[行业应用] ChatGPT application examples: bulk customer acquisition + generating sales scripts + automated deal closing through chatbot conversations

Disclaimer: This content is for study and research only. No tools are provided, and no services are provided—please don't ask.

The overall workflow is: crawl information from self-media such as videos and images across the internet, analyze the potential customer needs in it, then use the ChatGPT API to analyze the content and reply content. Outgoing content can include WeChat or other contact information depending on the situation. Then wait for customers to add the trained WeChat bot bound to ChatGPT, let the bot chat on its own, and close deals automatically.

Specific steps

Step 1: Use a crawler to obtain content lists and content comments.
Step 2: Process the data. The purpose of this part is to remove redundancy and adjust it into a format that ChatGPT can more easily accept.
Step 3: Use the ChatGPT API to give targeted replies to the processed content.

Notes

There are several difficult points in this design that need attention. I will give two examples of common problems.

1. The API and the ChatGPT web version are not the same model, and the API model has not been trained, so using prompts to operate the API is somewhat difficult. Look this up online; it is fairly complex, and this site has not yet provided these materials. I will write about it later when I have time.
2. Pay attention to the API model you choose and to the price. Before use, start with a small amount of data, and limit the number of inputs and outputs per unit time, so you don't waste tokens if it isn't set up well.
3. API models have different capabilities, so you need to choose a model that suits your project needs. Newer models are certainly easier to use, but they are also more expensive. Find your threshold so you can save money.