The constraint
The trap in agency economics is that the fix for too much production is more people, and that fix arrives slowly. A senior hire takes months to find and longer to sound like the agency, while the deliverables keep landing on the same desks. The people best placed to grow the business end up spending their weeks producing for it instead.
What made this workload worth building against was its shape. The content changes with every client, but the structure of the work barely moves, and work that repeats in structure while varying in content is the kind agents handle well, provided a person stays on the judgment. MKTG’s production line was almost entirely made of it.
The approach
The build order came out of the map, not the other way around. Looking at deliverables rather than job titles showed where the hours concentrated, which of the three workflows carried the most weight, and what a good result looked like for each one.
We built one agent family per workflow rather than a single general assistant, because the three workflows fail differently. Research fails on depth and sourcing. Copy fails on voice. Reporting fails on the numbers. A bounded agent with one job can be reviewed against a clear standard; a do-everything assistant cannot.
Every family ends at the same review point. The agents changed who does the first pass on a deliverable. They did not change who is accountable for the last one.
The outcome
Around 120 hours a week saved, measured across the agency, and the shape of the week changed with the number. Deliverables now start from agent output instead of a blank page, and the senior hours that were hardest to hire for are the ones that came back.
MKTG agreed to be named as a client, so we can publish the agency and the number together. What sits behind them is better discussed than written; book a call if you want the detail.