We work where complexity is the problem.
Not a strategy shop, not an AI vendor, not a think tank. A single firm that does the work of all three, with systems science as the spine.
The gap between advising, building, and publishing.
Institutions today face problems with too many moving parts to reason about on intuition. A government deciding how to allocate compute and energy. A foundation measuring change across a whole sector. An enterprise rewiring itself around AI. These are not slide problems. They are systems, with feedback, delay, and emergence, where the obvious intervention often makes things worse.
The market answers this badly. Strategy firms produce decks that never touch production. AI vendors ship pilots that stall. Think tanks publish papers no one operationalizes. Each holds one piece of the puzzle and none holds the whole.
Disnesta exists to close that gap. We treat every engagement as a complex adaptive system, study it with the discipline of a research practice, frame the decision so a leader can act, and build the AI that makes the decision live. Systems science is the spine that connects all of it.
The systems science lineage.
The firm's method descends from a real tradition: general systems theory, cybernetics, and system dynamics — the line that runs from Bertalanffy and Ashby through Forrester and Meadows. That tradition built the tools for reasoning about feedback, delay, and leverage long before "systems thinking" became a slide title.
Disnesta applies that lineage to the defining systems question of this decade — the redistribution of work between humans and machines — and to the institutions that must govern it. The indices, the advisory method, and the systems we build are all expressions of the same discipline.
Commitments we hold ourselves to.
We build what we recommend. Advice we would not ship ourselves is not advice we give.
We publish in the open. Our research stands on its own, before any engagement.
Senior people do the work. The team that scopes the problem is the team that solves it.
We optimize for the system, not the deliverable. The goal is a system that works after we leave.