OpenAI: Frontier Firms Generate 8.3x More Tokens
OpenAI published a post on its own site titled "How AI-native companies turn workflows into operating capability", filed under its AI Adoption section and dated September 1, 2026. Everything below comes from that post.
Transcript
OpenAI published a post today about three startups building agents into familiar work.
OpenAI says frontier firms, the top ten percent by AI usage, generate eight point three times as many output tokens per active user as typical firms.
OpenAI's post says first day onboarding at Basis now takes thirty minutes instead of two hours with Codex and a reusable skill.
Clay says one engineer gave each account a subagent that updates a deal folder overnight, saving about an hour of inbox triage each night.
At Exa Labs, OpenAI says Codex opens pull requests and runs tests before human review. Every figure here comes from OpenAI, not independent testing.
Inboxsmith builds AI phone and email agents for small business. Please like and subscribe for more news.
Sources
Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:
- Basis, Clay, and Exa Labs use agents for onboarding, account management, and developer integrations.(OpenAI's official post, How AI-native companies turn workflows into operating capability, published September 1, 2026)
- Frontier firms (those with the top 10% of AI usage) now generate 8.3× as many output tokens per active user as typical firms, up from 2.6× in January.(OpenAI's official post, How AI-native companies turn workflows into operating capability, published September 1, 2026)
- At Basis, which builds AI agents for accounting firms, first-day onboarding now takes 30 minutes instead of two hours, giving HR more time for culture and support.(OpenAI's official post, How AI-native companies turn workflows into operating capability, published September 1, 2026)
- On day one, employees receive immediate access to Codex and a company-specific onboarding skill (a reusable set of instructions and resources for a specific workflow).(OpenAI's official post, How AI-native companies turn workflows into operating capability, published September 1, 2026)
- Each subagent reviews primary sources and updates its deal folder overnight.(OpenAI's official post, How AI-native companies turn workflows into operating capability, published September 1, 2026)
- The workflow saves her roughly an hour of inbox triage each night, according to Clay.(OpenAI's official post, How AI-native companies turn workflows into operating capability, published September 1, 2026)
- Codex now monitors for high-priority integration opportunities, gathers the relevant context, creates pull requests, runs tests, and prepares weekly updates using sources such as Slack and Notion.(OpenAI's official post, How AI-native companies turn workflows into operating capability, published September 1, 2026)
- Exa turned that sequence into a defined workflow for Codex, with clear priorities, access to the necessary sources, and human review before anything ships.(OpenAI's official post, How AI-native companies turn workflows into operating capability, published September 1, 2026)