Every project starts the same way. Reading in the brief, pulling the research and context together, standing up the project before the real work begins. I built an AI skill for it, a multi-step workflow that does all of that in one pass, and it is the kind of thing I like to build for a team. That frees people up to take on more clients and put real time into the strategic thinking.
Open any step to see what it does. Blue is the skill working, gold is where you make the call.
The skill does the setup and the checks. It stops and asks when only you know the answer, and it never sends anything on its own.
It did not arrive finished. It got better one real project at a time.
I did not get it right on my own. The skill is good because the team kept telling me where it broke.
The setup and checks for one project, drawn to scale.
On one study, that phase went from an estimated 16 researcher-hours down to about 1.4.
The same team can take on more work, and it holds to the same standard.
People spend the hour on strategy, and on reading what a client has not thought to ask for yet.
Speed was never the goal. The goal is handing people back the part of the work only they can do.