The Okta Enterprise AI Index
How 20,000+ businesses are building their AI stacks
Products
Our platforms secure all types of identity from AI agents to your customers, employees, and partners.
Skip to content
Agentic AI Security
Identity & Access Management
Privileged Access Management
Identity Threat Detection & Response
Customer Identity & Access Management
Identity Governance & Administration
Integrations
Solutions
Okta is the neutral, powerful, and extensible platform that puts identity at the heart of your stack.
Customer Success Stories
Developers
Okta and Auth0 deliver flexible, secure access. Build fast with our extensible platforms for customers, workforce, and non-human identities.
Okta Platform
- Okta Integration Network
- Developer Center ↗
- Product Releases
- Product Documentation ↗
- Community ↗
- Okta Platform Status ↗
Company
Our platforms secure all types of identity from AI agents to your customers, employees, and partners.
About Okta
Careers & community
Values & Impact
Products
Our platforms secure all types of identity from AI agents to your customers, employees, and partners.
Agentic AI Security
Identity & Access Management
Privileged Access Management
Identity Threat Detection & Response
Customer Identity & Access Management
Identity Governance & Administration
Integrations
Solutions
Okta is the neutral, powerful, and extensible platform that puts identity at the heart of your stack.
Customer Success Stories
Developers
Okta and Auth0 deliver flexible, secure access. Build fast with our extensible platforms for customers, workforce, and non-human identities.
Okta Platform
- Okta Integration Network
- Developer Center ↗
- Product Releases
- Product Documentation ↗
- Community ↗
- Okta Platform Status ↗
Company
Our platforms secure all types of identity from AI agents to your customers, employees, and partners.
About Okta
Careers & community
Values & Impact
Generative AI once sat on IT’s restricted list. In just a few years, enterprises have moved from blocking the technology on corporate networks to enabling it across their workforces. Today, leaders aren’t debating if this technology belongs in their business. They’re deciding which AI platforms, and how many, they need to stay competitive.
Where are businesses concentrating their AI investments? To find out, we looked at real-world numbers: anonymized application access data from more than 20,000 organizations on the Okta Platform, spanning June 2022 to June 2026.
The data offers a snapshot of the tools organizations are deploying specifically for developers and knowledge workers. It captures enterprise AI usage from the last four years across six functional categories experiencing high generative AI momentum: foundational models, developer tooling, enterprise search, collaboration and productivity, audio transcription and meeting AI, and creative suites. (Read the Methodology)
For leaders trying to measure their own progress, this data reveals how the broader market is actually deploying AI.
Key findings
Two speeds of growth: AI-native startups are growing fast, but AI-enhanced incumbents still hold most of the enterprise market.
From autocomplete to action: AI is shifting from reactive chat prompts to autonomous agents that can act on our behalf.
Best tool for the job: Enterprises aren’t locking into single-vendor setups but are instead running more than one AI platform at once.
How we mapped the enterprise AI landscape
To identify which AI platforms businesses are using, we analyzed anonymized Okta Single Sign-On (SSO) data, tracking more than 100 distinct AI products, and consolidating them into 74 product suites.
We then needed an equitable way to compare a startup launching from scratch with a tech giant that already has millions of existing users. Our solution: a framework based on baseline growth velocity. First, we ranked the top 30 performers by measuring their growth in enterprise accounts from June 2022 through June 2026, as shown in Figure 1. Then, under this framework, we categorized companies into two distinct groups (see Methodology for more detail):
AI native
Pure-play disruptors that multiplied their corporate customer base by more than 4x over this period. These are companies such as Anthropic, OpenAI, and Cursor, where generative AI is the foundational core of their products.
AI enhanced
Established incumbents growing below the 4x threshold. These are companies such as GitHub/Microsoft 365, Google Workspace, Figma, Slack, Adobe, and Notion, which achieved enterprise success before the generative AI era but have since adapted their platforms to lead the market forward.
