The gap between AI leaders and laggards in the enterprise is widening. Frontier firms, those at the forefront of AI adoption, are now leveraging 3.5 times more AI intelligence per worker compared to typical organizations, a substantial increase from just a year ago when the ratio was 2x. This isn't merely about sending more messages; the real advantage lies in the depth and complexity of AI use.
New research from OpenAI News, detailed in their B2B Signals report, reveals that message volume accounts for only 36% of this frontier advantage. The majority stems from richer, more intricate AI interactions where employees are delegating complex work, providing more context, and expecting more substantive outputs.
The Deepening AI Divide
This shift signifies a move beyond basic AI assistance to AI as a core component of execution. Frontier firms are using AI to tackle challenging tasks, transforming it from a question-answering tool into a work-execution engine.
The adoption of advanced tools, particularly those enabling agentic workflows, marks a significant marker of maturity. These are systems where AI can use tools, process files, and complete longer-horizon tasks. Frontier firms are 16 times more likely to use tools like Codex, which assists with coding, indicating a strong push towards delegating complex technical work.
Organizations like Cisco have seen substantial gains, reducing build times by approximately 20% and saving over 1,500 engineering hours monthly by integrating Codex as a collaborative team member. This highlights the potential for AI agents to fundamentally redesign workflows.
