Hello Everyone,
We are currently using ChatGPT#4 for our chatbot and have implemented a feature that involves sharing specific links with users based on their input. However, we require assistance in establishing a system that monitors the successful delivery of these links to our users and facilitates seamless integration with our backend infrastructure. We kindly request your guidance on the following aspects:
Monitoring Link Delivery:
- How can we implement a reliable mechanism within ChatGPT to ascertain whether the shared links are effectively delivered to the users?
- Is there a recommended approach for ChatGPT to communicate with our backend team, providing notifications or data on the status of link delivery?
Backend Integration:
- Are there industry-standard methods, such as webhooks or API endpoints, for ChatGPT to seamlessly interface with our backend system?
- What steps should we follow to establish the integration, ensuring that relevant information regarding link sharing is transmitted to our backend infrastructure?
Individual Link Status:
- Is it feasible for ChatGPT to furnish our backend team with individualized information for each shared link, including its delivery status and pertinent details?
- What would be the most appropriate method for capturing and archiving this information for future reference and analysis?
We wholeheartedly appreciate any insights or code example that you can provide on these matters. If there are any pertinent resources or established best practices available, kindly share them with us.
Thank you.
1 Like
Monitoring Link Delivery:
- How can we implement a reliable mechanism within ChatGPT to ascertain whether the shared links are effectively delivered to the users?
You can use regular expressions to detect whether links are present in the response. Then, before you send the response to the client, you can wrap them with a link tracking solution, e.g. bit.ly or something else, and send the wrapped links instead of the raw ones.
- Is there a recommended approach for ChatGPT to communicate with our backend team, providing notifications or data on the status of link delivery?
I’d suggest the above for this too.
Backend Integration:
- Are there industry-standard methods, such as webhooks or API endpoints, for ChatGPT to seamlessly interface with our backend system?
There are various ways for this. I am guessing the best way would be with plugins, where the API can be instructed to call a plugin. Currently that’s not really available for developers yet, so you’d need to code it yourself.
A nice way I have found is to ask the language model to use a specific syntax, then use regular expressions to detect the presence of the syntax, and handle it before sending a response to the client. For example, I use a prompt to ask the LLM to say “something()” which gets parsed by my code in the backend and injects the results into the response and the conversation history.
- What steps should we follow to establish the integration, ensuring that relevant information regarding link sharing is transmitted to our backend infrastructure?
Depends on the implementation above.
Individual Link Status:
- Is it feasible for ChatGPT to furnish our backend team with individualized information for each shared link, including its delivery status and pertinent details?
This is only my personal opinion, but doing this completely automated would not really fall into the purpose of the language model, IMO. The method I suggested above should be simple enough to begin with, but I’m guessing eventually there may be services or plugins which will help you accomplish this without writing any code.
- What would be the most appropriate method for capturing and archiving this information for future reference and analysis?
Using my described method, you can collect the bit.ly etc analysis into a database and have proper analytics.
2 Likes
Thank you very much for the detailed response!