Yet another example would be variation. Salesforce happily creates merges of data into a sentence. But if we sent that sentence out, recipients begin to notice that it’s always the same. It would be very nice to allow chat, GPT to generate a variation and send that out instead.
Thank you so much for your kind words. Please help us to spread the word on social media so more and more person can take advantage of this FREE content.
The second application of a flow GPD element is to allow a prompt to pre-fill empty input. We are asking to user. It seems it would be nice to present one screen to have pre-filled context for a thank you note (as an example) and then let chat GPT, generate a thank you based on date, amount, donor, demographic, frequency of donations, and recent impact by the org. The user would then get back the propose thank you and be able to edit it before sending.
On unofficialsf, Andre Tawalbeh provided Apex code that calls an older model of chatgpt platform in a flow element. That’s nice, but I want to call the newer version and allow back-and-forth before a final answer is approved. I attempted to define a HGTV call out with the new beta flow feature, but somehow the Jason parsing is failing fellow, MVP, Hall of Famer, here, would love to get someone to collaborate with me, so nonprofits all over the world could have a chat GPT component to use much as you have shown a component there, except one that preserves context, or allows resetting contacts for a new exploration.
This is interesting though the number of api requests is also limited unless you can show an example if sending multiple questions in 1 call. Do you have a similar example with EinsteinGPT instead of ChatGPT of is there none? I also look for an example of Einstein Discovery in code.
Using VS Code to deploy to our Sandbox environment; however, what is the easiest way to deploy instead of going into each file and trying to deploy? Thanks
is it possible to perform Dynamic Translation with in Einstein Bot where a user who speaks one language can chat with a support agent who speaks another.
One possible solution for this use case is that you can use javascript library language to covert speech to text and then send that text to chat GPT and get a response , but you have to mange the time-out errors.
Platform openai is not free right? I created an account and I got an error after my first request: 429: You exceeded your current quota, please check your plan and billing details.