Activity
Mon
Wed
Fri
Sun
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
What is this?
Less
More

Memberships

AI Automation Agency Hub

328.5k members • Free

25 contributions to AI Automation Agency Hub
What if Chatbots could be HACKED? What about Client's Data privacy?
Recently I have been working with my Italian E-Commerce client and he was very much concerned about his customers' data privacy and whether AI Chatbots could be hacked. This is a very genuine concern and I convinced him too on this matter but I want to know your views on this.
5 likes • Mar '24
Cyber security and data protection should be one of the highest concerns. Using backend servers to store your secrets is a must when building for production. Reverse proxies and data validation techniques should always be used to ensure data protection and compliance. Many ways to protect your chatbots but it's 100% needs to be taken into consideration.
Vapi.ai Voice Agent and Outbound Sales Code
Build Voice agents using vapi.ai - Visit their site and sign up to gain access to their voice agents. Create a phone number or import one from twillo. Next create your inbound call agent with the easy to use UI, give it a prompt and actions. Enter the server url which is the FastAPI template in the zip file or github link below. The server will record all messages and write to json file for easy to view, plus you can view all the logs and conversations on the vapi dashboard. Once your create the agent and publish it the agent is live and ready to go. Give it a test call. Make sure you have the fastapi server running if you want to record the messages on the backend. Outbound Sales Script - Ive also included the Outbound Sales python script for auto pilot outbound sales calls. Create your outbound sales agent the process as above then follow the README.md. Enter your leads with contact details into the csv and watch it work. It includes book appointment function to include with your make or Zapier webhook, just create a new webhook and input the server_url in the outbound.py code, along with the vapi assistant id and phone number id. Follow the docs at - Create Phone Call - Vapi GitHub Link - https://github.com/mcd0056/Vapi-Voice-Agent
Date/Time collection in chat
What is the best way to collect date and time information in a chat. Currently suing botpress. And I need to save a date & time field for appointment setting. I am not finding a good way to do so.
1 like • Mar '24
If your using botpress, first you'll need to set two variables for date and time and have a question block to ask for the preferred date and time they want. Then ask gpt in the code block to write the code to extract the time and date from the saved variables workflow.time & workload.date something like that. Then parse that information using ask ai to format to ISO 8601 it's the standard datetime format. Another way is ask the user in two nodes. So first question have what day of week would you prefer, capture that day in the variable. Then what time would you like, capture that time in the variable and ask ai to format date and time to ISO format and output the result. Use that result to book appointment using zapier or make.
Voiceflow Problem
Hi Everyone, I am having a problem in Voiceflow with a Response AI. I initialy gave it a prompt with Variables and an example of what I want its output to look like. I have now deleted that Example and just have my Prompt + Variables, but the output is still using the Example about 70% of the Time. When I run a Test of the response AI it works most of the Time as I want it too, but when I run a Test of the entire Chatbot, it still outputs wrong, from using the Example I gave it. I have made sure that it is set on Prompt only and NOT Prompt and memory. What else could I do.
0 likes • Mar '24
Have you retrained the agent. Click the train button it sounds like its still using memory it has stored. So retraining it might help, if that doesn't work export the agent and start a new one and upload the file into the new agent see if makes a difference.
How to deploy Langchain server
I have built a langchain agent api using FastAPI. It works fine for one request, but when 2 users request simultaneously, it takes very long and till then I get a "Time Out" response in my frontend. Can anyone please help me deploying such an application in production, so that multiple users can use it? I am using the free open ai api with rate limits. Can that be an issue?
1 like • Mar '24
Is your fastapi endpoint setup to use asynchronous functions? The free openai key has rate limits also and that's probably causing the drama. Use a paid version of openai api key and asynchronous functions it will handle many concurrent requests.
1-10 of 25
Alex McDonald
4
57points to level up
@alex-mcdonald-1993
Python and Openai Developer

Active 805d ago
Joined Feb 23, 2024
Powered by