Jiyeon Lee

Hey everyone!
I ve been working on an AI service leveraging LLMs, and I ve hit a bit of a wall.

While the potential for generative AI is exciting, I m seeing real challenges with the economics. The user value of my service seems modest or at least uncertain right now it s hard to clearly define the problem we re solving and how much users would pay for it.

At the same time, the cost of running LLM-based services (especially at scale) is significant. With high compute costs and pricing models that don t always align with what users are willing to pay, it feels like a big risk to just launch and iterate.

Has anyone else faced this dilemma? How do you balance the uncertain user value against the large, unpredictable costs of LLM services? Any frameworks or lessons learned would be super helpful.
Thanks in advance!

I m Jiyeon, working at the intersection of business and AI, with a background in IT consulting and AI-powered platform solutions. My journey has been about bridging technology and user needs, always aiming for practical, high-impact solutions.

Currently, I m focused on LLMOps, building and refining AI evaluation and observability frameworks, and collaborating with global partners to push the boundaries of generative AI services.
Product Hunt has always been a great place for me to spot fresh ideas and connect with other builders and visionaries. If you re into AI platforms, enterprise adoption challenges, or just love talking about how to make tech more accessible and meaningful, let s chat!

I am Head of Marketing at Global AI Platform's US office in Silicon Valley, working on GTM for our app's US public availability launch this Summer'25. We re in the final stages of beta testing our mobile app (focused on meal and weekend planning), and we want to be intentional and avoid rushing just because we feel ready.

We re working on defining exit criteria the metrics, signals, and checkboxes that say:

Yes, it s time to move from beta to full release.