A framework · Michael Borck
A framework for using AI as a collaborator, not a delegate.
The dominant pattern of AI use in professional work is delegation. A problem arrives. The work is handed to a model. A polished result comes back. The work appears to be done.
The cost of this pattern is not in the output. The output is often serviceable. The cost is in what stops happening in the head of the person who used to do the work — the small, quiet, judgment-forming activity that used to happen between problem and answer, and that is the actual substance of professional practice.
Conversation is the alternative posture. Conversing with an AI is not the same as asking it for an answer. It is brainstorming with it, ideating against it, interacting with the output it produces, pushing back when it produces something that looks plausible but is wrong. The work that used to happen quietly stays where it should: in the head of the practitioner, with the model as a counterpart rather than a substitute.
Conversation preserves judgment. Delegation outsources it. The distinction is not philosophical. It is operational, and it is the difference between a firm whose use of AI strengthens its practice and one whose use quietly hollows it out.
The conversation loop
Define your question. Arrive with a problem worth thinking about, not a task to outsource. The quality of what comes out of the loop is bounded by what you bring in. Before prompting, check whether you are describing the problem or your attempted fix; if it is the fix, step back one level.
Explore possibilities. Generative and divergent — alternative angles, adjacent ideas, framings you had not considered. The AI's breadth across domains is genuinely useful here. Resist the temptation to stop at the first interesting output.
Push back and refine. This is where conversation actually happens, and where delegation is most tempting. Correct course. Redirect. Disagree on substance. Every response sharpens the output, and every one requires your judgement. Stay in this stage longer than feels necessary.
Make it yours. Take the best of what emerged and edit, restructure, combine, cut. Bring it into your voice, your context, your standards. The result should be something you would put your name on without hesitation.
Most good work loops more than once. The arrows go forward, but the feedback arc is always available — Brainstorm again with a sharper question, return to Ideate when iteration reveals a gap. The sign that you are done is not that the AI has stopped producing output; it is that your thinking has landed somewhere solid.
The books
Both books are free to download at books.borck.education under CC BY-SA 4.0, also available on Amazon. Each comes with a companion chatbot, a navigable wiki, and GitHub access to source material.
Applying this in a regulated Australian small firm is a different kind of work than reading it. That is what borck.consulting is for.
Apply this in your business →