Industrial-era organizations were designed around people, hierarchy, and the coordination costs of human communication. When AI agents can plan, execute, and coordinate work on their own, those design assumptions no longer hold. This program studies what replaces them: how organizations are built, run, and governed as agents move from tools to participants, and what the path from today's organizations to autonomous ones looks like in practice.
Reusable organizational components. Marketplaces for "orglets," organizational building blocks that can be composed rather than built from scratch.
Autonomous orchestration. How work is coordinated when agents, not managers, do the orchestrating.
Strategy and governance. How direction is set and accountability maintained in autonomous organizations.
Shared context across organizations. How context is shared across organizational boundaries when organizations collaborate while safeguarding security, privacy, and safety.
Inter-organizational design and evolution. How organizations compete, connect, collaborate, and change together in autonomous ecosystems.
Higher-order organizations. Organizations whose purpose is to generate other organizations.
People and organizations. How the relationship between individuals and the organizations they belong to changes.
A new programming language for networks of AI agents, combining a 3D/XR user interface with support for multi-user, collaborative applications. Today's AI-agent platforms rely largely on text and 2D interfaces. Kepos explores what becomes possible when networks of agents are embodied in a three-dimensional space that people can enter and work in together.
Autonomous Knowledge Acquisition System
AI agents for continuous research on trends and taxonomies, knowledge-graph development, and pattern identification. The starting point is the knowledge captured in more than 50 Post-Industrial Forum events featuring experts and senior leaders, including a former prime minister.
In this project we will develop new principles and solutions for organizational design in semi-autonomous and autonomous organizations, as well as in ecosystems with multiple organizations.
Post-Industrial Organization Modeling Language
The PIOML project builds on earlier organizational redesign done at PII during our first decade. Helping mostly larger organization with organizational redesign, we developed the Lean Systems Framework, with a rich modeling language that had both a visual and algebraic notation. The current project completes the work of giving the language a formal semantics, so that organizational models can be checked, transformed, and reasoned about by people and by AI agents.
AI-Guided Organizational Transformation
How enterprises that have been slow to adopt AI agents can catch up by applying agents to the transformation effort itself. Organizational transformation is knowledge-intensive and involves social as well as technical challenges. AGOT continues the Institute's organizational transformation methodology, now with AI agents doing much of the work and PIOML as the modeling language: agents gather data from repositories, interview stakeholders, build models of the organization through the lenses of workflow, structure, information architecture, and culture, identify opportunities, and propose redesigns.
Visiting researcher
Paul Collins, Sonoran University