> Site index: [llms.txt](https://ortelian.com/llms.txt) with all pages and descriptions.

# What is AI native?

Source: https://ortelian.com/notes/what-is-ai-native/
Markdown: https://ortelian.com/notes/what-is-ai-native.md
Author: Laurens Nys
Published: 2026-11-16

An AI-native company is one where agents carry whole jobs across the company's systems, and people set the rules and take the exceptions. Most companies are not there yet.

An AI-native company is one where agents carry whole jobs across the company's systems, and people set the rules and take the exceptions. That is a different thing from a company whose people use a chat tool. In [Notion's 2026 study](https://www.notion.com/resources/inside-the-ai-transformation) of 6,118 people in 10 markets, 57% were still at the stage where AI only helps one person, 31% were at the stage where AI can read the company's systems, 10% were at the stage where agents carry jobs, and 2% were AI-native. Those percentages are Notion's. The stage names are ours. We set this out on the [Ortelian homepage](https://ortelian.com/).

## What AI native means

"AI native" describes a company, not a tool. In an AI-native company, agents run the processes the business depends on. They work from one model of the company, so every agent sees the same customer, the same order and the same rules. People decide what an agent may do, and they step in when a case needs a person.

It is not a chatbot rollout. It is not a pile of AI features added to the tools you already have. And it is not the same as being "AI enabled," which we set out in [AI native vs AI enabled](https://ortelian.com/notes/ai-native-vs-ai-enabled/).

## What is an AI-native company?

A company can use a lot of AI and still not be AI-native. The test is whether AI owns actual work, not only answers questions.

We use four stages. They map to the four levels in Notion's study, but we name them for what the company can actually do:

1. AI helps one person. People use a chat tool to write, to sum up and to think. AI sees only what each person shows it. Notion found 57% of respondents here.
2. AI reads your systems. AI tools are connected to the company's data. Tasks go faster, but a person still carries every job, and each tool can read the company in its own way. Notion found 31% here.
3. Agents carry jobs. Every record is held once, in one model. An agent takes a job from start to end, and stops where a person must decide. Notion found 10% here.
4. AI-native. Agents run the processes the business depends on. People set the rules and take the exceptions. A new job can start from the model that the last job built. Notion found 2% here.

88% of respondents were still at the first two stages. The hard step is the one between stage 2 and stage 3: the company needs one shared picture of itself before an agent can carry a job.

## The company OS under the name

That shared picture is what we call a company OS: one model of the company, the rules that apply, and the actions an agent may take, built on the systems the company already uses. It is the layer agents work through, not a second copy of the company's records. How to build one, without a migration, is in [how to build a company OS for AI agents](https://ortelian.com/notes/company-os-for-ai-agents/). What it is, and what it is not, is in [what a CompanyOS actually is](https://ortelian.com/notes/what-a-companyos-actually-is/).

[Ortelian's technology](https://ortelian.com/technology/) is that model, plus the controls on what each agent may read and change, and the agents that run a job from start to end. The systems stay as they are.

## Why so few companies are there

A chat tool can make one person faster and leave the company's numbers unchanged. McKinsey's [state of AI in 2026](https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai) survey of 1,719 people found that only 6% of respondents both attributed 5% or more of EBIT to AI and called the impact significant. Those organizations were much more likely to redesign the work: nearly three-quarters of them said they had fundamentally redesigned workflows because of AI, against one-quarter of the rest.

An AI-native company has done that redesign. It holds the company in one model, decides where a person checks the AI, and measures the result against a standard. An established company gets there one job at a time.

## Questions people ask

### What does AI native mean?

It means agents carry whole jobs across the company's systems, from one shared model of the company, and people set the rules and take the exceptions. It does not mean everyone has a chat tool.

### What is an AI-native company?

A company at the last of four stages: AI helps one person, AI reads the systems, agents carry jobs, then AI-native. Notion's 2026 study put 2% of respondents there.

### Is AI native the same as AI enabled?

No. AI enabled adds AI to work that already exists. AI native changes the work so an agent can carry the job. The comparison is in [AI native vs AI enabled](https://ortelian.com/notes/ai-native-vs-ai-enabled/).

### Do you have to migrate systems to become AI-native?

No. The company OS sits on top of the systems you already use. The records stay where they are.

## Sources

- Notion Research,

  [The State of Global AI Transformation](https://www.notion.com/resources/inside-the-ai-transformation)

  (The Great Renovation), 2026. Fielded via Qualtrics, 6,118 AI decision makers and AI users in 10 markets, March to May 2026. The 57%, 31%, 10% and 2% figures are the share of respondents at each of Notion's four levels. The stage names in this note are Ortelian's.
- McKinsey,

  [The state of AI in 2026](https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai)

  , 25 August 2026. Global survey, 1,719 participants, 4 May to 8 June 2026.
