Based on my understanding of your scenario description, highlighting the following approaches:
You may leverage Azure OpenAI - this guide shows how to perform document summarization by using the Azure OpenAI GPT-3 model. Kindly take a look at the document architecture and workflow.
Query-based document summarization
You may use Azure Cognitive Search to extract information from unstructured PDFs and create document summaries. With Azure Cognitive Search feature "Skillsets" that allows you to define a set of skills to extract information from unstructured data.
To create document summaries, you may use the "Text Analytics" skill to extract key phrases and sentiment from the PDFs, and then use the "Custom Entity Recognition" skill to identify important entities in the text. You can then use this information to generate a summary of the document.
You can find more information on how to create a skillset for document summarization in the Azure Cognitive Search documentation.
You can also use the Cognitive Search Enrichment Pipeline to perform additional processing on the extracted content, such as entity recognition or sentiment analysis, to further enhance the quality of the document summaries.
Kindly let us know if you have further questions.