Get date from call
# support
m
Good morning, on my outbound assistant it is expected that the called customer can schedule an appointment or alternatively can ask to be called back on a certain day and time. Do you have any suggestions on how I can extract the time and date from the call? Is there a tool or a function that can extract it from the json? Best Regards Mattia
v
You can continue this conversation by mentioning me in the message.
@Mattia Good morning, Mattia! To extract the date and time from the call, you can utilize natural language processing (NLP) techniques to parse the spoken content. If the voice content is transcribed into text, you can then apply date and time recognition algorithms to identify and extract the datetime information. For your outbound assistant, you could implement a system that listens for specific phrases or patterns that indicate scheduling intent, such as "schedule an appointment," "set up a meeting," or "call me back on." Once such phrases are detected, the system would focus on extracting the subsequent date and time information. If you're looking to integrate this functionality into your system, you might need to use a specialized NLP service or library that can handle date and time extraction from text. Unfortunately, there isn't a specific tool or function within VAPI that directly extracts datetime from JSON. However, you can design your system to capture the user's spoken input and then process it with an external NLP service that offers datetime parsing capabilities. If you have the transcribed text and need assistance with implementing the NLP part, please let me know, and I can guide you through the process or suggest some services that could help with this task. Best regards, Your Support Bot
m
@User Good morning, can you suggest me which tools to use? My goal is to make the assistant recognize if the customer asks for an appointment or if he asks to be called back and take the two dates so as to insert them into my database or alternatively switch to Make / Python
s
@Mattia For your use case, there are two main approaches you can take: 1. During the call: Use prompting to have your assistant collect appointment/callback details in a structured way. You can then use tools to take actions in real-time, like updating your database. 2. After the call: Use the end-of-call report which contains a structured analysis of the conversation. You can extract appointment times, dates and other relevant information from this JSON data. Here are the key resources to help implement this: - Prompting Guide: https://docs.vapi.ai/prompting-guide - Tools Documentation: https://docs.vapi.ai/tools-calling - End-of-call Report: https://docs.vapi.ai/server-url/events#end-of-call-report The end-of-call report is particularly useful as it will contain all the structured data about dates, times and appointment details that were discussed during the call, making it easy to integrate with your database or automation tools like Make/Python.
3 Views