How I Work

I advise institutions on federal accountability, and I write publicly about it. Both bodies of work use AI. This page states where it is used, what verification stands behind the result, and who is accountable for it.

Research

Two AI agents deliver structured research feeds daily, each running prompts built for what it does well, one qualitative and one quantitative. I call these the 3P Hunts: Powerful, Purposeful, Thought-Provoking. Some are timed to a date when policy shifts are expected, as on July 1, 2026. Others begin when a study I encounter sends me back to its primary sources. The hunts decide what earns sustained attention; they do not decide what is true. Nothing from either feed enters a draft or a client document until I have verified it independently against the primary source, and statutory claims are checked against the Federal Register or the regulation itself.

Writing

The essays in the Mission Preservation Series and The Changed Ground connect federal policy, labor market data, and institutional strategy rather than summarizing other people’s reports. Once the question, the argument, and the evidence are set, I commission a first draft from an AI model and edit it into my own voice. Isolating the question takes longer than any other stage, sometimes weeks. The argument and the final word are mine. A full account, including which models I use, appears in “The Work Behind The Work.”

Client work

Client deliverables follow a stricter standard than the essays. Every figure and citation in a brief or report is verified against its primary source before it reaches you. No number enters a document by inference, and a claim I cannot source is removed rather than softened. Where a document carries a statutory reading, the citation sits on the page so you can check it against the text. I write the analysis and the recommendations, and I am accountable for them.

The instruments

The estimators and risk instruments published here are decision-support tools built on public federal and state data. They model exposure under stated assumptions. They do not determine an institution’s compliance status, and they are not legal advice. Verify any result that will inform a decision against your own institutional data and against the governing regulation.

Questions

If any of this needs to be more specific for a procurement office or a general counsel, write me at sue@suemukherjee.com and I will answer in writing.