The dual-engine market: Innovation versus scale
The fastest-growing enterprise AI applications
Figure 1: The top 30 fastest-growing enterprise AI applications, ranked by enterprise account growth from June 2022 to June 2026. Index growth scores are calculated by normalizing to the maximum enterprise account growth achieved within the measured timeline (index max = 100). Applications are classified as "AI native" when account growth exceeds 4x and "AI enhanced" otherwise.
The data shows the enterprise market isn’t a zero-sum battle. AI-native startups are defining new capabilities, while AI-enhanced incumbents are putting the tech in front of enterprise customers they already have. Both startups and incumbents are growing, but from different starting points.
Seeing the two groups represented in the Top 30 suggests enterprises aren’t taking sides. Instead, businesses are likely adopting specialized startups for new or novel use cases while simultaneously expanding their use of established incumbents.
The 3 paradigms of enterprise AI
The enterprise AI landscape hasn't just grown. It has fundamentally evolved. As illustrated in Figure 2, we have experienced three generative AI paradigms over the past four years, each redefining how businesses operate and how platforms capture market share.
Top 10 enterprise AI applications’ growth trajectories
Figure 2: Growth trajectories of the top 10 enterprise AI applications from June 2022 to June 2026. Data is normalized to the maximum monthly enterprise account volume achieved by any single application within the measured timeline (index max = 100).
The autocomplete paradigm (begins June 2022)
GitHub Copilot’s general availability (GA) in June 2022 marks the first massive, mainstream generative AI success in real enterprise workflows. GitHub’s first-mover advantage gave it an early lead and cemented its initial mindshare among developers and corporate IT leaders alike.
The chat paradigm (begins November 2022)
The “Big Bang” launch of ChatGPT in November 2022 introduced conversational chatbots to the public, kickstarting the generative AI era. While consumer adoption was immediate, formal enterprise deployment lagged. This gap resulted in shadow IT, with employees using personal accounts for work. The arrival of enterprise versions, such as ChatGPT Enterprise, Claude Enterprise, and Gemini Enterprise, pulled these tools out of the shadows and into official processes. Meanwhile, AI-enhanced incumbents saw steady customer growth as they rolled out new AI features. OpenAI stood out among AI-native startups, but the pre-existing scale of giants such as GitHub and Google Workspace allowed incumbents to add more total accounts.
The agentic paradigm (begins May 2025)
Cursor’s developer-driven hockey-stick adoption and Claude Code's explosive growth in spring 2025 kicked off a new paradigm. During this period, AI expanded from "conversational assistants" that merely ask questions to a vision of "autonomous agents" that act with minimal oversight. Operating within deeply integrated environments, AI can now navigate codebases, execute multi-step goals, and resolve complex tasks without constant human prompting.
This evolution in product capabilities quickly translated to market growth: The top three AI-native apps (Anthropic, OpenAI, and Cursor) added more new accounts than any incumbent. According to Figure 2, in March 2026, Anthropic overtook OpenAI in enterprise accounts, and our underlying usage data shows it surpassed OpenAI in monthly active users (MAU) the following month. Even so, the incumbents still dominate total volume. The top-performing AI-native platform, Anthropic, reached less than 50% of the account volume achieved by the leading AI-enhanced application, Microsoft 365. Similarly, our usage data reveals it had an MAU less than one-tenth of Microsoft 365's.
The rise of the multi-vendor AI platforms
Businesses have likely made up their minds about AI’s value. What they probably haven't landed on is a favorite platform. Figure 3 highlights that as of June 2026, the majority of enterprise customers use more than one AI platform simultaneously. The move away from single-platform deployments is picking up speed: The number of companies using only one AI provider shrank by 1.2 percentage points in June compared to May as businesses expand their AI stacks. The actual scale of this multi-vendor landscape is likely even larger, as these figures exclude companies accessing models through centralized IT gateways, cloud providers, or AI model hosting services. For the time being, enterprises are keeping their options open.
