Governance
Agreement, permissions, and hard limits.
- Example “Two partners approve any contract.”
yulen.aiv1.0
AGENT
Your agents follow the company's rules.
Clear rules for people and agents.
Yulen turns the shareholders' agreement, the permissions, and the budgets into written rules that people and AI agents follow, with a record of every decision. The first segment is new companies, and organisations that create companies again and again.
PERSON Example of the interface. Not a customer's data.The problem
The agent acts. The rule is still in a PDF.
A company already hands tasks to AI agents. Who approves, how far spending can go, and what is never done still live in an agreement, a chat, or someone's head. An agent cannot follow a rule that is not written down.
Work
Tone, deadlines, and what is never done.
- Example “Reply to a customer in under 24 hours.”
Plan
The quarter's goals and the channel.
- Example “This quarter, win paying customers.”
Money
Collections, payments, and suppliers.
- Example “Invoices are collected at 30 days.”
- Example “Only suppliers that are already approved.”
An example of use, not a Yulen customer: the agreement, the collections, and the rules of an accounting firm can live in a file an agent reads.
How it works
Three steps. A person approves anything outside the plan.
1Write the rules
Agreement, permissions, budget, and what is never done. Each rule is a file with a path and a version.
2Agents work inside them
An action inside the rules goes ahead. Anything outside is asked of the person who has permission. You choose the AI. There is no published automatic recommendation.
3Every decision is recorded
The record says what was done, with which permission, and who approved it. If you change provider, the context leaves in an open format.
Why now
Agents are moving from trials into operations.
In 2025, organisations are already trying agents. What is missing is a written account of what they are allowed to do.
>40%
of agentic AI projects will be canceled by the end of 2027 because of cost, unclear value, or insufficient risk controls. This is a prediction, not an observed result. Gartner, June 2025.
62%
of the organisations surveyed are at least experimenting with AI agents. 23% are scaling an agentic system in some function. Survey of 1,993 people, 25 June to 29 July 2025. McKinsey, November 2025.
Yulen is designed to help with the GDPR and the EU AI Act, and to prepare audits such as ISO 27001. Yulen is not certified.
Who it's for
New companies, and the people who create them again and again.
Yulen starts with whoever is setting up a company, and with the organisations that repeat that work: venture studios, programmes, accelerators, universities, and funds.
New companies
One company, with its rules written from day one.
People who create companies repeatedly
Several companies, each with its own context and its own rules.
What is Yulen?
Yulen turns a company's rules (who decides, how much is spent, what is never done) into written files that people and AI agents read and follow. Every decision is recorded.
Who is it for?
The first segment is new companies and organisations that create companies again and again: venture studios, programmes, accelerators, universities, and funds.
Is it a platform for building agents?
No. Yulen does not build agents. It writes the rules inside which the company's agents and tools work.
What happens if an agent steps outside the rules?
The action does not run. The person who has permission is asked, and the answer is recorded.