Publications
Every paper on this site as a plain list, newest first.
AI assistance in knowledge work: separating adoption, workflow gains and realised value
· Analytical perspective
Generative AI can improve knowledge work. Published evidence shows faster and better professional writing in randomised tasks, customer-support gains under a quasi-experimental staggered rollout, and higher-rated product-innovation proposals relative to solo work without AI in a field experiment. Other evidence records worse answers outside a tested frontier, limited changes in measured office work patterns, and near-null average earnings and recorded-hours effects. These findings concern different interventions, populations, dates and outcomes. They are not interchangeable verdicts on whether AI works. This paper asks how an enterprise can test whether AI assistance improves a defined knowledge-work workflow, and distinguish that improvement from adoption, perceived time savings and realised financial value. It argues for evaluating acceptable outcomes across the full production boundary, including downstream recipients' checking and rework, then documenting what happened to any capacity released. Quality, service, accessibility, learning and sustainable workload can justify investment without becoming cash savings. Existing methods already supply the foundations, so no new branded framework is warranted. The proposed contribution connects them through three reporting artefacts that are not field-validated. The appropriate decision is not always to scale, nor always to stop. It is to identify which claim the evidence supports and what remains untested.
Download PDF →The enterprise is where AI co-existence gets real
· Perspective
A work-in-progress perspective, not a peer-reviewed paper, not an empirical study, and not a completed account of displacement. It responds to the published excerpt and publisher descriptions of Ethan Mollick's Co-Existence, not to a reading of the full book. I treat large enterprises as an early setting in which authority, work and accountability have to be rearranged together, not as the first users of AI and not as a representative sample. The argument is that organisational fit is the missing piece: a faster calculation and a faster organisation are not the same thing. Insurance and retail scenes are anonymised conversations, not audited case studies. Published evidence is cited inside its original limits. The 2027 expectation is a bet that publication-cleared examples of bounded production autonomy will appear, and it is explicitly open to being wrong. Disclosure: I work at Strategy, formerly MicroStrategy. My interest in semantic definitions overlaps with my employer's business. The argument is mine, not a statement on its behalf.
Download PDF →