OpenAI has introduced B2B Signals, a recurring research report that uses de-identified, aggregated enterprise usage data from OpenAI products to examine why business AI adoption is moving beyond tool access toward deeper workflow use.

For many companies, the first phase of AI adoption was about access: which employees had tools, how often they used them, and whether experimentation was spreading. OpenAI’s first B2B Signals report raises a more difficult question for enterprise leaders: whether AI is changing how work gets planned, delegated, completed, and measured.

OpenAI says frontier firms, meaning companies at the leading edge of enterprise AI adoption, now use 3.5 times as much intelligence per worker as typical firms, up from 2 times in April 2025. The report says message volume explains only 36% of that gap, with most of the difference coming from deeper usage rather than more frequent prompting alone.

In short, B2B Signals is OpenAI’s attempt to measure enterprise AI maturity by looking beyond access. The report argues that leading firms are pulling ahead because workers are using AI for more complex tasks, richer context, delegated workflows, and production use cases across the business.

Enterprise AI maturity refers to how deeply an organization integrates AI into real work, not simply whether employees have access to AI tools.

OpenAI’s first B2B Signals report defines enterprise AI maturity as a question of workflow depth, showing that access to AI tools is no longer the clearest measure of adoption.

OpenAI introduced B2B Signals as a business extension of OpenAI Signals, built to track how AI use develops inside companies over time.

The report is based on de-identified, aggregated enterprise usage data from OpenAI products. OpenAI says the analysis examines how deeply AI is used inside firms, which tools and tasks are associated with frontier adoption, and where business use cases are growing across industries, products, and functions.

OpenAI also included a privacy note with the report. The company says message content was classified using automated systems, and no OpenAI employee reviewed individual enterprise, business, or API customer data as part of the analysis.

That privacy detail is important for how readers should interpret the report. B2B Signals is not a company-by-company customer study. It is a usage-pattern report built from aggregated product activity across OpenAI’s enterprise ecosystem. That gives OpenAI a way to analyze how businesses are adopting its tools while giving enterprise leaders a benchmark for comparing access, workflow depth, agentic tool use, and production deployment.

The first release focuses on depth of use, agentic workflows, and patterns across industries and business functions. OpenAI says future updates will track progress on these measures as enterprise AI adoption continues to develop.

The report’s central argument is simple: access is no longer enough to separate leading AI adopters from everyone else.

OpenAI says frontier firms now generate 3.5 times as much AI output per worker as typical firms, compared with 2 times in April 2025. The report uses tokens generated to estimate how much work employees are asking AI systems to perform. OpenAI notes that tokens are not a direct measure of business value, but uses higher AI output as one sign that some companies are starting to use AI to change real workflows.

The report also says message volume explains only 36% of the gap between frontier firms and typical firms. OpenAI’s interpretation is that the rest of the difference comes from how employees are using AI, not just how often they are using it. Workers at frontier firms are asking AI systems to take on more complex work, process richer context, and produce more substantive outputs.

OpenAI summarizes the difference this way: typical firms are using AI to answer questions, while frontier firms are using AI to execute complex work.

The key point: enterprise AI maturity is becoming harder to measure through access alone. A company can provide AI tools without changing how work is planned, delegated, reviewed, or completed. OpenAI’s report argues that the stronger indicator is whether AI is becoming part of the workflow itself.

For business leaders, this creates a measurement problem. Seat counts and adoption dashboards may show whether employees can access AI. Message volume may show whether they are trying it. But those numbers may not show whether teams are using AI to complete harder work, reduce bottlenecks, improve handoffs, or build new operating habits, which are the kinds of changes that determine whether AI becomes part of the business rather than just another tool employees can access.

That is where the report becomes more useful than a basic adoption update. It asks leaders to look at the quality of AI use, not only the quantity.

OpenAI says the largest gap between frontier firms and typical firms appears in advanced and agentic tools, where AI starts taking on more delegated work.

According to OpenAI, Codex shows the largest difference between frontier firms and typical firms, with frontier firms sending 16 times as many messages per worker as typical firms. The report says ChatGPT Agent, Apps in ChatGPT, Deep Research, and GPTs show similar directional patterns, suggesting that frontier firms are adopting tools that help workers code, delegate multi-step tasks, apply company context, and conduct more complex research.

