AI Agents as Roles and Trade-offs

Spencer Graham and Nicholas Naraghi, co-founders of Hats Protocol, pioneering the design and experimentation of decentralised, programmable organisations, join us in this episode to explore how these new forms of collaboration can enable new ways to organise, govern, and create value collectively.
They discuss the future of organisational design, including how AI agents can take on roles, how frameworks and reusable templates accelerate experimentation, and why adopting different role-based “Hats” can help individuals contribute meaningfully in new, decentralised ways.
They also speak on how decentralisation can lower risk, increase system “hardness,” and improve predictability, while reflecting on what it means to distribute responsibility in a world where the boundaries of firms are increasingly fluid.
Tune in to discover a more participatory way of organising that helps solve the principal-agent problem.

Together, Spencer and Nicholas have been pioneering new ways of structuring DAOs and digital-native organisations, making roles programmable, modular, and resilient for several years now.
In their work, they bring deep experience in building DAOs, governance frameworks, and infrastructure that allow organisations to operate with greater transparency, adaptability, and distributed decision-making.
As we explore role-based structures to enable meaningful participation, we learn what it means to build adaptive systems capable of tackling complex challenges in a decentralised world.

Key Highlights
👉 Decentralised organisations reduce the cost of organising by embedding rules, roles, and incentives directly into software, minimising the need for traditional bureaucratic structures.
👉 AI agents can take on organisational roles, augmenting human capabilities and enabling more modular, scalable coordination.
👉 Lowering coordination costs increases the responsibility for individuals to participate meaningfully in organisational life.
👉 Roles within organisations can be made programmable and modular, allowing for flexible experimentation and adaptation.
👉 Reusable frameworks and templates accelerate organisational experimentation, letting groups test new coordination methods quickly.
👉 Individuals may act like “micro-organisations” with AI agents representing them, but collaboration will always remain necessary for complex problem-solving.
👉 Tokenisation and algorithmic governance allow individuals to earn ownership and rewards proportional to the value they create in an organisation.
👉 Participating in decentralised organisations requires embracing uncertainty, both in outcomes and in coordination dynamics.

Topics /chapters
(00:00) The Benefits of Programmable Organizations
(01:39) IntroducingSpencer Graham and Nicholas Naraghi (Hats Protocol)
(03:35) From DAOs to Roles: The Birth of the HATS Protocol
(10:11) The Principal-Agent Problem
(15:00) Getting Buy-In on Protocols
(22:03) Separating Tech from the Principal-Agent Problem
(30:31) When Organizing Becomes Cheap: What New Organizations Will Emerge?
(40:43) Uncertainty with Autonomy
(46:05) Will “Organising” become a necessary skill?
(49:05) Breadcrumbs and Suggestions

Remember that you can always find transcripts and key highlights of the episode on our website: https://www.boundaryless.io/podcast/protocol-Hats

Episode recorded on Oct 01, 2025

Find out more about the show and the research at Boundaryless at https://boundaryless.io/resources/podcast/

Get in touch with Boundaryless:

• Twitter: https://twitter.com/boundaryless_
• Website: https://boundaryless.io/contacts
• LinkedIn: https://www.linkedin.com/company/boundaryless-pdt-3eo

Music

Music from Liosound / Walter Mobilio. Find his portfolio here: https://blss.io/Podcast-Music