at the Electrification of Factories in History as a Model for AI-Driven Organizational Transformation
Sivulka draws an analogy from the late 19th-century industrial revolution. When electricity began replacing steam engines, many textile factories simply replaced their steam power with electric motors, but their factory structures and production processes remained unchanged. As a result, over nearly 30 years, factory capacity saw little significant improvement.
It wasn’t until the 1920s, when factories underwent comprehensive redesign—introducing assembly lines, equipping each machine with its own motor, and re-planning workflows—that electricity truly led to substantial productivity growth.
Sivulka believes that current AI development is at a similar stage: companies are just “replacing the motor” but haven’t yet “redesigned the entire factory.”
From “Personal AI” to “Organizational AI”: How to Maximize Efficiency?
He refers to current AI models and applications as “Personal AI,” and introduces the concept of “Institutional AI,” explaining the difference between the two.
Personal AI: Tools to Enhance Individual Efficiency
Sivulka states that most current AI applications are “Personal AI,” limited to individual productivity tools. For example, employees use AI to write reports, organize data, or generate presentations:
These tools can indeed improve personal efficiency but often lack unified processes and collaboration mechanisms, leading to AI-generated content that can’t be integrated and potentially increasing internal chaos and information noise.
Organizational AI: Intelligent Systems Embedded in Organizational Processes
He proposes another form called “Institutional AI.” This type of AI isn’t just a single tool but is deeply embedded into a company’s decision-making, processes, and governance structures, helping the organization generate real value.
In this framework, AI can play various roles, such as analyzing risks, coordinating information across departments, or proactively discovering new business opportunities.
The Seven Pillars of “Organizational Intelligence”: The True Blueprint for Corporate AI
Sivulka further outlines seven core elements of “Organizational Intelligence,” believing these capabilities will form the foundation of future enterprise AI systems.
Coordination: Prevent Organizational Chaos Caused by AI
First, if each employee uses AI tools independently, the outputs and processes may conflict. One key task of organizational AI is to establish collaboration and management mechanisms, enabling humans and AI agents to work together under clear division of labor.
Personal AI causes organizational chaos; organizational AI can unify workflows and assist task allocation. Signal: Finding Value in AI-generated Noise
Generative AI drastically reduces the cost of content creation, leading to a flood of variable-quality information. One important function of organizational AI is to identify and filter out truly valuable “signals” within vast data and AI-generated content.
Personal AI cannot filter information effectively; organizational AI can identify signals. Objectivity: Avoid AI Reinforcing User Biases
Sivulka notes that many current AI models tend to cater to user opinions, potentially exacerbating internal cognitive biases. Future enterprise AI systems need to act more objectively, akin to auditors or supervisors, questioning decisions and highlighting potential risks.
Personal AI may reinforce “echo chambers” and biases; organizational AI can focus on the truth. Competitive Advantage: Combining General Models with Vertical Applications
He cites the concept of the innovator’s dilemma, pointing out that current general large models provide foundational capabilities, but vertical applications—still in early stages—often offer significant marginal benefits. For example, image generation platform Midjourney or speech AI company ElevenLabs build technological moats by focusing on specific domains.
He believes that for companies, combining general AI with organizational-specific professional AI is the key to establishing a competitive advantage.
Outcome-Oriented: From Cost Savings to Revenue Generation
Sivulka emphasizes that many current AI products focus on “saving time or labor costs,” but what companies truly care about is “whether it can increase revenue.” Therefore, future AI systems should demonstrate value by discovering new business opportunities or boosting sales:
A truly defensible organizational AI must directly drive revenue. For example, identifying the only worthwhile M&A target from thousands of potential candidates, rather than just enabling analysts to build financial models faster.
Implementation Capability: Integrating AI into Business Processes
The practical deployment of AI often requires redesigning business workflows and management systems. Sivulka points out that companies like Palantir (PLTR) are gaining market attention partly because they help organizations integrate AI systems with existing operations and drive internal change management.
Proactive Action: Moving Beyond Human Prompts
Most current AI systems require human input to operate. Sivulka believes that a more important future capability is for AI to proactively monitor data, detect anomalies, and issue early warnings. For example, before a fund manager even opens financial reports, the AI might have already identified deteriorating operational cash flow in an invested company and automatically issued a risk alert based on loan covenants.
AI Era Corporate Competition: Who Can “Rebuild the Factory” First Will Win
In conclusion, Sivulka emphasizes that while personal AI tools will remain the main entry point for companies to access AI, truly gaining a competitive edge will require integrating organizational AI systems. He predicts that future enterprises will likely use both general AI assistants and specially designed organizational AI systems—where the former boosts employee productivity, and the latter consolidates information, drives decision-making, and creates business value.
He ends with a historical analogy from the industrial revolution, stressing that companies face the same challenge today:
Those factories that first adopted electrification ultimately lost to those that redesigned their production lines. We now have electricity; it’s time to rebuild the factory.
This article, “AI Increasing Employee Productivity by 10x Doesn’t Equal 10x Company Value: Where Has the Productivity Gone?” originally appeared on Chain News ABMedia.
