AI saves me real time most days. Not by doing my thinking for me, but by clearing the work around the work, so I can spend my attention where it matters.
These are the habits that have earned their place for me as a surgeon, professor and clinical leader. They translate well to anyone who reads, writes, analyses and decides for a living: leaders, academics and researchers.
Brief the model like a sharp junior
The single biggest lift in output quality comes from how you ask. Give the model a role, the context, the constraints, and one good example of what you want. "Summarise this" gets you mush. "You are reviewing a grant for a sceptical panel. Summarise the methods in 150 words, flag the two weakest assumptions, and keep the tone neutral" gets you something useful.
Build a prompt library
Most professional work repeats. Reference letters, meeting briefs, grant sections, literature summaries, board papers. Save the prompts that work and reuse them. A handful of good templates will do more for your week than any single clever trick.
Use it as a thinking partner, not an oracle
Ask the model to argue the opposite case. Ask what a critic would say. Ask for the three weakest points in your own draft. The value is not the answer it gives, it is the sharper thinking it provokes in you.
Ground research in real sources
General chatbots invent citations. For anything that has to be true, use tools built to cite their sources: Perplexity to explore, Elicit to build evidence tables, Consensus to weigh findings, NotebookLM to question your own documents. Treat every fact and number as unverified until you have checked it.
Turn your own documents into a knowledge base
Load your papers, policies or reports into a tool like NotebookLM and ask questions of them. It is faster than hunting through PDFs, and it keeps the answers anchored to what you actually wrote.
Automate the admin, not the judgement
Let AI capture and summarise meetings, draft routine replies, and tidy your notes into actions. Hand it the repetitive work. Keep the decisions, the relationships and the accountability for yourself.
Protect what is confidential
Never paste identifiable patient data, confidential material or anything you could not email to a stranger into a public AI tool. Know whether your tool trains on your inputs, and use enterprise or local options when the stakes are high.
AI is a force multiplier for judgement, not a replacement for it.
Start with one habit, not ten. Pick the task you do most often that you enjoy least, and put AI on that first. Build from there.
David J McCormack
Surgeon, professor and health-AI leader. Get in touch →