OpenAI connects this pattern to the rise of delegated AI work. As AI systems become more capable of using tools, working across files and codebases, and completing longer tasks, companies need to build the habits and systems required to delegate meaningful work to AI agents.

This is where adoption becomes less about individual productivity and more about organizational behavior. Asking AI for help with a paragraph, summary, or answer is one kind of use. Asking AI to work across a codebase, support a multi-step research task, or complete part of a production workflow requires a different level of trust, context, review, and process design.

OpenAI points to Cisco as one example. According to the report, Cisco uses Codex to speed up complex software work across a large enterprise engineering organization. In production workflows, OpenAI says Codex helped reduce build times by about 20%, save 1,500-plus engineering hours per month, and increase defect-resolution throughput by 10 to 15 times. OpenAI quotes Cisco’s team as saying the biggest gains came when they treated Codex as “part of the team.”

That example gives the report its clearest practical meaning. The difference is not simply that Cisco used an AI tool. The difference is that Cisco used Codex inside production workflows where engineering work, review cycles, and defect resolution already mattered.

The report does not show whether every enterprise will see Cisco-like results. But it does show the kind of integration OpenAI wants business leaders to study: AI that becomes part of how teams execute work, not just a faster interface for asking questions.

The Codex gap may also reflect an important enterprise advantage rather than a universal measure of AI maturity. Large companies are more likely to have internal engineering teams, existing codebases, technical review processes, and production workflows where AI coding tools can be applied. For smaller businesses, deeper AI adoption may look different. Access to an advanced coding agent does not automatically create software capability if a company lacks developers, technical oversight, or systems where the tool can be safely used.

In those cases, AI maturity may be better measured by how effectively a business uses AI in customer service, marketing, operations, documentation, or decision support rather than by whether it can deploy coding agents.

OpenAI’s report also shows that enterprise adoption is not developing in one uniform pattern.

The report says companies are deploying API use cases across in-app assistants, coding and developer tools, and customer support. These are areas where AI can become part of products, services, internal systems, and customer-facing workflows.

OpenAI also says AI use is broadest in writing and communication, but function-specific use is growing. The report identifies several examples:

This pattern matters because maturity may look different from one department to another. A finance team does not need the same AI workflow as an engineering team. A customer support team does not measure value the same way a security team does. As AI use becomes more specialized, companies may need different adoption strategies for different teams instead of one generic AI rollout plan.

OpenAI also says there is no single AI adoption leaderboard. Some industries lead in broad ChatGPT adoption. Others show stronger Codex use, API intensity, or message intensity. That means organizations may have different entry points depending on their needs. Some may scale access. Others may deepen usage. Some may adopt agentic tools. Others may build AI directly into products and systems.

The report uses Travelers Insurance as an example of AI embedded into a business process. According to OpenAI, Travelers built an AI Claim Assistant with OpenAI that guides customers through first notice of loss, answers policy questions, gathers information needed to start a claim, and creates claims directly inside Travelers’ systems. Travelers expects the assistant to handle about 100,000 first notice of loss calls in its first year.

The Travelers example shows how enterprise AI can move beyond internal productivity into customer-facing operations. It also raises a more demanding question: whether companies have the data access, governance, system integration, user experience design, and oversight needed to place AI inside real business workflows.

OpenAI’s report describes a gap between frontier firms and typical firms, but it also cautions against treating that gap as permanent.

Without that clarification, readers could easily interpret the report as a simple winners-and-losers story. OpenAI’s argument is more useful than that. The report suggests some firms are building momentum by using AI more deeply, more broadly, and in more delegated workflows. But many organizations are still early in the process of moving from broad access to integrated use. The value of the frontier, in this context, is that it offers a view into which practices may help companies build AI capability over time.

OpenAI identifies education and learning as one of the clearest areas of frontier advantage. The report says the task-level advantage is largest in education and learning, suggesting that leading firms may use AI not only to complete work but also to build employee skills, habits, and confidence.

That finding may be especially relevant for leaders who think of AI adoption mainly as a software rollout. If employees do not know how to use AI well, access alone may produce uneven results. OpenAI’s report points toward enablement as a core part of maturity: employees need training, examples, governance, and permission to use AI in meaningful work.

OpenAI says organizations can move toward the frontier by measuring depth of use, building governance that enables production use, treating enablement as core infrastructure, identifying frontier teams and scaling their impact, and moving beyond chat toward delegated work with agents.

