Building conversational interfaces with LLM APIs
In this section, we will create a chatbot that allows users to take pictures of invoices and save the structured data to a Google Sheet. For this chatbot, I will use Twnel, a messaging application that allows us to create conversational automations to automate processes with the extended supply chain. However, a similar implementation can be done with other chatbot builders, such as UChat, Botpress, and others.
The chatbot workflow for the chatbot that uses AI to structure unstructured text extracted from an image using OCR looks like this:
Figure 8.2: Structuring data from an image chatbot flow
This chatbot uses AI to structure unstructured text extracted from an image using optical character recognition (OCR). The workflow is as follows:
- The user sends an image to the chatbot through a messaging platform. This platform needs to allow users to either upload existing images or take new pictures...
