Filter Generic Advice · Where to Next

Before ranking, Copilot must throw things out. That is the skill. A list of twelve “you could use AI for email” ideas is a failed interview.

Drop what could apply to anyone in your title.

Drop or demote: already working, generic to the job title, multi-month transformation, or impossible with your stated tools and data rules.

If almost everything got dropped, your interview was thin or you already have a tight practice. Either thicken context or pick a narrower slice of the week and re-run.

Core principles

Generic-to-title is a fail, not a maybe.
Unsafe or off-limits data is a drop, not a stretch goal.
Prefer experiments you can finish in one sitting.
Keep a “not now” list so good ideas are not lost — they are just not next.

If candidates are still generic: Personal Brain

Check yourself

When should you throw out a suggested Copilot use case?

Do this in Copilot

From the interview summary, strike anything generic, already working, or blocked. Keep at most eight candidates.

If Copilot pads the candidate list, send it back: “Too generic. Strike and ask.”

Apply the filters

Here is the interview summary: [PASTE]. Filter hard. Drop or demote anything that is (a) already working, (b) generic advice for anyone in my job title, (c) a multi-month transformation, or (d) impossible or unsafe with my stated Copilot / data limits. Return two lists: Candidates (max 8, one sentence each, specific to my world) and Not now (with the filter that hit). If Candidates looks generic, say so and ask for the missing context instead of padding the list.

Techniques: Constraints

Open Copilot

Did you run this in Copilot? Mark complete when you have tried it.