I run an e-commerce operation and I rely on large-language-model platforms—ChatGPT, Claude, and Gemini—to handle customer service queries around the clock. I now need those responses tuned so they are sharper, more accurate, and far more engaging for shoppers who reach out. Your main objective is to lift response accuracy and user engagement. That means auditing my current prompt flows, rewriting or restructuring them where necessary, and introducing any proven frameworks or tagging methods that help the models understand context better. I am open to prompt engineering, function-calling setups, or lightweight fine-tuning if that moves the needle. You will work directly with live customer service transcripts pulled from our store, test iterations in each model environment, and document clear before-and-after metrics—think CSAT, click-through on suggested links, or reduced clarification follow-ups. Previous success stories doing this for other brands are essential; I’ll ask to see concrete examples. Deliverables • A revised prompt/playbook set for ChatGPT, Claude, and Gemini • A short report showing measurable gains in accuracy and engagement on customer service queries • Guidance for my team so they can maintain and expand the system themselves If you can start straight away and have the track record to prove it, let’s talk.