importing plain language and summarizing then writing from a diverse set of data based on AI prompts

Ross Youngs 0 Reputation points
2023-03-25T16:47:07.24+00:00

What I would like to be able to do is import articles, internal summaries, white papers, and other documents. and then use an AI prompt to write a new article or to write a specific question with a well-written result. This will be for summarizing grant applications and other work where I want AI to use best practices to provide a result.

Word
Word
A family of Microsoft word processing software products for creating web, email, and print documents.
822 questions
Azure Cloud Services
Azure Cloud Services
An Azure platform as a service offer that is used to deploy web and cloud applications.
693 questions
Azure OpenAI Service
Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
2,936 questions

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. Joel Borellis (gametolearn) 195 Reputation points
    2023-03-28T20:59:13.0366667+00:00

    If I understand you want to use your own corpus of information/data and have the AI look to that information when constructing an answer/completion? If yes, this is a use case for embeddings with semantic search. Meaning you can create embeddings (vector representations of your white papers, articles, etc.) and then when you construct the AI prompt you do a search against your data and the results of that search are used for the prompt to the Completion request. Here's some more details from Microsoft and OpenAI's docs on embeddings. I think this is what might help you "import" your own corpus of information.

    Embeddings with Azure OpenAI - https://video2.skills-academy.com/en-us/azure/cognitive-services/openai/how-to/embeddings?tabs=console

    OpenAI Embeddings documentation - https://platform.openai.com/docs/guides/embeddings/what-are-embeddings

    1 person found this answer helpful.
    0 comments
  2. kobulloc-MSFT 26,246 Reputation points Microsoft Employee
    2023-03-25T17:48:21.49+00:00

    Hello, @Ross Youngs !

    My recommendation is outside of the scope of Azure but falls into a bit of a gray area so I'll do my best to help you get started.

    How would I import natural language documents and articles and then use AI prompts to write new articles or Q&As?

    I've added the Azure OpenAI Service tag as this is a rapidly changing field with be other approaches worth exploring (like Azure Cognitive Service's document summarization) but Microsoft Power Apps and Microsoft Power Automate may give you the combination of AI functionality that you are looking for.

    There is a workflow that is outlined in detail in this Power Automate blog that has undoubtedly become more simplified in recent months but the general workflow is likely still pretty similar:

    https://powerusers.microsoft.com/t5/Power-Automate-Community-Blog/Summarize-Text-from-large-documents/ba-p/1524459

    The Summarize Text API in Cognitive Services extracts key sentences in a document as they relate to the topic so you can quickly understand what the document is about. To use this API, you need to upload the document to a SharePoint library, OCR and extract the text of the document using Power Automate, and then send the text to the summarization API via an Azure Function. The resulting summary text is then updated in a multi-line text field in the SharePoint document library where it is indexed and searchable.

    Workflow for extracting text summarization:

    1-Overview.png

    An example workflow in Power Apps:

    User's image

    0 comments

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.