What's your rate? Billing for human effort outside the billable hour.

Many clients today expect you to work through an API key they provide. That way, data stays in their network and they have more control over what an AI has access to. But this means the cost of goods has largely been accounted for and some clients have started to push back on the hourly rate. It’s a fair concern: What are you really charging for?

The billable hour used to be a proxy for effort. Now the work is prompting and babysitting agents, and the human piece is the judgment calls in between. There’s no shortage of tools that measure AI usage. While they help managers account for token cost, they miss the human effort of steering the agent. So how do we account for judgment and work done while being honest about how much of it was ours?

In this talk, I redefine work as phases of effort instead of units of time. I’ll demo a tool I built that reimagines the invoice as phases of work, separating effort that led to an outcome from sessions that produced no outcome. From there, I propose a tiered rate structure based on effort and outcomes. After all, failed effort is still human effort. You’ll leave the talk with a practical framework for pricing services as a freelancer and a deeper understanding of how to make AI-assisted work visible to clients.