Neutrality permeates the AI enterprise ecosystem
Figure 3: The percentage of multi-vendor AI platform adoption across enterprise organizations as of June 2026. For this ecosystem analysis, platforms are aggregated at the parent-company level for enterprise AI applications (e.g., “Microsoft” includes both GitHub and Microsoft 365).
Staying flexible, staying secure
Four years in, the enterprise AI landscape is still wide open and fundamentally shifting. Forward-thinking businesses aren't waiting for it to settle. They’re running diverse, multi-vendor stacks to drive real productivity today.
But this reality comes with fresh considerations. As AI evolves from software that takes orders to agents that can act autonomously across these systems, organizations face a widening security gap. To close it, they must treat these new digital workers as first-class identities, answering three fundamental questions: Where are my agents? What can they connect to? And what can they do?
Making this diverse ecosystem work requires a connecting fabric where humans, systems, and AI agents can collaborate across any platform. In a multi-vendor future, a neutral identity layer is a critical part of that foundation, helping organizations not just turn on their AI tools, but securely deploy them.
Methodology
This analysis is based on anonymized enterprise SSO data from the Okta Platform (specifically, Okta Workforce Identity, excluding Auth0). We tracked 109 distinct AI products, tools, and brand aliases, which were selected and grounded using AI-driven discovery, domain expertise, and leading industry analyst reports.
This raw data was mapped to 72 parent companies and normalized into 74 consolidated vendor product suites (which we refer to as “applications” throughout this report) across six categories: foundational models (10), developer tooling (14), enterprise search (5), collaboration & productivity (22), audio transcription and meeting AI (6), and creative suites (17).
Our scope focuses on ready-to-use software and coding assistants. AI model delivery via hyperscaler cloud platforms, AI infrastructure, systems-of-record software, AI embedded within data platforms, and cybersecurity AI are excluded from the study. By focusing on these six categories experiencing high generative AI momentum, we can offer clearer insights into the evolution of the enterprise AI application landscape.
To establish a standardized, equitable comparison between emerging AI-native pure-plays and entrenched AI-enhanced incumbents, we measured total enterprise customer acquisition from June 2022 to June 2026. We ranked the top 30 performers by net growth in unique enterprise companies. To normalize the data across different baseline scales, we indexed this growth by assigning a benchmark score of 100 to the application with the highest net company growth, scaling all other vendor scores proportionally. To objectively classify vendors, we used a data-driven threshold: companies with a customer growth ratio greater than 4x between June 2022 and June 2026 were classified as “AI native,” while those with a ratio below 4x were labeled “AI enhanced.” This multiple was selected through a combination of empirical business understanding and machine learning. Specifically, to accurately categorize companies in the growth spectrum (<10x) where AI-enhanced and AI-native separation is less obvious, a two-cluster k-means algorithm identified a clean mathematical division at the 4x threshold. This boundary reflects whether AI serves as a foundational product offering or as an integrated capability within an existing product.
While these findings offer valuable insights into enterprise adoption trends, they reflect behavior within the Okta Platform and may contain inherent biases that do not fully capture global enterprise adoption.
Timeline resources
- Autocomplete paradigm: GitHub Copilot GA (June 2022)
- Chat paradigm: ChatGPT launch (November 2022); Google Bard launch (March 2023); Claude.ai launch (July 2023); ChatGPT Enterprise launch (August 2023): GitHub Copilot Chat GA (December 2023); GitHub Copilot Enterprise GA (February 2024); Claude Enterprise (September 2024); Gemini Enterprise GA (October 2025)
- Agentic paradigm: Claude Code GA (May 2025); Cursor 1.0 (June 2025); OpenAI Codex app GA (February 2026); Claude Cowork GA (April 2026); Google Antigravity 2.0 (May 2026); Microsoft Scout preview (June 2026)
Acknowledgments
Thanks to Aditi Tupsakhare for refining data queries, Laurie Isola and Lauren Everitt for editing the article, and Nomi Coltrane for elevating the visual design.