The open question is how much of this advantage comes from AI tools themselves, and how much comes from organizational readiness. Companies with stronger data systems, technical teams, governance practices, and leadership support may be better positioned to benefit from advanced AI tools. The report shows a usage gap, but it does not fully separate technology access from management quality, employee skill, workflow design, or existing digital readiness.

That limitation does not make the report less useful. It makes the interpretation more specific. B2B Signals is strongest as a view into enterprise usage patterns inside OpenAI’s ecosystem. It is not, by itself, proof that deeper AI use always creates lasting competitive advantage.

Q: What is OpenAI B2B Signals?
A: OpenAI B2B Signals is a recurring report that uses aggregated enterprise usage data from OpenAI products to examine how businesses are adopting AI. The first report focuses on AI workflow depth, agentic tools, and how companies are moving from basic AI access toward deeper operational use.

Q: What does OpenAI mean by frontier firms?
A: In this report, frontier firms means companies at the leading edge of enterprise AI adoption. OpenAI says these firms use AI more deeply, more broadly, and in more delegated workflows than typical firms.

Q: Why isn’t giving employees access to AI tools enough?
A: Access only shows whether employees can use AI. It does not show whether AI is changing how work gets planned, delegated, completed, or measured. OpenAI’s report suggests that enterprise AI maturity depends more on workflow depth than on seat counts or experimentation alone.

Q: Why does AI workflow depth matter?
A: AI workflow depth matters because it shows whether AI is being used for real work, not just lighter experimentation. OpenAI says frontier firms are asking AI systems to handle more complex work, process richer context, and produce more substantive outputs.

Q: Does higher AI use prove that a company is getting more business value?
A: No. OpenAI says tokens generated can help estimate how much work employees are asking AI systems to perform, but tokens are not a direct measure of business value. Higher AI output may show deeper use, but it does not prove lasting competitive advantage by itself.

Q: Why does Codex show such a large gap between frontier firms and typical firms?
A: OpenAI says Codex shows the largest gap, with frontier firms sending 16 times as many messages per worker as typical firms. That may show stronger adoption of agentic tools, but it may also reflect an enterprise advantage: large companies are more likely to have engineering teams, codebases, and technical review processes where coding agents can be used safely.

Q: What should companies measure instead of basic AI access?
A: Companies may need to measure workflow depth, agentic tool use, function-specific adoption, production deployment, and employee enablement. Those signals can show whether AI is becoming part of how work gets done, rather than just another tool employees can access.

OpenAI’s B2B Signals report turns enterprise AI adoption into a practical measurement question: not whether companies are using AI, but whether AI is becoming part of how important work gets done.

Key point: The report’s strongest interpretation is that the next enterprise AI gap may come from workflow depth. Companies that treat AI as a basic productivity tool may see limited gains, while companies that redesign work around AI may build stronger operating habits over time.

Who should care: Business leaders, CIOs, CTOs, AI program owners, operations leaders, engineering teams, and strategy teams should pay attention because the report challenges simple adoption metrics. Access, seat counts, and message volume may not show whether AI is changing workflows, delegation patterns, employee capability, or production systems. Smaller businesses should also read the report carefully, because enterprise AI maturity may look different for companies without internal engineering teams, large codebases, or technical review processes.

Why this matters now: Many companies have already provided AI access and encouraged experimentation. OpenAI’s report suggests the harder phase is now emerging: teaching teams how to use AI for complex work, agentic delegation, function-specific tasks, and customer-facing systems.

What decision this affects: The report affects whether organizations treat AI adoption as a software rollout or an operating model decision. Leaders may need to invest more in workflow redesign, employee enablement, governance, agentic tools, and measurement systems that track depth of use rather than activity alone.

In short, B2B Signals does not prove that deeper AI use automatically creates lasting advantage. It does show that OpenAI sees enterprise AI maturity moving beyond access and experimentation toward workflow integration, delegated tasks, and production use. For business leaders, the practical question is whether their organization is merely using AI, or learning how to work differently because of it.

The companies most likely to benefit from AI may not be the ones with the most tools, but the ones that build the clearest habits for turning AI access into real organizational capability.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing support, AEO/GEO/SEO optimization, image concept development, and editorial structuring support from ChatGPT, an AI assistant. All final editorial decisions, perspectives, and publishing choices were made by Alicia Shapiro.