How frontier enterprises are building an AI advantage
Observe that frontier firms use 3.5x more AI per worker and 16x more Codex messages; consider measuring depth and agentic workflows in your organization.
Measure your firm's AI depth and agentic usage to benchmark against frontier firms.
Summary
OpenAI’s new B2B Signals report shows that frontier firms now use 3.5 times as much AI intelligence per worker compared to typical firms, a jump from 2× in April 2025. The gap is driven largely by depth, not volume, with frontier teams sending 16 times as many Codex messages per worker, and leveraging advanced tools like ChatGPT Agent, Apps in ChatGPT, Deep Research, and GPTs. Cisco’s engineering org reports a 20% reduction in build times, 1,500+ engineering hours saved per month, and a 10–15× increase in defect‑resolution throughput when Codex is treated as a team member. In production, AI is moving into specialized workflows such as in‑app assistants, coding tools, and customer support, with IT and security teams focusing on procedural guidance and finance teams using AI for analysis. Travelers Insurance’s AI Claim Assistant, built with OpenAI, is projected to handle 100,000 first‑notice‑of‑loss calls in its first year, illustrating the scale of delegated work. The report emphasizes that the frontier advantage is not a fixed gap; organizations can move toward it by measuring depth, building governance, scaling enablement, and shifting from chat to delegated agentic workflows.
Key changes
- Frontier firms use 3.5× more AI intelligence per worker than typical firms (up from 2× in April 2025).
- Frontier firms send 16× as many Codex messages per worker compared to typical firms.
- Cisco’s use of Codex cuts build times by ~20%, saves 1,500+ engineering hours per month, and boosts defect‑resolution throughput by 10–15×.
- Travelers Insurance’s AI Claim Assistant is expected to handle ~100,000 first‑notice‑of‑loss calls in its first year.
- AI is increasingly embedded in specialized workflows: in‑app assistants, coding tools, customer support, IT/security procedural guidance, finance analysis.
- The B2B Signals report focuses on depth of use, agentic workflows, and emerging patterns across industries and functions.