Enterprise AI is entering a new phase.
For the past few years, businesses have largely treated AI as an assistant: something employees use to brainstorm, summarize documents, write emails or answer questions.
That model is changing.
The emerging enterprise model is AI that actually executes work.
From Asking to Doing
OpenAI's latest enterprise research shows how quickly that transition is happening.
The company describes a shift from assistants that help employees think through tasks to agents that can use tools, access company context and complete multi-step workflows.
The difference is important.
An employee might previously have asked AI how to prepare a presentation.
An agent can potentially gather information, work through relevant files, create the presentation and return it for human review.
AI becomes less like a search box and more like a digital worker.
The AI Leaders Are Pulling Away
The research also suggests that enterprise AI adoption is becoming increasingly uneven.
OpenAI identifies the top 10% of companies by AI usage as "frontier firms."
In June, those companies generated 8.3 times as many output tokens per active user as typical firms.
That gap was just 2.6 times in January.
In other words, companies are not simply adopting AI at different speeds.
The leading organizations are beginning to use it at a fundamentally deeper level.
The advantage appears to come from connecting AI to the systems and workflows employees already use.
Agents Are Moving Beyond Engineering
Software development was an early proving ground for AI agents.
But enterprise adoption is now spreading across other knowledge-work functions.
According to OpenAI, weekly active enterprise Codex users have grown since February by:
108× in legal
41× in sales
41× in recruiting
26× in marketing
5× in engineering
That suggests the agent opportunity is much broader than coding.
A sales agent can work with customer information.
A recruiting agent can help process candidates.
A legal workflow can involve large volumes of documents.
Marketing teams can use agents for research, content and campaign workflows.
The common denominator is not the industry.
It is repeatable knowledge work.
The Workforce Gap Could Change
One of the more interesting findings concerns seniority.
OpenAI says its analysis of millions of conversations found that early-career employees use AI more heavily than executives.
Six months after adoption, early-career employees sent an average of 13 more messages per week than executives.
That could have important implications for companies.
The employees gaining the most practical experience with AI today may not necessarily be the people with the most seniority.
Organizations therefore have an opportunity to identify their strongest AI users and turn their individual workflows into company-wide practices.
What This Means for Miami
This shift could be particularly significant for Miami's growing technology and business ecosystem.
South Florida has a large concentration of companies in sectors such as finance, real estate, healthcare, hospitality, logistics and professional services - industries containing enormous amounts of repetitive knowledge work.
The opportunity isn't necessarily to replace entire departments with AI.
It is to identify the processes inside those businesses where an agent can reliably perform several steps that currently require employees to move between spreadsheets, emails, documents and software systems.
That creates a new competitive question for Miami companies.
Who is using AI and who is actually redesigning their business around it?
The companies that answer the second question first could gain a much larger advantage than those simply giving employees access to another chatbot.
Reporting Source: This article builds upon research published by OpenAI and adds analysis of what the development means for Miami and South Florida.
