The State of Enterprise AI 2025 Report

Executive Summary

OpenAI's "The State of Enterprise AI 2025 Report" indicates that artificial intelligence is transitioning from consumer-facing applications to core organizational infrastructure. With over 1 million business customers, OpenAI reports a massive surge in usage intensity: ChatGPT message volume has grown 8x and API reasoning token consumption per organization has increased 320x year-over-year.

Scaling and Deepening Integration

Enterprise AI is moving beyond simple experimentation into repeatable, multi-step workflows. This shift is driven by two primary mechanisms:

Custom GPTs and Projects

Custom GPTs and Projects allow organizations to codify institutional knowledge and automate workflows through internal system integrations. Weekly users of these tools have increased 19x year-to-date, and approximately 20% of all Enterprise messages are now processed via a Custom GPT or Project. For example, BBVA utilizes more than 4,000 GPTs to embed AI-driven workflows into daily operations.

API and Developer Workflows

As firms transition to production deployments, API consumption has scaled rapidly. More than 9,000 organizations have processed over 10 billion tokens, and nearly 200 have exceeded 1 trillion tokens. The 320x increase in average reasoning token consumption suggests a systematic integration of more intelligent models into expanding products and services.

Measurable Business Impact and Productivity

AI is delivering material productivity gains and expanding the technical capabilities of non-specialist workers.

Operational Gains

Survey data from nearly 100 enterprises shows that 75% of workers report improved speed or quality of output. On average, ChatGPT Enterprise users save 40–60 minutes per active day, with data science, engineering, and communications workers saving 60–80 minutes. Specific functional gains include:

IT: 87% report faster issue resolution.
Marketing and Product: 85% report faster campaign execution.
HR: 75% report improved employee engagement.
Engineering: 73% report faster code delivery.

Expansion of Technical Work

AI is acting as an equalizer, enabling 75% of users to complete tasks they previously could not, such as programming support, spreadsheet automation, and custom agent design. This is particularly evident in coding-related messages, which have grown by 36% among non-technical teams (outside of engineering, IT, and research) over the last six months.

Industry and Geographic Trends

While adoption is broad-based, the pace of acceleration varies by sector and region.

Sector Growth

The median sector grew more than 6x year-over-year, with the technology sector leading at 11x. Other fast-growing sectors include healthcare (8x) and manufacturing (7x).

API usage patterns differ by industry:

Technology: Focuses on in-app assistants, search, and agentic workflow automation.
Professional Services: Prioritizes coding and developer tools to accelerate delivery.
Finance: Often begins with customer support to realize immediate ROI, followed by coding tools for system migration and compliance.

Global Expansion

International adoption is accelerating. Markets such as Australia, Brazil, the Netherlands, and France have seen business customer growth exceed 143% year-over-year. Japan currently has the largest number of corporate API customers outside of the U.S.

The Adoption Gap: Frontier Workers vs. Median Users

A significant divide is emerging between "frontier workers" (the 95th percentile of adoption intensity) and the median worker.

Message Volume: Frontier workers generate 6x more messages than the median worker.
Specialized Tasks: In data analytics, frontier workers use the data-analysis tool 16x more than the median. In coding, they send 17x as many messages.
Outcome Correlation: Users engaging across approximately seven task types report five times more time saved than those using only four.

At the firm level, frontier firms generate 2x more messages per seat and 7x more messages to GPTs than the median enterprise, indicating deeper organizational integration.

Real-World Case Evidence

Case studies demonstrate that AI is being applied to specific operational challenges to drive revenue and efficiency:

Intercom: Used the Realtime API to reduce latency by 48% for its Fin Voice agent, resulting in 53% of calls being resolved end-to-end.
Lowe's: Deployed Mylow and Mylow Companion, answering nearly 1 million questions monthly and doubling conversion rates for online visitors.
Indeed: Used GPT-powered job matching and career coaching to increase started applications by 20% and improve downstream success (interviews/hires) by 13%.
BBVA: Automated over 9,000 annual legal queries via a chatbot, redeploying 3 FTEs and achieving 26% of the Legal Services division's annual savings KPI.
Oscar Health: Integrated member-facing chatbots with internal data to answer 58% of benefits questions instantly and handle 39% of benefits messages without human escalation.
Moderna: Reduced a core analytical step in Target Product Profile (TPP) development from weeks to hours.

Strategic Framework for AI Maturity

Leading firms consistently employ five key practices to maximize AI impact:

Deep System Integration: Enabling secure access to company data via connectors.
Workflow Standardization: Promoting the creation and sharing of repeatable solutions (e.g., GPTs).
Executive Sponsorship: Setting clear mandates and securing resources for experimentation.
Data Readiness: Codifying institutional knowledge into machine-readable routines and running continuous evaluations.
Change Management: Combining centralized governance with distributed enablement via AI champions.