OpenAI with Azure- Use Office calls/meetings wisely

Β·

2 min read

Have you gone through the KT sessions ..? Hushh ..!! After couple of hours of KT session, everybody has same thing in mind,What we learnt from the call.

Again, everybody goes through the recording whenever we need a summary or some solution.

What if I tell you , we are able to extract the solution by skipping all of the above or if we skip the call, we will be able to generate a summary from recording. Sounds cool.... right ?

This real time scenario indicated GenAI use cases respectively.

Convert Speech to Transcript:

  • Convert meeting conversation into text.

Approach:

Using Speech Service of Azure, We can easily convert real time voice into text or any audio file (.mp3,.wav) into text. Basically Azure OpenAI co-develops the conversion APIs with OpenAI and deploys the model on azure instance.

This service supports multiple language frameworks like Python,C# ,.net

Getting Started with Speech to Text β€” Taygan

Reference:

https://speech.microsoft.com/portal

Demo :

https://speech.microsoft.com/portal/speechtotexttool

Transcript Usage:

  • From the Transcript, Create Summary of points discussed in meeting.

  • After analysing Summary , If User want to schedule another meeting.

  • Create agenda for the next meeting from summary

Approach :

Creating a summary from a transcript using OpenAI can be achieved by leveraging the capabilities of natural language processing models like GPT-3.5.

To use OpenAI's GPT-3.5, you'll need to have access to the OpenAI API. You'll need an API key, which you can obtain from the OpenAI platform.

We can send prompt to get precise summary with relevant words and sentences.

There are three different Algorithms along with Hypertuning parameters.

Explaining each algorithm and fine tuning itself is a big topic.

Please refer Language Service from to get brief.

Reference:

https://language.cognitive.azure.com/

Demo:

https://language.cognitive.azure.com/tryout/summarization

We can able to create any of this functionality without giving single penny by using free account of Azure upto certain limit. One can explore this anytime as a curiosity as well no need to have technical expertise.

And there you have it, folks, the AI-powered rollercoaster ride through the OpenAI wonderland! πŸŽ’πŸ€–βœ¨ . This is just a single ride, many more upcomings are waiting.

As we wrap up this journey through the world of artificial intelligence and language models, remember that the future is as bright and exciting as a disco ball on New Year's Eve. So, keep those creative ideas flowing, your neural networks buzzing, and may your data always be labeled correctly. Until our next data-driven dance party with OpenAI, stay funky and code on!" πŸ’ƒπŸ•ΊπŸŽ‰

Β