Hi. We're interested in exploring some of the ChatGPT functionalities within the EM platform. Specifically, does EM have the ability to import a file containing Part Numbers and Long Descriptions, and then generate eCommerce Titles and Technical Specifications?
For instance, I recently used ChatGPT with prompts like "generate tech spec names and values from description" and "generate eCommerce title from the descriptions" on a sample dataset that I provided.
ChatGPT Sample.xlsx (11.5 KB)
. Attached are some screenshots of the results and the Excel file that contains the data.
Currently, EM does not offer a direct method to inject arbitrary data into the OpenAI context window, such as attaching example files. You can work around this limitation to some degree by injecting the file contents directly into the prompt text using parameters, although this solution may be somewhat limited.
Could you provide more details on how you envision the overall layout of the workflow?
In your example, you have a formatted text document as a result containing the desired data. Do you expect EM to generate a data table with columns such as 'Product Type', 'Package Quantity' etc based on the loose text description of the product?
Hi. Ultimately, I am looking to import the file I attached into EM and run the same type of commands and generate the results in a csv, xls file using the ChatGPT action. In my example, I used the web version of ChatGPT to create the command(s) and copied the data from the excel file (attached) into the chat.
