Build a SAAS AI Product with AutoGen | A Customer Survey App

Hello and welcome to an explanation and tutorial on building your first AI SAAS product with AutoGen! We will be creating a team of 4 AI Agents to create a Customer Feedback Survey Application! We will also learn how to set it up so you can build it yourself!

Other Source Information:

Chapters:
00:00 Introduction
00:19 The Plan
01:30 The Config File
01:51 The Agents Code Setup
05:18 Running the Code
06:34 First Test
07:09 Second Test
07:57 Final Test!
09:23 Conversation Review
10:33 Outro

If you have any issues, let me know in the comments and I will help you out!

Comment your games or software you build, I can't wait to see what you make!

[FaithfulStreaming]

You're thinking in 2030! Nice work on these videos!

[mohammadalblooshi]

exactly what i was searching for for past few months, thank you very much and keep it up

[Andromeda_]

Thank you, Tyler! As always, your content is cutting-edge and insightful. Keep up the fantastic work!

[MrAndi]

Finally someone is doing a more advanced tutorial, love it! Please keep doing those videos!!

[theit-unicorn]

Thank you, Tyler! I was looking for something like this. You rock! I really appreciate you sharing the code as well. Keep up the awesome work, I can't wait to see what's next from you.

[justnft]

Great video ! Thank you for showing this, please do a test for a mern stack app if possible ! Thanks again .

[parthasarathyvenkatadri]

Could there be a tester agent who would do the testing of the software.... By running the code and try various inputs and find if there are any errors ..

[VOKorporation]

Thx🎉please make an example in which there will be an application (like a chat) with autogen. to use it not in the pycharm terminal. for example, the same flask

[AnsaarSyed]

Hey @TylerReedAI I saw the whole video there is something i observed is you don't describe each agent using `agent.description=""`, any reason you didn't do that?

[camburley]

I’m not sure I understand why you’d need to identify all the agents/assistants ahead of time. The concept we seem to be pushing towards is being the userProxy and saying the goal (basic prompt) and letting the AI figure out the rest. Why shouldn’t the AI decide which teammates are necessary and the path to completion? What if it makes sense to have 3 critics instead of 1?

[MrMoonsilver]

Is there a possibility to run an "Autogen Inference Server" with an API? I think that could be really powerful.

[Illtronic-hqvg]

I use GPT all the time for coding, but without using this group. I simply ask for what I need, use the code, and return the errors (if any) and ask it to be debugged. Is there any advantage to your method?

[AIinIM]

thanks for this video. Could you make a video with using autogen with Microsoft guidance?

[warkid]

This is great. I'm kinda busy will try later but can you try this with the new code based LLM like deepsake please

[bivaskumardash]

Can you build this use case with latest Auto Gen 2.0 with less code approach

[AnsaarSyed]

Also i am using mistral-7B with your smae code, i keep getting an error`{'error': "'messages' array must only contain objects with a 'content' field that is not empty."}`, also i dont see the view how each agent communicate to each other, it straight gives me the code, also doesn't ask for any human input, do you know why?

[ordinarygg]

Imagine being a customer of his survey, ahahahahahaahah)))