reggz

5 ms·

1.
by reggz 5mo ago
yeah you’re not wrong, but the issue is usually mixing pricing lanes. The clean comparison is plan-only vs plan+overflow vs usage-priced baseline for the same seat/activity assumptions. Buying credits is fine if it's part of your
2.
by reggz 5mo ago
The signal I’d care about is whether they can scope the task, prompt the model, review the output, explain tradeoffs and clean up what the model got wrong.
3.
by reggz 6mo ago
can treat each internal agent like a small product: one owner, one budget, and one event ID that bundles retries and tool calls into the true cost per successful outcome. If you only track provider spend, nobody knows what one useful outcom
4.
by reggz 6mo ago
I’d break it into a few separate numbers instead of trying to estimate one blended API cost. We usually look at cost per workflow run, runs per active account, and the heavier paths separately, then keep retries and tool calls as their own
5.
by reggz 7mo ago
I think there's the work and there's the compute bill. AI mostly changes the work side since you can finish faster but the compute cost is still real and can vary a lot. The way I’ve seen it handled is a fixed project fee with “no
6.
by reggz 7mo ago
If you're paying for model calls, retrieval, or other backend work on each request, charging earlier is a lot easier to justify because usage can literally cost you money while you're still figuring things out. But I agree that it
7.
by reggz 7mo ago
In practice I’ve found agentic search does improve accuracy on harder queries, but the cost climbs pretty quickly once you move away from a single pass. More steps, more chances to call the model again, more room for retries. It compounds f