Start with the archive
The Studies page is the primary archive surface, bringing together current working papers, committed instruments, and preserved legacy records.
Open pagePublic research archive
MKAI is a public research archive on AI and organisations.
Founded by Richard Foster-Fletcher, it publishes restrained, citable records on how AI changes organisational judgement, governance, reporting, and deployment. The archive is built to be read carefully, cited properly, and checked against its stated evidence standard.
Follow the argument
The Research map groups the published record by the questions it addresses, so readers can move across connected studies rather than isolated posts.
Open pageCheck the standard
The evidence standard states what enters the archive, how records are versioned, and how corrections and scope limits are handled.
Open pageRecent records
Comparative Output-Coding Study on Senior Strategic Prompts
Published instrumentInstrument
Committed instrument, 6 July 2026
The committed prompt set and coding scheme for a comparative output-coding study of model responses to senior strategic prompts.
The Mandate Study
Published working paperWorking paper
Published July 2026
A public-record study of how large organisations require, incentivise, and formalise AI engagement through mandates, incentives, training, workflow integration, and board-level expectation.
How Senior Executives Access AI Capability Beyond Sanctioned Tools
Published working paperWorking paper
Published July 2026
A study of how senior executives at large organisations may reach AI capability beyond the tools their institutions formally sanction.
The Unread Instruction
Published working paperWorking paper
Published July 2026
A documentary study of whether major board-governance and AI-oversight frameworks require anyone to know the organisation-authored system prompt shaping an enterprise AI deployment.
The Missing Override Architecture
Published working paperWorking paper
Published July 2026
A documentary reading of public administration and compliance documentation across seven enterprise AI deployments to test whether rank confers a query-time override to the governed layer.
The Liability Transfer
Published working paperWorking paper
Published July 2026
A documentary reading of responsibility and acknowledgement clauses in the public contractual terms of seven enterprise AI deployments.
Archive note
Existing published inquiry and examination URLs remain in place. Where a record first appeared under an earlier archive section, that original address is preserved and still serves as the canonical public citation point.
The record should stay stronger than the narrative built on top of it.