I'm running an AI product that's starting to scale, and I'm noticing our OpenAI/Anthropic bills growing faster than I'd like. We're at the point where costs are becoming a real line item in our budget. Curious how others are dealing with this:

Are LLM costs a top concern for you right now, or is it more of a "figure it out later" thing? What strategies have actually worked to reduce costs? (prompt optimization, caching, cheaper models, etc.) Have you found any tools that help you track/optimize costs effectively, or are you building custom solutions? At what point did costs become painful enough that you had to actively address them?

I'm trying to understand if this is a real problem worth solving more systematically, or if most teams are just accepting it as the cost of doing business. Would love to hear what's working (or not working) for you.