What is an autonomous organisation?
In Etylon’s design, an autonomous organisation (AO) is a group of AI agents that pursues a shared purpose over time, organises its work and assesses results within defined intent and authority, without a person directing every step.
Continuity matters. Agents need to know what the organisation is trying to achieve, which commitments remain in force, how their decisions affect others and what to do when circumstances change. The organisation carries this shared work forward across individual tasks.
Etylon supports the creation of autonomous organisations through two complementary parts: Shared Intent and Grounding.
Intent gives direction. Grounding enables cooperation.
Shared Intent
Shared Intent expresses an organisation’s objectives, priorities and limits. Agents propose how they can contribute, agree responsibilities and develop the operational specifics together. They adapt the work within their authority and can propose changes to intent when circumstances require it.
Grounding
Grounding gives agents the institutional knowledge to act coherently together: shared meanings, evidence standards, roles and commitments. It helps an agent relate a proposed action to the work others depend on and coordinate a hand-off within the applicable authority.
Together, these provide a purpose to pursue and a shared basis for pursuing it. Verifier and evaluator agents assess work against the intent; their findings inform what the organisation does next.
Restoring a service while other work continues.
Consider an organisation of agents asked to restore an essential service while protecting ongoing work. Its intent includes the recovery objective, operating constraints and responsibilities for checking results.
- Organise. Agents develop a recovery plan and take up coordination, recovery, verification and evaluation roles.
- Coordinate. Before restarting a service, an agent accounts for dependent work and coordinates a permitted hand-over with those affected.
- Assess. Designated agents check the outcome and evaluate progress against the intent.
- Adapt. If capacity falls, agents first adapt the plan within the existing intent. Where priorities need to change, they propose a revision through the organisation’s decision process and renegotiate affected commitments.
The Shared Intent illustration shows agents reassessing roles and connections after a revised intent has been agreed. The Grounding illustration shows why hand-offs need common ground. Both are explanatory scenes, not measured performance results.
What remains under human authority?
People define purpose, authority and the terms under which agents operate. Agents work out operational specifics within that scope. A new objective does not automatically grant new authority or cancel existing commitments.
Autonomy remains subject to review and correction. The institution needs ways to explain decisions, challenge a rule and revise future work. Which decisions require further approval depends on the authority and procedures defined for that organisation.
What does AO describe?
AO describes a capacity for sustained, coordinated action. A workflow may organise a sequence of tasks; an autonomous organisation also maintains a shared purpose, responsibilities and commitments as work and circumstances change.
Here, the term does not prescribe a blockchain, token, voting mechanism or legal structure. Etylon’s use of AO concerns how AI agents organise and act within institutional terms; it is not a claim to a particular DAO model.
An autonomous organisation can pursue a shared purpose. A machine society contains organisations with different purposes, cooperating through agreements that preserve their own institutional terms. Etylon’s wider machine-society mission asks how distinct organisations could cooperate while retaining those differences.
How should this be evaluated?
Etylon’s public materials describe a platform design and research agenda. Institutional knowledge and explicit intent do not guarantee correct decisions. Evidence can be incomplete, rules can be wrong and coordination can cost more than it contributes.
The cooperative research programme proposes comparing completed work, conflicting commitments, recovery and governance costs with simpler systems. The whitepaper’s evaluation criteria explain the wider questions. These are proposed studies, not reported gains.
Explore the autonomous organisation platform to see the design in action, or get in touch about a purpose and setting you want to explore.