The Scope Problem

I was running a quarterly review with an agency last month and something odd came up in the numbers. Revenue per client was flat, but the team was doing about a third more work than they had been six months earlier. Nobody had noticed because they were getting it done.

The AI tools they'd adopted were working exactly as intended. Research was faster. First drafts were faster. Production timelines had compressed by about 40 per cent. So the team, being good at what they did, kept saying yes. Yes to the extra revision. Yes to the additional format. Yes to the scope extension that would have taken three days before but now took one.

The client was delighted. The team was busy but coping. The margins, when I finally got someone to calculate them properly, had dropped by eleven points in six months.

Nobody had repriced. Nobody had renegotiated scope. The efficiency gains from AI had been quietly absorbed by the client, who hadn't asked for them and probably would have paid for them.

The founder's response was interesting. She said they'd assumed faster meant more profitable. It should have been, but only if you hold the scope constant and take back the time. Instead they'd been giving away time that wasn't free.

They rebuilt their scoping process to reflect the new delivery speeds, set clear boundaries on what was included, and started quoting for the value rather than the hours. Revenue per client went up by about 20 per cent within a quarter.

There's a section on pricing and scope in the AI frameworks at scaleatspeed.com/ai